UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A simulation analysis of the passenger check-in system Arnett, John Douglas 1971

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1971_A4_5 A75.pdf [ 5.78MB ]
Metadata
JSON: 831-1.0101735.json
JSON-LD: 831-1.0101735-ld.json
RDF/XML (Pretty): 831-1.0101735-rdf.xml
RDF/JSON: 831-1.0101735-rdf.json
Turtle: 831-1.0101735-turtle.txt
N-Triples: 831-1.0101735-rdf-ntriples.txt
Original Record: 831-1.0101735-source.json
Full Text
831-1.0101735-fulltext.txt
Citation
831-1.0101735.ris

Full Text

A SIMULATION ANALYSIS OF THE PASSENGER CHECK-IN SYSTEM by JOHN DOUGLAS ARNETT B.A., Simon Fraseir U n i v e r s i t y , 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF BUSINESS ADMINISTRATION i n the Faculty °f COMMERCE AND BUSINESS ADMINISTRATION We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1971 In presenting th is thesis in par t ia l fu l f i lment of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library shal l make i t f reely avai lable for reference and study. I further agree that permission for extensive copying of th is thesis for scholar ly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publ icat ion of th is thesis for f inanc ia l gain shal l not be allowed without my wri t ten permission. Department The University of B r i t i s h Columbia Vancouver 8, Canada Date £>£*>7i*:*nA£*?. / f 7 / The objective of this thesis is to provide a comprehensive manage-ment tool that will aid in policy formulation and evaluation of Air Canada's Passenger Check-in System. The tool is a computer simulation model that was verified to ensure that representative information could be derived from the model concerning the state of the real system. The model can be used to deter-mine the effects of a given policy or passenger arrival composition on the three performance objectives of the system; namely, the utilization of the baggage and ticket facilities, a minimum passenger waiting time in the system, and the checking-in of passengers in accordance with procedural policy. The simulation model describes the state of the system and sub-sequent assessment of the effect of a policy on the objectives with their statistics: 1. the average utilization of the facilities, 2. the percentage of passengers exceeding two and a half minutes in any one queue, 3. the average transit time per passenger (summation of delay times in the system). An analysis utilizing the simulation model was undertaken in three areas of policy management. They are: 1. To determine the implications of the behaviour of the Passenger Check-in System on policy formulation. 2. To determine the facility policy that should be imple-mented in order to achieve the present service policy as well as the maximum capacity of the system. 3. To formulate alternative operating policies and test for viabil ity prior to implementation of the policy. The results of the analysis were as follows: 1. The service policy that a minimum of 15% of the Revenue passengers be allowed to exceed 2.5 minutes of waiting time has been formulated correctly. 2. The nature of the system is such that greater utilization of facil i t ies wi l l not be achieved by a nominal increase in the allowable percentage of passengers exceeding 2.5 minutes. 3. The facil ity policy and associated methodology has been formulated so that the objectives of the system wi l l be attained. A. The maximum capacity of the system occurs when the arrival rate is in excess of 200 passengers per 15 minute period. 5. The use of a single queue at the Revenue ticket counter wi l l ensure greater attainment of the system objectives than the use of multiple queues. 6. The combining of the baggage and ticket operations at one counter is a viable alternative in the present system. Page LIST OF TABLES v i LIST OF FIGURES v i i CHAPTER I INTRODUCTION . . . 1 The Sytems Concept 1 A Specific Application: The Airport System 5 The Thesis Defined 11 II SYSTEMS IDENTIFICATION 13 A Review 13 The Passenger Check-in System 16 A. Definition and Isolation of the System 16 B. Boundaries of the System 28 C. Formulation of the Objectives of the System 28 III SYSTEMS ANALYSIS 31 A. Statement of the Problem 32 Objectives of the Analysis 33 B. Design of the Analysis 35 Part 1 - Why Simulation 35 Part 2 - Construction of the Model 36 Part 3 - Verification 51 Chapter Page C. Analysis of the System 68 Step 1 70 Step 2 78 Step 3 82 IV CONCLUSIONS 90 The systems Approach Concept 90 Conclusions of the Systems Analysis . . . . . . . 91 Areas of Future Research 93 BIBLIOGRAPHY 94 APPENDIX I - Air Canada Facilities APPENDIX II - Flowcharts of the Passenger Check-in System APPENDIX III - Program Listing and Sample Output APPENDIX IV - USER'S MANUAL APPENDIX V - Report to Air Canada Table Page 3.1 Group Type Distribution 42 3.2 Group Size Distribution . 43 3.3 Mean Service Times for Baggage Counters 48 3.4 Parameter Allocation 54 3.5 Simulated Arrival Stream as Compared to Observed . . . . 61 3.6 Comparison of System Statistics 62 3.7 Averages of System Statistics 65 3.8 Compatibility Tests for Optimal Policies 67 3.9 System Statistics from Analysis: Step 1 - Revenue Passenger System 72 3.10 Facility Policy for Various Arrival Rates 79 3.11 Facil ity Policies and System Statistics 80 3.12 Revenue/Standby Ticket Operations in Full Separation . . 83 3.13 Comparative Statistics for Revenue Ticket Counter under Various Operating Policies 86 3.14 Standby Operations Compared 89 Figure Page 1.1 The System as a Black Box 2 1.2 The Airport as a Black Box 6 1.3 Systems in Operation at the Airport 7 1.4 Strata of Airport Systems 9 2.1 Translation of General to Specific System 17 2.2 Arrival Stream 21 2.3 Revenue Passenger Baggage System 22 2.4 Revenue Passenger Ticket System 23 2.5 Standby Process Stream 26 3.1 Distribution of Load Factors 41 3.2 Arrival Stream Specification 45 3.3 Check-in Process Specifications 46 3.4 Probability Distribution for Ticket Counter Service Times 49 3.5 Search for Steady State - X = 100 57 3.6 Search for Steady State - X = 50 58 3.7 Effect of Arrival Rate on Transit Time . 73 3.8 Arrival Rate and Corresponding Utilization 74 3.9 Arrival Rate and % Passenger Units exceeding 2.5 Mins. . 76 3.10 Average Utilization and % Passenger Units exceeding / 2.5 Min. Waiting 77 Figure Page 3.11 Multiple or Single Queues 85 3.12 1 Standby Agent Performs Both Ticket and Baggage Operation 88 I wish to express my sincere thanks to Assistant Professors D. H. Uyeno and J. B. Sidney of the Management Science Department, Faculty of Commerce and Business Administration, the University of British Columbia. Dr. Uyeno, the committee chairman, was of great assistance in the preparation of my thesis. Dr. Sidney was a motivating force in the undertaking of this thesis. Once again, thank you both. I would also like to express my gratitude to Air Canada, in particular, Mr. C. Morgan, Passenger Service Co-ordinator, Vancouver. His assistance and advice was of great value. INTRODUCTION THE SYSTEMS CONCEPT The "systems approach" concept has been utilized since the middle of the 1930's as a method to handle complex and dynamic problems. The methodology employed is commonly pragmatic in nature; and its output is generally designed for the ultimate user. Due to this pragmatic approach, there are two logical divisions in the procedure. First , there is the identification or definition of the problematic system. Second, there is the system analysis or the determination of a solution for the system. The first stage has three parts: A. Definition and isolation of the system; B. Determination of the boundaries of the system; C. Formulation of the system's objectives. The first part, one of major importance, is the explicit defini-tion of the system. For the purpose of explaining this primary step of the system approach, the common analogy of the 'Black Box' wi l l be used. A Black Box is a system which receives certain inputs and produces out-puts. The conversion or transformation of the inputs into outputs is accomplished through the interaction of the components located within the Black Box. (See Figure 1.1). Outputs Figure 1.1 - The System as a Black Box This definition should identify the elements of flow (inputs and outputs) and the conversion process (relationship of inputs to out-puts). The behaviour of the system should be defined as to its degree of predictability. The second part is the determination or setting of the bound-aries of the system that wi l l be analyzed. Consider once again the Black Box concept. Any Black Box can be sub-divided into smaller components or sub-systems. By the same token, each system can be sub-divided into smaller sub-systems or i t may be a sub-division of a larger system. Thus, there is a hierarchy of systems and the specific system under consideration must be identified as to its limits. The third and final step in the identification of a system is the establishment of the operational objectives of the system. This Inputs BLACK BOX Hare, Van Court, Systems Analysis (New York: Harcourt, Brace and World, 1967), Chapter I. formulation of standards wi l l be employed to measure the system's per-formance. It is also a logical key to the development of a useful and efficient analysis. The system must be defined as to what i t is supposed to do and in what manner. The second stage of the systems approach is the analysis of the system defined. It also has three parts: A. The statement of the problem and the formulation of the objectives; B. The design of the analysis; C. The determination of viable alternatives. A statement of the problem wi l l generally be directed towards the objectives of the system. Are these objectives being met? Is there a problem in the control of the system? The formulation of the objectives of the analysis w i l l give formal direction to the analysis, define the ultimate user of the analy-sis and highlight the areas of research. This is an explicit statement of how the analysis wi l l be carried out. The second area is the design of the analysis. A model of the system may be formulated to test the system under various operating conditions in order to determine those conditions which lead to optimal performance. Often these models are computer-based. This model must be consistent with the f irst . The result of the third part is the determination of what might be accomplished in the real system. The presentation of the viable alternatives wi l l enable the decision maker to appraise the problematic system and implement a decision. In summation, then, the Systems Approach i s : 1. Systems Identification A. Definition and isolation of the system. B. Determination of the boundaries of the system. C. Formulation of the system's objectives. 2. Systems Analysis A. Statement of the problem and the objectives of analys i s . B. Design of the analysis. C. Determination of viable alternatives. Turning the discussion from the theoretical to a specific appli-cation should clarify the systems approach. The application of the system identification to the Airport is simplistically modelled in Figure 1.2. The input/output flow elements of the system are defined by their state - either arriving or departing. In describing a real system, there is an associated time element - t^. The arrivals (t^) must occur before the departures (t^) in the modelled system as well as in real systems (t^ _< t^). Just as the general system identification was expanded to reveal its components or sub-systems, so might the Airport system. The two major sub-systems are identified by the kinds of flow elements - passen-gers and aircraft. Just as the Black Box might be divided into s t i l l smaller, more functionally oriented components, so can each of the major sub-systems of the Airport. Consider the Passenger Flow System. There is a three-fold sub-division of this system in operation at an airport. Sequentially defined, there are the following sub-divisions: 1) the Check-in System; 2) the Baggage Flow System; and 3) the Gate Procedure. This division is represented in Figure 1.3. The three systems interact continuously to form the Passenger Flow System. Passengers may either arrive or depart by plane. The Check-in System is the process of validating tickets, allo-cating excess baggage charges and checking baggage. This system initiates part of the Baggage Flow System. The boundaries of this system are ARRIVALS Time Element < t. AIRCRAFT SYSTEM PASSENGERS SYSTEM Figure 1.2 The Airport as a Black Box MAJOR SUB-SYSTEMS DEPARTURES AIRPORT Flow Elements: Passengers and Aircraft AIRPORT AIRCRAFT FLOW SYSTEM GATE & TAXI PROCEDURE 7 MAINTENANCE CARGO HANDLING CHECK-IN BAGGAGE FLOW GATE PROCEDURE Figure 1.3 Systems in Operation at the Airport PASSENGER FLOW SYSTEM defined by the state of the passengers. As an input to the system, passengers queue at one of the available counters and become outputs when check-in is completed. The Baggage Flow System transforms terminal baggage checked by passengers to on-board cargo and vice versa. The third system operates at the Departure Lounge of the airport. Passengers receive a final ticket validation and seat selection is made. The output of the system consists of specific passengers on board specific flights. If the flow element were defined as an aircraft, then there are three sub-systems present: Gate and Taxi Procedure, Cargo Handling, and Maintenance. These three systems may be described in connection with the state of the aircraft in the system. For example, consider the Gate and Taxi Procedure System. When an aircraft arrives i t must taxi by a definite path and terminate at a particular gate. When the aircraft departs, the reverse procedure is followed. The other systems are as their name would suggest. The hierarchy of the airport systems may be classified into various strata (see Figure 1.4). A system which has been identified as being of the third strata, is of the appropriate size and dimension for a viable system analysis (the Passenger Check-in System, etc.). The systems have succinct and readily definable boundaries, objectives, inputs and outputs. If optimal methods are formulated and implemented for each sub-system, then the entire system wi l l advance closer to a state of optimality. R. Mountjoy, "Airport Simulation Models," AGIFORS Proceedings, 1969, pp. 619-42. AIRCRAFT MAINTENANCE CARGO GATE, TAXI PROCEDURE ON-LOAD OFF-LOAD Figure 1.4 Strata of the Airport Systems GATE PASSENGER CHECK-IN REVENUE Strata 1 Strata 2 BAGGAGE Strata 3 STANDBY Strata 4 TICKET BAGGAGE The justification for a study of airport systems is quite apparent i f a major airline terminal has been observed during periods of peak volume. The main problem results from simultaneous arrivals and depar-tures of aircraft. The resultant volume of passengers and aircraft tends to 'clog' the systems at the airport. Specifically, there are several reasons for a study to be under-taken. 1. There has been an increased demand for air transportation. Airlines have met this demand with expanded fleets and increased capacity of aircraft. As a result, greater volumes of passengers must be pro-cessed through air terminals. 2. There must be more sophisticated technologies employed to co-ordinate, plan and control the volumes of traffic at airports. The question remains as to the selection and implementation of these tech-nologies for viable results. . 3. The air transport industry is a highly competitive business. The airline which supplies quality service both in the air and on the ground, while maintaining lower costs, wi l l benefit. This can only be achieved through efficient use of resources in their systems. The general direction of this study has been indicated in the previous two sections of Chapter I. What remains is to define the exact nature and scope of the thesis as i t wi l l be presented. The systems approach wi l l be applied to a readily available and familiar system - the Passenger Check-in System. The methodology used in this thesis wi l l be applicable, in the most part, to any major airline check-in system. The specific details of the thesis w i l l be derived from the Air Canada system, Vancouver International Airport, Richmond, British Columbia. Chapter II wi l l present the first part of the systems approach - the Passenger Check-in System Identification. The elements, boundaries and objectives w i l l be discussed to enable a greater understanding of the system's internal operations. Chapter III w i l l present the detailed analysis of the system. As part of the analysis a computer simulation model wi l l be constructed. Experiments wi l l be performed on the model in the area of service policy formulation and evaluation. Appendix III contains a program listing of the model and Appendix IV, a User's Manual for the model. Chapter IV contains a summary and conclusions. The conclusions of the thesis w i l l be pertinent to both the present system and future research on the system. Appendix V contains a management summary directed towards the Management of Air Canada. There is a twofold objective of this thesis: 1. To provide a comprehensive tool that wi l l make possible the testing of alternative mangement policies. 2. To f u l f i l l the requirements for the Master of Business Administration degree at the University of British Columbia. SYSTEM IDENTIFICATION This chapter wi l l deal primarily with the identification of the Check-in system. To begin with, a brief review of previous studies concerning the Passenger Check-in system would be appropriate. This is a logical step for any analyst since i t could provide insight into the methodology employed for problematic areas. A REVIEW The five studies reviewed below were extracted from the AGIFORS Proceedings (The Airline Group, International Federation of the Operations Research Society) from 1967 to 1969.1 1. Tit le : Optimal Staffing at Airline Terminals, 1969, pp. 15-42. Author: K. C. Khanna and H. Takamori (airline not mentioned). Objective: To determine the manpower requirements and shifts for the various functions at the airline terminal. Technique: The problem was formulated as an optimization over a large number of small integer problems. Subsequent use of branch and bound methods. 2. Tit le : Airport Simulation Models, 1969, pp. 619-642. Author: R. Mountjoy, United Airlines. An annual symposium of the major airlines is held for the purpose of discussing studies on problems encountered and their respective solutions. Objective: To handle inter-airline questions concerning (a) runway capacity; (b) taxi times and terminal configurations; (c) airline schedules, terminal facil i t ies and station policies affecting customer service; (d) airport facil it ies response to increased traffic forecasts. Technique: Subdivide the Airport system into several simu-lation models to handle complex inter-relationships. 3. Tit le : Airport Manpower Planning Systems, 1969, pp. 681-696. Author: V. K. Wozniuk, Air Canada. Objective: to provide management with accurate short-range forecasts (up to two or three years) of manpower requirements at certain airport positions to be used for planning of budgets and agent training. Also to provide management with a means of evalu-ating the proposed schedules and various standards. Technique: Sub-divided AMPS (Airport Manpower Planning System) into three areas: (a) Check-in Counter - arrival stream and its properties; - service time distribution; - queue discipline and system type. (b) Departure Gate - departure requirements for five-minute intervals. (c) Arrival Gate - same as departure gate. Uses classical queueing methods to determine requirements. 4. Tit le : Manpower Planning for Airport Operations, 1968, pp. 245-267. Author: S. O'Broin, Air Lingus. Objective: To determine manpower requirements for the Passenger Flow system at various levels and service standards. Technqiue: Uses classical queueing theory for requirements. Scheduling is determined by an algorithm. 5. Tit le : Passenger Handling Model, 1967, pp. 52-80. Author: W. Gensema, Royal Dutch Airlines (KLM). Objective: To determine the number of service counters required for a specific standard of service. Technique: The derivation of the model uses both classical queueing theory and Monte Carlo simulation. The studies that have been cited here, are not exercises in the use of pure theory, but are practical approaches to the determination of viable solutions for problems. The techniques varied depending upon the problem, but when queueing situations were encountered, classical queueing theory was utilized extensively. Unfortunately, the use of this body of mathematic theory becomes extremely complex, i f not impossible, when applied to large systems. Many of the objectives of a detailed systems analysis are unattainable, because of the assumptions and limita-tions of the theory. While the theory is applicable for static analysis of a homogeneous system, the theory does not represent the dynamic behaviour which characterizes a real system. The queueing behaviour of real and complex systems is extremely difficult to analyze with classical queueing theory due to the simultaneous movement of elements in the various queues. On the other hand, a simulation model of a queueing situation has the ability to handle heterogeneous elements in queues and to reveal the true behaviour of a dynamic system. THE PASSENGER CHECK-IN SYSTEM The system identification division, as previously stated, contains three main steps: A. Definition and isolation of the system; B. Determination of the boundaries of the system; C. Formulation of the objectives of the system. Each of these steps wi l l be presented and explained in turn as the study is directed towards the application of the systems concept to the Passenger Check-in System. A. Definition and Isolation of the System To the casual observer at the airport, the arrival of passengers for check-in, prior to flight departure, is at times confusing to say the least. However, there are patterns of behaviour that can be discerned so that the whole system may be defined. As stated in Chapter I, the check-in system may be described generally as a flow system with input elements, a transformation or conversion process, and output elements. The translation of the general system definition to that applicable for the check-in system may be seen in Figure 2.1. The input elements are passengers but the process is defined as the arrival stream. The con-version of inputs to outputs is the check-in process consisting of baggage and ticket counters. The output element of the system may be defined simply as a passenger that has been checked in . The Arrival Stream. The arrival or input stream for the check-in system may be defined by the number of passengers entering the system. ARRIVAL STREAM PASSENGER ARRIVAL INPUT ELEMENT CHECK-IN PROCESS CONVERSION PROCESS DEPARTURE STREAM PASSENGER, CHECKED-IN OUTPUT ELEMENT Figure 2.1 Translation of General to Specific System The passengers for a particular flight usually arrive in the 1 1/2 hour 2 time period prior to departure. The number of passengers that arrive in a given time interval within that period is a function of the load factor (ratio of the number of passengers to maximum capacity of the aircraft)and the amount of time until the aircraft departs. It was observed that a certain percentage of the passengers arrived in each of the six 15-minute periods. This percentage was observed to remain approximately constant for a particular flight even though the load factor varied. For example, i f the arrival pattern for a DC-8L (198 passenger capacity) is such that 30 per cent of the passengers arrive in a certain 15-minute period, then 47 passengers wi l l arrive i f the load factor is 80 per cent and 35 i f the load factor is 60 per cent. More wi l l be said about this later. The arrival pattern of passengers for different flights overlap due to the scheduling of departure times. Thus the pattern might be such that in a 15-minute time interval 50 passengers for flight X arrive as well as 45 passengers for flight Y. The arrival stream consists of 95 passengers for the particular 15-minute period. Another characteristic of the arrival stream is that passengers travel in groups. The groups themselves are independent upon arrival at the airport i f the assumption is made that the small number of groups that do arrive by bus do not affect the arrival pattern significantly. The group size may range from one to six passengers on the most part. S. O'Broin, "Manpower Planning for Airport Operations," AGIFORS Proceedings, 1968, pp. 245-67. The occurrence of groups larger than six is a very small proportion of the total group arrivals. The arrival rate of groups may be approximated by a Poisson distribution since the group arrivals are independent. The number of group arrivals is then the number of passengers divided by the average group size. The passenger groups arriving at the airport are of two types: independent or dependent. The distinction is made because of the method of processing passengers at the check-in counters. The passengers of a group may be checked-in independently or dependently. If the group is of the independent type, the passengers of that group are processed individually. A group of size one, w i l l , by definition, be independent. Another example of an independent group would be two or three business associates travelling together. When they arrive by taxi, they do so as a group, but are processed at the counters individually. The dependent groups that arrive at the airport are families. They are processed at the check-in counters as one. Due to the method of processing the passengers, the flow elements wi l l be defined as "passenger units." A family or one businessman in a group represents one passenger unit. The final characteristic of the arrival stream is the identi-fication of the passengers travelling at fu l l fare (Revenue) or reduced fare (Standby). Revenue passengers have confirmed seats while Standby passengers travel only i f space is available. This characteristic deter-mines whether the group wi l l enter the Revenue or the Standby Passenger System. It was observed that passengers within groups rarely were of both classifications (Revenue and Standby in one group). Families were always directed to the Revenue system. The flow chart of the arrival stream appears in Figure 2.2. In summation, the arrival stream may be identified statistically by: 1. The number of passengers entering the system in a specific period; 2. The group size and type distribution for a particular airport; 3. The average number of Revenue passengers as a percentage of total for a particular flight. The Check-in Process. In Figure 2.2 i t was noted that a passen-ger group may enter either the Revenue or Standby System. The two systems differ because of the operating policy employed by Air Canada. The operating policy for a system is the set of methods used to direct or control passenger movement through the system in a prescribed manner. As this thesis identifies the Revenue and Standby System, these operating rules wi l l be clarified. The Revenue System. Figures 2.3 and 2.4, the flow charts of the passenger flow for the Revenue check-in process, indicate the operating rules which are in effect. As groups of passengers enter the airport, the first operating rule is that the passengers must have a ticket before their baggage may be checked. A small percentage of Revenue groups arrive non-ticketed. If the group is non-ticketed, they go to the Revenue Tickety System - another component of the Revenue System. (This system wi l l be explained later.) GROUP ARRIVAL Figure 2.2 Arrival Stream CURBSIDE CHECK-IN NO M UPTS BAGGAGE COUNTER CHECK-IN YES -J RPTS Figure 2.3 Revenue Passenger Baggage System RPTS = Revenue P_assenger T_icket System Figure 2.4 Revenue Passenger Ticket System RPBS = Revenue Passenger Eiaggage System Continuing the flow, the next operating rule is that i f a passen-ger group does not have excess baggage, a "red cap" may be employed to tag the baggage and deliver i t to the Air Canada faci l i t ies . This is called "Curb-side Check-in" since i t takes place at the terminal side-walk. The passengers do not have to be served at the check-in counters, and may proceed to the appropriate departure lounge or explore the terminal shops. A small percentage of passenger groups use this check-in procedure. Passengers having a ticket then select an agent for service at the Revenue baggage counter. The queueing discipline may be described as the selection of the shortest queue. It was observed that business associates or friends travelling together tended to select a queue as a group. Families must queue for a single facility as they are processed dependently. The changing of lines was minimal because passengers had to move their baggage. When the baggage agent serves a passenger unit, the baggage is tagged with the appropriate flight number, the ticket is validated and the weight of the baggage is checked. If the weight exceeds a certain limit, the passenger unit is defined as having "excess baggage." The next operating rule is that i f there is excess baggage, the passenger unit proceeds to the ticket counter to pay the excess charge. If there is no excess baggage, the passenger unit has completed the check-in process and proceeds to the Departure Lounge. The facil ity utilization (the percentage of time an agent was utilized for service to the total time interval) is dependent on the facility policy and the demand for service at the counters. The facil ity policy which specifies the number of counters that wi l l be available for passenger service is set by management. The service time for the baggage counters (the length of time an agent is employed in service with a passenger) is dependent upon the size of the passenger unit and whether or not there is excess baggage. As indicated in Figure 2.3, there are two procedures by which a passenger might enter the Revenue Ticket System - a passenger group arriving at the airport non-ticketed or a passenger unit having excess baggage charges to pay. This distinction is evident in Figure 2.4 as the passengers flow through the system. The queueing discipline in the Revenue Ticket System is similar to that of the baggage counter. When a passenger purchases a ticket, only the Revenue ticket counter may be used. The Standby ticket counter may also be utilized for excess payment. The service times are dependent upon the nature of the service required. The mean service time for ticket purchases was greater than that for excess baggage payments. Once the ticket is purchased, the passenger unit proceeds to the Revenue baggage counters. If excess payment has been made, the passenger unit has com-pleted the check-in requirements. The Standby System. The operating policies of the Standby System are represented in Figure 2.5. The procedure is similar to that of the Revenue Passenger System, except there is no option for curbside check-in. Passengers must have times and priorities concerning their Standby status determined at the baggage counter. The ticket procedure is the same as that of the Revenue System. Non-ticketed groups proceed to the "Standby PURCHASE TICKET BAGGAGE CHECK-IN PAY EXCESS CHARGE Figure 2.5 Standby Process Stream Ticket Purchase" counter. Passenger units with excess charges to be paid may uti l ize the Revenue ticket counters as well as the Standby ticket counter. To aid in the conceptualization of the system, a floor plan of the facil it ies with the flow of passengers superimposed on the floor-plan appear in Appendix I. In summation, the Check-in System may be statistically defined by: 1. The operating policies in effect; 2. The distribution of service times at the various facil it ies and the type of service demanded; 3. The percentage of passenger units who arrive non-ticketed or use curbside check-in, or have excess baggage; 4. The number of facilities that are available for use in the system. The Output. The output of the Check-in System are passenger units that have met a l l the requirements of the check-in procedure. The passenger units then proceed to the departure lounge for a final ticket validation and seat selection. If time permits, passengers may browse in the airport shops. Associated with the output, there are system statistics. These statistics describe the flow elements of the system, given the definition of the arrival stream and check-in process. The transit time statistic is the time a passenger unit has spent travelling through the system. Another statistic is the aforementioned facility utilization. Other statistics may be formulated depending on the objectives of the analysis. B. Boundaries of the System The second step of the system identification is the determination of the boundaries of the system. The essence of this step is to structure the system for a concise analysis by defining what is exogeneous to the system when the analysis is undertaken. The input boundary should be considered as the mean arrival rate of passengers. No attempt w i l l be made to statistically determine the functional relationship between the expected and actual passenger arrivals. This figure - the number of passenger arrivals - w i l l be an exogeneous variable to the system. The procedures of the agent wi l l be dealt with as just a dis-tribution of service times. If certain procedural changes were to be made which would effect this distribution, the system could be restruc-tured to reflect the change and subsequent state of the system. The output boundary is defined by the state of the passengers as they pass through the system. Once the passengers have completed the requirements at the check-in counters, they are deemed to be an output element. No examination wi l l be made of the Departure Lounge at this time. The results of the systems analysis could be used as inputs to a much larger systems analysis. C. Formulation of the Objectives of the System The final step in the identification of the Passenger Check-in System is the formulation of the objectives. These objectives wi l l be vised to measure the performance of the system and to determine the desir-ability of alternative methods of operating the system. There are three main performance objectives for the Check-in System: 1. To attain a high utilization of manpower in the system. 2. To achieve a minimum customer waiting time in the queues. 3. To check in the passengers properly. There is undoubtedly a non-linear relationship between manpower utilization and customer waiting times. The cost of service is related to the utilization of manpower. A low utilization is representative of a high cost of service. Only when there is a small number of passengers in a system and the resultant demand for service is low, does this relation-ship become unrepresentative of variable costs. The functional relation-ship w i l l be examined in Chapter III. The strategies or service policies that have been formulated by Air Canada are the means of achieving the systems objectives. These service policies are reflected by the operating rules described earlier and are listed below. 1. Separate the Standby passenger service from that of the Revenue passenger. Since Revenue passengers comprise a large per-centage of travellers and pay ful l fare for their tickets, they are the most important. If Revenue passengers become dissatisfied, hopefully i t w i l l not stem waiting for service while a passenger who has paid a lower fare is being served. 2. Staff the f a c i l i t i e s such that 85 per cent of the Revenue passenger units and 75 per cent of the Standby passenger units do not wait longer than 2.5 minutes in any one queue. This i s the standard of service set by Air Canada. It should be mentioned that due to the importance of the Revenue passengers, the 15 per cent who wait longer than 2.5 minutes is a maximum number. The Standby passengers are of lesser importance and the 25 per cent should be considered as a working figure and not a s t r i c t maximum as i n the case of Revenue passengers. 3. Use a 15-minute staffing horizon to e f f i c i e n t l y and effectively allocate manpower and forecast manpower requirements. This strategy i s employed because the demand for service fluctuates greatly over the day. A 15-minute planning horizon w i l l allow for adequate 3 staffing and transfer of agents to various functions in other systems. The attainment of the objectives may be measured s t a t i s t i c a l l y by: 1. F a c i l i t y u t i l i z a t i o n ; 2. The percentage of passenger units waiting in queues longer than 2.5 minutes; 3. The transit time in the system. Since the system has been defined as to i t s inputs, outputs, conversion process, boundaries and objectives, the second step of the system analysis w i l l be undertaken. The definition of the system under study i s of extreme importance. If a viable analysis is to be accom-plished a thorough understanding of the system both descriptively and st a t i s t i c a l y i s imperative. 3 V. K. Wozniuk, "Airport Manpower Planning Systems," AGIFORS Proceedings, p. 686, 1969. SYSTEMS ANALYSIS As stated in Chapter I, the analysis of a system consists of three steps: A. A statement of the problem and the formulation of the objectives of the analysis; B. The design of the analysis; C. The understanding and determination of viable alternatives for the problematic system. The statement which defines the problematic areas of the system is generally derived from the stated objectives of the system. It should define the problems encountered in operating the system and the possible source. The second part of the step is the formulation of explicit objec-tives for the analysis. These objectives wi l l define exactly what wi l l be undertaken in the analysis. The result of this step wi l l be the methodology used to rectify the problematic system. The second step is a direct result of the previous section. The design of the analysis is the mode or vehicle by which the analytical objectives wi l l be achieved. This is the construction of a laboratory tool that can be used for experimental analysis under controlled conditions. A cr i t ica l part of this step wi l l be the verification of the modelled system. The third step is the analysis of the system as set out in the objectives. An important part of this analysis is the interpretation and presentation of the results. The ultimate user of the results must be kept in mind since the analysis is undertaken for the decision-makers who control the system. (The results of this step is the determination of viable policies for the real system.) The three steps could be considered as a system, the inputs being the statements of the problem and the objectives, and the con-version process being the modelled system. The output would be the ability to test for and determine viable solutions for the problems of the system. A. Statement of the Problem The purpose in defining the objectives of the Passenger Check-in System in Chapter II was to provide a means of evaluating the per-formance of the system. Alternatively stated, i t is desired to determine the degree of attainment of the stated objectives. Policies are formu-lated and implemented in the system as a means of achieving a desired state. The inherent problem of operating a real system is the measure-ment of the effectiveness of implemented policies and the determination i f proposed policies wi l l be viable in the system. The measurement of existing policies to determine i f they do in fact achieve their objective is a very real problem in a dynamic system. It is imperative for the management to know i f their policies achieve their purpose. Efficient control of the system can then be maintained. If new policies are formulated, then i t is necessary to fore-cast their feasibility in the system. The decision-makers have two avenues to follow. One is to formulate the policy and test for feasibility by implementation in the real system. This could be costly and disruptive i f the hypothesis is false. The other avenue is to construct a model of the system so that controlled testing of the policy may take place prior to the implementation. The model used could either simulate or emulate the system. A justification for this wi l l be presented shortly. Specifically, Air Canada's management needs to know the u t i l i -zation of facil it ies and the associated percentage of passengers waiting in queues longer than 2 1/2 minutes. If alternative methods of handling passengers are formulated, what are the effects on the system? What is the processing capacity of the existing facil it ies at the Vancouver Air-port? And finally, in the present system, has Air Canada an optimal faci l i ty policy for the various arrival rates encountered in a 15-minute time interval? These questions wi l l form the basis for the objectives of the systems analysis. Objectives of the Analysis The objectives of the analysis wi l l be used to direct the analysis of the system. If experiments are performed on the modelled system, the answers to questions relating to the system wi l l be known. Also, the relationship of the dynamic elements in the system wi l l be revealed. The first objective is to provide a laboratory tool for the management of Air Canada that wi l l enable the evaluation of policies, both implemented and proposed. This tool w i l l be a computer-based simu-lation model. It w i l l be presented in the second step of the systems analysis - the design of the analysis. The second objective is to determine the relationship between certain variables or systems s t a t i s t i c s . This w i l l aid in the formu-lation of future policies and the understanding of present ones. This part of the analysis w i l l examine the effects of the arr i v a l rates on passenger unit transit time (as determined by the percentage of passenger units waiting in queues longer than 2 1/2 minutes) and on f a c i l i t y u t i l i -l i z a t i o n . The third objective of the analysis i s to provide a f a c i l i t y policy (P^) for the various a r r i v a l rates ( A ^ ) . This policy w i l l attain the system's objectives as stated i n Chapter II. The fourth objective i s to determine, with the use of the simu-lation model, the maximum capacity of the existing f a c i l i t i e s given system objectives. When forecasts exceed this capacity, plans must be made to adjust for the volume. The f i n a l objective of the systems analysis i s to determine i f there are alternative methods of handling passengers in the system. Thus new operating rules w i l l be formulated, tested, and compared to present operating policies. Since the statement of the problem and the objectives of the analysis has been made, the next step in the systems analysis w i l l be presented. B. D e s i g n o f t h e A n a l y s i s The d e s i g n o f t h e a n a l y s i s w i l l be p r e s e n t e d i n t h r e e p a r t s . The f i r s t p a r t w i l l p r e s e n t a j u s t i f i c a t i o n f o r u s i n g a s i m u l a t i o n model i n t h e a n a l y s i s o f t h e P a s s e n g e r C h e c k - i n System. The s e c o n d p a r t w i l l d e s c r i b e how t h e s t a t i s t i c a l d e f i n i t i o n s and o p e r a t i n g r u l e s as i d e n -t i f i e d i n C h a p t e r I I a r e c o n v e r t e d i n t o a s i m u l a t i o n m o d e l. The t h i r d p a r t w i l l b e t h e v e r i f i c a t i o n t h a t t h e computer model does i n f a c t ' s i m u l a t e ' t h e r e a l s y s t e m . P a r t 1 - Why S i m u l a t i o n The j u s t i f i c a t i o n f o r u s i n g t h e o p e r a t i o n s r e s e a r c h t e c h n i q u e known as s i m u l a t i o n i s t h r e e f o l d . 1. The a n a l y t i c a l powers o f t h e t e c h n i q u e overshadow t h a t o f mathe-m a t i c a l t e c h n i q u e s t h a t c o u l d be u s e d i n a s ystems a n a l y s i s o f t h i s t y p e . I n t h e r e v i e w o f p r e v i o u s s t u d i e s w h i c h was p r e s e n t e d i n C h a p t e r I I , i t was p o i n t e d o u t t h a t when complex q u e u e i n g systems a r e e n c o u n t e r e d , c l a s s i c a l q u e u e i n g t h e o r y i s l i m i t e d . S i m u l a t i o n has t h e a b i l i t y t o h a n d l e complex and dynamic s t a t e s i n s u c h a s y s t e m . The v e r y n a t u r e o f s i m u l a t i o n t e c h n i q u e s f o r c e s t h e u s e r t o b e a n a l y t i c a l . Each r e l a t i o n -s h i p must be d e f i n e d e x p l i c i t l y . Thus even b e f o r e t h e model i s c o n s t r u c t e d a sound u n d e r s t a n d i n g o f t h e s t r u c t u r e o f t h e s y s t e m i s r e a l i z e d . 2. The s e c o n d r e a s o n f o r u s i n g s i m u l a t i o n i s t h e v e r y n a t u r e o f t h e s y s t e m under s t u d y , w h i c h may be c l a s s i f i e d as a " f l o w " s y s t e m . The f l o w e l e m e n t s o r p a s s e n g e r s ' f l o w ' t h r o u g h t h e s y s t e m . F o r t h i s r e a s o n , t h e P a s s e n g e r C h e c k - i n System i s r e a d i l y a d a p t a b l e t o many a v a i l -a b l e s i m u l a t i o n l a n g u a g e s , i n t h i s c a s e t h e I.B.M. computer s i m u l a t i o n language - GPSS/360 (General Purpose Simulation System). 3. The third reason is that once the model is constructed i t becomes a very useful management tool to control the system. Relatively inexpensive experiments may be made under controlled conditions. If a policy is to be tested, the decision maker simply specifies the in i t i a l state of the system or flow charts the new path of the passengers in the system. The analysis is then easily performed. Part 2 - Construction of the Model The construction of the model w i l l be divided into two parts. An outline of each part wi l l be given below and then discussed in detail. (a) General information wi l l be discussed concerning the nature of the model, the prime elements of the computer language, and the model structure. (b) In Chapter II, the systems identification was divided into three areas - the arrival stream, the check-in process and the output. Each area had certain statistical or quantitative definitions. This section wi l l convert those definitions into operational elements of the model. (a) General Information The computer language used to code the model was GPSS/360. During the late spring of 1970, the University of British Columbia Com-puting Centre obtained a new compiler called GPSS-V: OS Version. A few changes have been made in the compiler to enable the undertaking of extremely large and complex simulation models. The computer facil it ies at U.B.C. consist of an IBM 360/67 Duplex utilizing MTS (Michigan Terminal System). The general principles of the GPSS/360 program have been extracted from the User's Manual and appear below. Block diagrams or flow diagrams are widely used to describe the structure of systems. They consist of a series of blocks, each of which describes some step in the action of the system. Lines which join the blocks indicate the flow of traffic through the system, or describe the sequence of events to be carried out. Alternative courses of action that arise in the system are represented by having more than one line leaving a block. Conversely, one block may have several lines entering i t to represent the fact that this block is a common step in two or more sequences of events. The choice of path, where an alternative is offered, may be a probabilistic event or a logical choice, depending upon the state of the system at the time of the choice. Both of these methods of selection can be used in the GPSS/ 360 program. The units of traffic that move through the system depend upon the system being simulated. Units might be messages in a communication system, electrical pulses in a digital circuit , work items in a production line, or any number of other units. These units upon which the system operates in the GPSS/360 program wi l l be called "transactions." The GPSS/360 program also has various other entities (facil it ies, storages, queues, tables, etc.) whose attributes are changed by the move-ment of trasactions through the various block types. Although a block diagram is a commonly used means of describing a system, the notation used in normal block diagrams depends upon the system and the person who is describing the system. For the purpose of the GPSS/360 program, certain conventions and systems concepts have been defined, each corresponding to some basic action or condition that generally occurs in systems. Statis-t ical variations may be introduced in the block diagram, and many statistical sampling procedures are provided. Levels of priority may be assigned to transactions and complex logical decisions may be made during the simu-lation. The GPSS/360 program operates by moving transactions from block to block of the simulation model in a manner similar to the way in which the units of traffic they represent progress i n the real system. Each such movement is an event that i s due to occur at some point in time. The GPSS/360 program maintains a record of the times at which these events are due to occur, then proceeds by executing the events in their correct time sequence. When transactions are blocked and cannot move at the time they should, the program moves them as soon as the blocking condition or conditions change. In order to maintain the events in the correct time sequence, the GPSS/360 program simulates a clock that i s recording the instant of time that has been reached in the model of the real system. The number shown by this clock at any instant i s referred to as the "absolute clock time." Another clock time, the "relative clock time" is one of the Standard Numerical Attributes which can be externally addressed by the analyst. A l l times i n the simu-lation model are given as integral numbers. The unit of system time which i s represented by a unit change of clock time i s implied by the user, who enters a l l data relating to times i n terms of the time unit he has selected. Whatever unit of time i s chosen, such as millisecond or tenth of an hour, i t must be used consistently throughout a simulation model. The GPSS/360 program does not simulate the system at each successive interval of time. Instead, i t up-dates the absolute clock to the time at which the next . most imminent event is to occur. The controlling factor in the amount of computing time that is used by the program i s , therefore, the number of events to be simulated, not the length of real-system time over which the simulation is being made.^ The time units of the Check-in model are seconds. This unit was used because of the nature of the service time distribution and length of simulated time. Detailed flowcharts of the system appear in Appendix II. IBM User's Manual GPSS/360, #GH20-0362-3, IBM Corporation Technical Publication Dept., 1970, pp. 4-5. (b) Conversion to GPSS The Arrival Stream. As stated in Chapter II, the arrival stream of the Passenger Check-in system may be defined statistically by: 1. The number of passengers who enter the system in a certain time interval; 2. The type and group composition distribution for a particular airport; 3. The ratio of Revenue passengers to total arrivals in a certain time interval; 4. The operating rules in effect at the airport and the percentage of passenger groups who are effected. Each of these quantitative definitions wi l l be discussed and illustrated. Each departing flight has an associated arrival pattern of its passengers. The distribution defining this pattern is the percentage of the load factor (L.F.) arriving in a particular period relative to time of departure. The '% of L.F.' was observed to remain approximately constant even though the L.F. changed. If the 15-minute time intervals prior to departure are defined as t^ / i =1, 6; (t^ = 0 - 1 5 minutes prior to departure) then the % of L.F. arriving in t^ for flight X may be defined. If flight X has a capacity of 198 passengers and i f L.F. = 80% for a particular sample Then i f 50 passengers arrive in t^, the % of L. F. would be 50 x 100% = 32% of L.F. 198 x .80 Assume the distribution of % of L.F. for flights X and Y were: Time Period % of L.F. t. X Y 1 t 1 10 5 t 2 20 30 t 3 35 40 t. 20 20 4 t 5 10 5 t _ 5 _0 0 100% 100% If the scheduling of flights is such that X departs 15 minutes before Y, the distributions of % of L.F. is depicted in Figure 3.1. The load factor depends upon the time of year, day of the week, des-tination of the flight and time of departure. In a discrete time period of the day T^ , assume the departure time of flight X is T. and flight Y is T, , .. Then the latal arrivals J 3 + 1 into the system will depend on t^(X) + t^ + ^(Y) , where t^(X) and t^(Y) are defined in the obvious manner. t 3(Y) If the load factor of X is 80% and Y is 60% and T - t 2(X) + Then the arrival rate of T will be 20% of (.80 x 198) + 40% of (.60 x 198) = 79 passengers. Thus, this variable, the total number of passengers who enter the system in a 15-minute time period T^  will be the exogenous to study. The thesis will not attempt to determine the expected number of arrivals at the Vancouver airport but justifies how this variable may be obtained. % of L.F. 40 30 20 10 " 5 10 % of 40 L.F. 30 20 10 35 20 20 40 10 t 5 t 4 t 3 t 2 t± 30 h 's fc4 fc3 fc2 ' l Flight X L.F. = 80% Flight Y L.F. = 60% j | i ! i T T T" T A i - 1 i l + l X± + 2 t. = 15 minutes I T. = 15 minutes l Arrivals of = t^ of Y + t 2 of X Figure 3.1 Distribution of Load Factors The next statistical definition of the arrival stream is the distribution of the group type and composition. The group type distribution is used to distinguish between indepen-dent and dependent groups. Independent groups may be groups of size 1 (a person travels alone) or a group wherein the members are processed independently at the check-in counters. The distribution appears in Table 3.1. 2 TABLE 3.1 GROUP TYPE DISTRIBUTION VANCOUVER AIRPORT, 1967 Group Type 1. Independent - Alone Others 2. Dependent - Family % Cumulative % 16.0 16.0 13.5 29.5 70.5 100.0 It should be noted that 70.5% of groups departing on Air Canada flights were family units. 3 Table 3.2 is the distribution of group size. A. T. Wiley, Director, Marketing Intelligence, Air Canada. A study was performed in 1967 to determine group composition and size for Vancouver Airport. A. T. Wiley, ibid. GROUP SIZE DISTRIBUTION VANCOUVER AIRPORT, 1967 Group Size % Cumulative % 1 16.0 16.0 2 34.9 50.9 3 15.4 66.3 4 17.2 83.5 5 8.9 92.4 6 7.6 100.0 The next step is to calculate the arr i v a l rate. As stated previously, since the groups arrive independently of one another, the arri v a l rate may be approximated by a Poisson distribution. If the mean number of passengers arriving i s denoted by A^, the mean group a r r i v a l rate A i s A P AG - y G y« = mean group size (2.9) passengers as calculated from CJ Tables 3.1 and 3.2. The interarrival rate, or time between arrivals is approximated by a negative exponential function and i s calculated as follows: Time Interval Mean Group Arrival Rate X Negative Exponential Function — x ln (R.N.) AG In (R.N.) = natural log of a Random number. The percentage of Revenue and Standby groups (passengers) can 1 be calculated from the flight manifests. The fourth statistical definition is derived from the allocation process in the arrival stream. The flowchart of this process appears in Figure 3.2. The process deals with the total group arrivals in the system and subsequent determination of Revenue or Standby groups. As groups flow through the system, only non-family groups may fly Standby. Thus 70 per cent of groups arriving bypass this allocation process to Revenue or Standby systems. The arrival stream at the airport may be explicitly specified by the analyst to reflect changes in the composition of passenger arrivals or operation characteristics. The Check-in Process. The discussion of the quantitative defini-tions of the check-in process w i l l centre on: 1. The allocation process in the Revenue and Standby systems; 2. The facil it ies and their characteristics; 3. The queueing behaviour at the faci l i t ies . The allocation process within the Revenue and Standby systems is graphically displayed in Figure 3.3. A quantitative dimension wi l l be added to the description. The allocation process within the two systems acts upon either groups or passenger units. The percentage speci-fications are exogenous to the system and must be defined as follows. Family Bypass 70% of Total Group Arrivals GROUP ARRIVALS S GROUP A >v FAMILY 1^ - »T *•» r\ a. No = 30% of a l l Groups XH30 = % Standby RPS = Revenue Passenger System SPS = Standby Passenger System Figure 3.2 Arrival Stream Specification XH = Halfword Save Value to Specify Percentage (Yes answers) in Parts per Thousand. XH31, 150 or 15% of Revenue Groups Arrive Non-ticketed Figure 3.3 - Check-in Process Specifications * % of groups using curbside check-in (XH33) % of passenger units with 'excess baggage (XH34) STANDBY SYSTEM % of groups arriving non-ticketed (XH32) % of passenger units with excess baggage (XH35) * (XH31) is the name of the variable for allocation process. * * Note: groups that arrive non-ticketed are not allowed to use the curbside check-in services. These percentages may be the averages for a l l flights in a certain period of the year or by the month. The averages do fluctuate because more excess baggage is incurred in the winter months due to the weight of winter clothes. A l l specifications except curbside check-in are percentages of the total. If curbside users are defined as a per-centage of the total, allowance must be made for the fact that non-ticketed passenger groups are not allowed to use curbside check-in. The facil it ies are the baggage and ticket agents'check-in counters. The number of facil it ies that are made available for passenger service is given by the facil ity policy. The staffing horizon used by Air Canada is a 15-minute interval. This policy must be specified in the model, section 1.1, by the following method. XH1 = number of Revenue baggage counters open XH2 = 16 - the number of Revenue ticket counters open XH3 = 16 + the number of the Standby baggage counters open. only one is operational at this point in time. The service times of the facil it ies may be approximated by the 4 Erlang distribution (K = 3 or 4). At the baggage counter, the mean service time was dependent on the size of the passenger unit and whether or not the passenger unit had excess baggage. In the model, exponential service times were used to impute a conservative element into the service times. The mean service time was dependent upon the size of the group and i f excess baggage was encountered, the mean increased by 30 seconds. Table 3.3 reveals the mean service times by groups. At the ticket counter, the service times were dependent upon the type of service and not the group size. Data collection problems were encountered when trying to identify the passenger unit size. Therefore two probability functions were formulated - one for ticket purchases, the other for payment of excess baggage. The distributions are in Figure 3.4. TABLE 3.3 MEAN SERVICE TIMES FOR BAGGAGE COUNTERS Group Size Mean Service Time (seconds) 1 45 2 60 3 90 4 125 5 160 6 180 V. K. Wozniuk, ibid. The queueing discipline at the check-in was on a first-come, first-served basis. When passenger groups arrived at the counters, the shortest queue was selected as a group. If groups ' spl i t ' into passenger units, those passenger units would follow the same queueing behaviour. This behaviour was programmed into the model as follows. Let = current length of queue i in passenger units. Let BV. = a boolean variable such that BV. = 1 1 l i f the facil ity ' i ' is in use, and '0' i f i t is not. Let the number of counters open be n. Therefore as groups arrive, the queue size (L^) is L. = QJ + BVJ . l i i L^ MIN. is selected as the queue for the group starting from LI. (If LI = L2 = MIN then LI is selected; also, i f Q1 = Q2 = 0 and BV± = 1 and BV2 = 0, then L2 is selected.) The Output. The output unit of the Passenger Check-in system was stated as being a passenger unit which has completed check-in. Associated with this flow element is a group of statistics which des-cribe the state of the system. This set may be used to compare the effects of various policies used in the system. The first statistic is the average transit time per passenger unit. The transit time of a particular passenger unit is the summation of its waiting times in a l l queues and the service times encountered while checking in . No consideration has been given to walking time to and from the various counters because of the sampling problems in-volved and the lack of useful information that would be derived. The next statistic deals with the faci l i t ies . The time that an agent is engaged in service as a ratio to total time available is the utilization of that faci l i ty. If the average utilization of a set of facil i t ies is high and service standards maintained, the cost of service is low because the manpower of the system is being used. More wi l l be said about this in the analytical experiments. The statistics which indicate the service standard is the per-centage of passenger units who wait in lines longer than 'x' minutes. Air Canada has stated that 85% of the Revenue passenger units must not exceed 2.5 minutes waiting time in any one queue; 75% of the Standby passengers would be a desirable amount to meet this same standard. In summary, then, the state of the system may be defined as: 1. Transit time per passenger unit; 2. Utilization of faci l i t ies ; 3. % of passenger- units exceeding 2.5 minutes waiting time in queues. Part 3 - Verification To simulate, is to duplicate the essence of the system or activity without actually attaining reality i t se l f . r George W. Morgenthaler, "Theory and Application of Simulation in Operations Research," in Progress in Operations R e s e a r c h , Vol. I, Russell L. Ackoff, ed. (New York: John Wiley & Sons, Inc., 1961), pp. 367, cited in Meier, Robert C , Newell, William T. , Pazer, Harold, Simulation in Business and Economics, (Englewood Clif fs , New Jersey, Prentice-Hall Inc., 1969), p. 2. The verification that a computer based model is , in fact, "dupli-cating the essence" of the system is of extreme importance. The requirement that a complete validation of the simulation model take place, is that the results of experimentation wi l l be used as inputs for the decision-making process that controls the real system. Thus, verification must proceed the utilization of the model. The methodology for verification consists of four sequential steps. The result of each step wi l l be the foundation or assumptions of the next step. The four steps are described below and wi l l be pre-sented in turn. Step 1: Logic verification. The logic of-the model wi l l be verified to ensure that i t represents the system correctly. This is a fundamental step in the verification since further verification is based on the assumption that the logic has been modelled. Step 2: Several runs wi l l be made to determine the length of simulated time necessary for the model to become stable, and thus, rep-resentative. Step 3: Since the logic and the time necessary to stabilize the system are known, a comparison of observed data with simulated results wi l l be made. This is a crucial step since i t wi l l determine i f the system has in fact been simulated. Step 4: The primary objective of Step 4 is to determine the feasibility of using independently determined facility (15 minutes) policies for planning over a longer range. Before this can be achieved, optimal facil ity policies must be known for given arrival rates. Step 1. As previously stated, the logic of the system must coincide with that of the simulation model. The operating rules of the system were verified by the personnel who work in and control the system. Flow charts of the system similar to those appearing in Appendix II were shown to the agents. Thus, i t may be stated that the basic model represents the logic of Air Canada's present Check-in Systems at the Vancouver International Airport, Richmond, B. C. The next step in the logic verification is testing the model to ensure that the parameter specification properly allocates passengers to the subsystems. The parameters used for verification steps 1, 2 and 4 are as follows: Revenue Stream: 85% of total Passengers are Revenue 15% of Revenue groups arrived at the Airport without a ticket 10% of Revenue groups used curbside check-in. Standby System: 15% of total passengers 15% of Standby groups arrived without a ticket 15% of Standby passenger units have excess baggage. The arrival and group composition were as previously defined for Vancouver. The model reacted as prescribed and allocated the passenger groups or units to the various sub-systems. An indication of the accuracy of this allocation process is presented in Table 3.4. It should be noted that close approximations are acceptable due to the random assignment of passengers and the length of time simulated. TABLE 3.4 PARAMETER ALLOCATION Parameter % Revenue Passengers % Revenue groups: - non-ticketed - using curbside r % Revenue passenger units - having excess baggage % Standby groups - non-ticketed Standby passenger units - excess baggage Allocation Specified Simulated 85.0% 15.0% 10.0% 15.0% 15.0% 15.0% 86.3% 14.2% 10.3% 15.1% 16.0% 14.0% Simulated time was 120 minutes with an arrival rate of 100 passen-gers per 15-minute period. The percentage "specified" are the parameter values that should be approximated by the column marked "simulated" - the resultant statistics obtained from the output. The essence of this step is to confirm that the logic and allo-cation of passengers in the system is taking place within reasonable limits. Step 2. The objective of Step 2 is to determine the length of simulated time necessary for the system to stabilize or reach a "steady state." The steady state of a simulated system is reached when fluctuations in system statistics from one time period to the next become nominal. Only when the system is at steady state are the statistics representative. Stochastic variation of a simulated system is the variation in a statistic at one particular point in time resulting from natural variations in system flows under normal operating conditions. Thus, when the perturbations associated with bringing a system from an "unloaded" state to normal operating levels (people in queues, etc.) are reduced and the statistics converge to an acceptable range about a single value, the system that is being simulated has stabilized. Representative statistics of the system's state may now be obtained. The length of simulated time to steady state wi l l be utilized in Steps 3 and 4 when representative statistics of the system must be known for verification. The method used to determine the steady state occurrence of a A^/P^ (an arrival rate and the associated facility policy P )^ was as follows: 1. For a given date (A.) and an associated facility policy (P.), the system was simulated for four hours. 2. At intervals of 60 minutes, the statistics were collected. 3. Steps 1 and 2 were repeated twice using new random number seeds. 4. A different A . /P . was selected and Steps 1, 2 and 3 were repeated. The statements in the GPSS program used in verification appear in Appendix IV. The results were plotted and appear in Figures 3.5 and 3.6. The average transit time in the system per Revenue passenger unit and the Revenue baggage counter utilization were selected to provide a succinct view of the system as simulated time progresses. These statistics reflect the parameter specification and system performance. The stochastic variation tends to smooth over time, as expected. In summation then: 1. The steady state of the computer model is reached in the 180 to 240 minute range of simulated time. In this range, the statistics are representative of the state of the system. Thus for future va l i -dation and experimentation, the statistics should be collected after 180 minutes of simulated time. 2. If Figures 3.5 and 3.6 are compared, i t is noted that when the arrival rate is high (X = 100) the statistics tend to smooth more quickly than at lower arrival rates (X = 50). There is probably an inverse relationship (non-linear) between the volume of arrivals and the length of time to steady state. Step 3. Since the optimal time to simulate the system is known, the next step in the verification is to determine i f the model does in fact simulate the real system. This step is of great importance because further analysis of the system is based on the assumption that the model does 'simulate' the system. MIN. 2.8 2.6 2.4 2.0. REVENUE PASSENGER TRANSIT TIME % 60 55 50 UTILIZATION 45 40 60 AVERAGE REVENUE BAGGAGE COUNTER UTILIZATION 120 1W 240" Simulate Minutes Figure 3.5 - Search for Steady State A = 100 Passenger/15 Minutes 3.0 4 2.8 2.6 . 2.4 2.0 -REVENUE PASSENGER TRANSIT TIME UTILIZATION 40 35 30 25 20 60 AVERAGE REVENUE BAGGAGE COUNTER UTILIZATION 120 Figure 3.6 - Search for Steady State 180 2%0 Simulated Time X = 50 Passengers/15 Minute Period It should be noted that the human systems are less predictable than mechanical systems. This is due to the fact that humans control their movements in the system. A mechanical sys em is very ordered because the machine speeds and movements of parts have a very small variance in their descriptive mean times. When simulated, the results more closely approximate or predict the state of the real ystem. The methodology of this step is as follows. 1. Observe the system and obtain sample statistics which reflect the state of the system. 2. The \^ and that was observed in the real system were used as parameters in a simu-lation. Simulated system statistics were then calculated. Various random number seeds were used to obtain averages. 3. A comparison of observed results with that of the simulated statistics was made. It should be noted that observed statistics have been used in the simulation model. First there are the service times. These were collected by observations and the mean times varied directly (non-linear) with the number of people per passenger unit. The type and group com-positions were obtained from Air Canada and are representative of the average arrival stream. Thus any or part of the variation in the simu-lated results from that of the observed might stem from the fact that the type and group composition of the sampled time interval was not the same as the data obtained from Air Canada (Tables 3.1 and 3.2). Another source of variation could be the use of a mean arrival rate. In the real system, the arrival rate changes continuously. The model, however, uses a constant average arrival rate for a given 15-minute The validation run was constructed as follows: 1. The arrival rate at the Revenue and Standby baggage counters was used. This was done because 1) curbside check-in does not affect the counter operations; and because 2) a l l non-ticketed groups must proceed to the baggage counters via the ticket counters. Thus the total arrivals would be the sum of the arrivals at the baggage counters plus the number of passengers using curbside check-in, or in this case, X = 58 passengers per 15 minutes 2. The facil ity policy that was in operation was: Revenue baggage 4 ticket 3 Standby baggage 1 ticket 1 The comparison of the observed arrival stream to the simulated appears in Table 3.5. The variation comes from the fact that the model produced 2.8 people per Revenue passenger unit while the observed was 2.6 people. However there were the same number of Revenue passenger units in the system. Thus the simulated system handled eight per cent more people. Therefore, the system statistics should be slightly higher than the observed. Statistics Observed Simulated Total number at baggage counters 58 64 (+8%) Number of Revenue passengers 46 51 Standby passengers 12 13 Revenue as a per cent of total 79% 80% Number of Revenue passenger units 18 18 Standby passenger units 12 13 Passengers per passenger unit (Revenue) 2.6 2.8 (+8%) The data collection in the stimulation for the system statistics occurs as follows. At intervals of 60 seconds, the number of passenger units in the queues was counted. Thus, the average as well as the maximum number of passenger units in queues at the Revenue and Standby counters may be calculated. At the same time, i f the facil ity was in use, i t was counted as '1 ' , i f i t was not in use, then a '0 ' . Therefore the approxi-mate average utilization for a set of facil it ies would be the sum of the ' l ' s divided by the number of observations. It is felt that these three statistics should determine whether or not the model does in fact simulate the real system. The comparison is given in Table 3.6. TABLE 3.6 COMPARISON OF SYSTEM STATISTICS Statistics Observed Simulated Maximum number of passen-ger units in queues Revenue baggage ticket 4 2 4 3 Standby baggage ticket 5 3 3 5 2. Average number of passen-ger units in queues Revenue baggage ticket Standby baggage ticket 2.1 1.8 4.4 2.0 2.6 2.3 3.7 2.5 3. Facility utilization Revenue baggage ticket Standby baggage ticket 47% 40% 73% 33% 48% 39% 65% 34% Variation could be caused from method of data collection. Thus, i t may be concluded that the simulation model as constructed, does in fact, represent the system at the Vancouver Airport. Step 4. The fourth step in verification of the model is to confirm that independently determined optimal facil ity policies wi l l be compatible from one period to the next. When the arrival rate A of period t is large and A + ^ of period t + 1 is small, the facility policy determined by simulation for A + ^ may no longer be optimal. The magnitude of A may be such that passengers are left in the system and thus wi l l affect the performance of the system in period t + 1. Thus, the methodology for this step is as follows: 1. Determine an optimal ^ / P ^ r o r ^ = 100, 50, 25. 2. Set A./P. at the optimal level and simulate until a steady state is reached. This signals the end of the first time period (t). 3. The passengers in the system (queues and f ac i l i -ties) w i l l remain, but a l l other system statis-tics wi l l be set to zero at the start of t + 1. 4. A./P. wi l l be changed to A_/P„ and simulated for 15 minutes. The statistics wi l l be collected at this point in time. The results were as follows: In the determination of a faci l i ty policy for each arrival rate, the passengers must be processed within acceptable norms of system per-formance; namely, 85% of the Revenue passenger units must not wait in queues longer than 2.5 minutes; 75% of Standby passengers must meet the same conditions. Thus, by a search technique of setting a A_^  and varying the P^, the following policies were obtained: X1 = 100 = 5, 3/2, 1 A2 = 50 P 2 = 4, 2/1, 1 A 3 = 25 P 3 = 2, 1/1, 1 The notation of = X^, X 2 / X 3 > X^ is defined as: X^ = no. of Revenue baggage counters open; X 2 = no. of Revenue ticket counters open; X^ = no. of Standby baggage counters open; X^ = no. of Standby ticket counters open. There is an approximate P^ for each A^ that achieves the service standard or norms of system performance. Steady State statistics for A^/P^ (optimal) appear in Table 3.7 and wi l l be used in the remainder of Step 4. TABLE 3.7 AVERAGES OF SYSTEM STATISTICS Statistic X. = 100 50 25 Transit Time: Per Revenue Passenger Unit 2.6 min. 2.4 Passenger 2.7 2.5 2.8 3.1 Standby Passenger Unit (in minutes) 2.9 3.1 2.1 2. Facility Utilization (%) Revenue baggage counters ticket counters 57.5% 48.2% 30.7 30.0 30.0 20.9 Standby baggage counters ticket counters 31.2 49.9 32.4 18.6 16.6 10.2 Average time in queues per passenger unit (nearest 10 sec.) Revenue baggage counters ticket counters 20 sec. 30 20 sec. 20 30 sec, 10 Standby baggage counters ticket counters 30 170 80 6 50 20 4. % of Passenger Units In Queues Longer than 2.5 min. at: Revenue baggage counters ticket counters 5.2% 6.4 3.2% 4.1 6.5% 0 Standby baggage counters ticket counters 6 - ° * 30.0 23.0 16.7 13.5 9.0 The change-over from """n fc t o ^2^2 ^n ^ + ^ W a S a S Run 1 X l = 100 changed to X2 = 50 P l = 5, 3/2, 1 P2 = 4, 2/1, 1 Run 2 A l = 100 changed to X2 = 25 P2 = 5, 3/2, 1 • P2 = 2, 1/1, 1 Run 3 X l = 50 changed to A2 = 25 P2 = 4, 2/1, 1 P2 = 2, 1/1, 1 The statistical results are presented in Table 3.8. As could be expected, the average transit time for the second period increased compared to the transit times when the policies were evaluated individually. This is due to the number of people remaining in the system from the previous period. The percentage of passenger units waiting in the queue at the Standby ticket counters was greater than the acceptable mean standard of 25% (A = 100 -> 25, 29.6% and A = 100 -> 50, 28.2%). However, the stated service policies derived in the earlier part of Step 4 have remained optimal, since 1) Air Canada does not have the capacity to increase Standby Ticket purchase counters to two positions; and 2) Standby passen-gers are of lesser importance than Revenue passengers. The degree of change in optimality of the t + 1 period policies depends also on the amount of slack existing at the previously optimal steady state policy. RUN I: A = 100 50 Statistic Averages Steady State Change-over 1. Transit time (minutes) per: A = 50 A = 100 -* 50 Revenue passenger unit 2.4 2.6 Revenue passenger 2.5 2.7 Standby passenger unit 3.1 3.2 2. % of Passenger units in Queues longer than 2.5 minutes at: Revenue Baggage counter 3.2% 3.5% Ticket counter 4.1% 14.0% Standby Baggage counter 23.0% 4.8% Ticket counter 16.7% 28.2% RUN 2: A = 100 -> 25 RUN 3: A = 50 -»• 25 A = 25 A = 100 -> 25 A = 50 -*• 25 1. Transit time (minutes) per: Revenue passenger unit 2.8 3.4 3.2 Revenue passenger 3.1 3.5 3.3 Standby passenger unit 2.4 3.0 3.3 2. % of Passenger units in Queues longer than 2.5 minutes at: Revenue Baggage counter 6.5 6.6 5.0 Ticket counter 2.3 15.0 4.5 Standby Baggage counter 13.5 18.5 22.5 Ticket counter 9.0 29.6 23.3 In summation to the verification of the computer model: 1. The representation of the system has been captured and computerized; 2. The results of further experimentation w i l l be plausible since the basic system has been simulated correctly; 3. Steady state conditions occur within approxi-mately 180 to 240 minutes of simulated time; 4. Optimal facil ity policies determined indepen-dently are compatible with one another. C. Analysis of the System In summary, the system approach has been applied to the Passenger Check-in system: 1. The system was identified as to its inputs, conversion process and outputs in Chapter II; 2. The system was defined quantitatively modelled and verified thus far in Chapter III. Thus what remains is the analysis of the system. There are two objectives of this analysis: 1. To provide a tool so that experimentation may take place; 2. To be able to formulate and evaluate policies. Since the model has been developed for experimental policy evalu-ation, the first objective is deemed completed. The capabilities of the model are vast. For instance, due to the method of construction, very slight changes in the program would allow the model to simulate a 24 hour period with fluctuating (continuous) arrival rates and facil ity policies. More specifically, the model was designed for the evaluation of policies (service, facility and operating) prior to implementation. There is no need to implement new policies in the real system in order to determine their viability; the model wi l l provide a faster and less costly evaluation of these new policies. The remainder of the systems analysis wi l l consist of three steps. They are briefly described below and wi l l then be presented in their entirety. 1. To determine the exact behaviour of the system and its impli-cations on policy formulation. This step wi l l , in effect, allow policies to be formulated in a more rational manner. 2. To determine the facil ity policy that should be implemented to achieve the required service standard at various arrival rates. The maximum capacity of the facil it ies wi l l also be determined. 3. To formulate alternative operating policies and to determine i f they are viable in the system. This wi l l be a multi-step process. As in the verification, there are certain exogenous variables that must be specified. In the experimental analysis, a l l steps wi l l have the same set of variables. They are: Revenue System: 85% of total passengers 15% of Revenue groups arrive without tickets 10% of Revenue passenger groups use curbside check-in 15% of Revenue passenger units have excess baggage. Standby System: 15% of Standby passenger units have excess baggage. Step 1. This part of the analysis wi l l explore the behaviour of the system to determine i f the service policy has been properly formulated. The service policy is as follows: To allow no more than 15% of Revenue passenger units to wait in a queue more than 2.5 minutes; to allow an average of 25% of the Standby passenger units to wait more than 2.5 minutes. The main elements which determine the state of the system are the arrival rate, X^, and the facility policy, P^. The behaviour or state of the system is reflected by three statistics: the average transit time per passenger unit, the average utilizatization of the various facil it ies and the percentage of passenger units waiting in a queue longer than 2.5 minutes. If the facil ity policy (P^) is held constant, denoted as P^, then the effect of the arrival rate on the subsequent behaviour of the system may be determined. The methodology of this experiment is as follows. 1. Set P^ = 4,2/1,1 and simulate with \^ = 40 until a steady state is reached. 2. Collect the system statistics and repeat step 1 using a new random number seed. 3. Repeat steps 1 and 2 until X^ = 160 using increments of 20. The statistical results have been tabulated and appear in Table 3.9. The selection of the Revenue Baggage Ticket System statistics for presentation wi l l be sufficient to reveal the justification of the ser-vice policy. The relationship between the arrival rate and the average transit time per passenger unit when the facil ity policy is held constant, may be seen in Figure 3.7. The interpretation of this curve is as follows: When is in operation and X_^  is in the range of 40 to 90 passengers per 15 minutes, the average transit time is between 2 and 3 minutes. The system is thus relatively stable or insensitive to the number of passengers in the system. When is greater than 90 passengers the curve becomes hori-zontal. This implies that the system becomes unstable and is sensitive to the number of passengers in the system. For example when X^ = 100 goes to X^ = 120 the transit time goes from 3.3 to 6.5 minutes per passen ger unit. This is a 20 per cent increase in volume and results in 100 per cent increase in transit time. Thus, i f a curve was derived for each P. at the various arrival I rates, the vertical portion of the curve would indicate the range of X^ such that P^ produces stability in the system. Figure 3.8 is a graph of the average utilization of the Revenue baggage counters as X^ varies. Note that the curve is almost linear in the range of X = 40 to X = 120. When X^ increases past 120, the curve becomes vertical in nature. At a very high arrival rate, utilization would approach 100%. Also from the previous figure, there would be an extremely long transit time. Thus, the average utilization for the Revenue baggage counters wi l l be in the range of 30 to 65% when 40 _< X _< and P = 4, 2/1,1. SYSTEM STATISTICS FROM ANALYSIS: STEP 1 REVENUE PASSENGER SYSTEM Arrival Rate A. = 1 Revenue Passenger Baggage System Transit Time Per Passenger Unit (Min.) Revenue Passenger Ticket System Counters: Average Utilization % of Passenger . Units exceeding 2.5 minutes waiting time Counters Average Utilization % Passenger Units exceed-ing 2.5 min. waiting time 40 2.3 25.% 2.3% 26.4% 4.2% 60 2.5 41.6 3.0 42.1 9.2 80 2.6 54.0 6.6 50.0 5.5 100 3.3 66.4 13.4 62.5 16.1 120 6.5 87.3 49.2 68.5 20.8 140 7.6 93.1 59.5 91.0 52.9 160 12.4 94.3 73.0 92.6 61.1 NO 0 10 20 3(!> 40 50 60 70 §5 90 100 Average Utilization in % Revenue Baggage Counter Figure 3.8 - Arrival Rate and Corresponding Utilization If the average service time is 1 minute and average time in the system is 3 minutes, then the average waiting time in queues would be 2 minutes. Figure 3.9 is a plot of the annual rates versus the percentage of Revenue passenger units waiting longer than 2.5 minutes at the baggage counter. This curve has a vertical and horizontal section. When X. l is in the range of 40 to 100 passengers, the Revenue baggage counter is able to process the passengers at such a rate that the 15 per cent rule is not violated. In the range above X = 100, the curve becomes horizontal in nature. The change-over occurs when approximately 13% to 15% of the passenger units exceed 2.5 minutes of waiting time. This curve then justifies the current service policy. If the percentage is increased, the system becomes very sensitive to the amount of passengers in the system for (at X = 120 and P^, 49% of Revenue passenger units exceed 2.5 minutes). The vertical portion of the curve represents the range within which P^ produces stability in the system. In Figure 3.10 the utilization and corresponding percentage of passenger units exceeding 2.5 minutes waiting time has been graphed (Figure 3.8 + Figure 3.9). For policy formulation, the effect of increasing the percentage of passenger units allowed to exceed 2.5 minutes waiting time on the facil ity utilization may be determined from the graph. Alternatively, th gives the trade-off of opposing system objectives. If the percentage was increased to 25%, the average utilization that could be expected would Arrival Rate X = 160 • 140 • 120 • 1 i ^ — 100 . 80 . 1 P^ held constant 60 . | Present Policy 15% 40 . 1 0 10 20 30 40 50 60 70 80 % Figure 3.9 • Passenger - Arrival Rate and % Units Exceeding 2.5 Minutes % passenger units exceeding 2.5 waiting minutes time Utilization % 100 . P. Held Constant and Arrival Rate 1 Varies 90 . | ^ ' RPBS 80 . 70 • 60 | 50 1 40 -30 . 1 Service Policy 15% of the Revenue Passenger | Units are Allowed to Exceed 2.5 Minutes 20 10 . 1 Revenue Passenger Baggage , System Counters = 4 0 5 10 15 20 30 40 50 60 70 80 % Passenger Units _^ 2.5 Minutes Waiting Time in Queues Figure 3.10 - Average Utilization and % Passenger Units > 2.5 Min. Waiting Thus, we have developed in this step a methodology for deter-mining the service policy. Once a change has occurred in the system, this step should be repeated to determine i f the service policy should be altered. Step 2. The second step is to determine the facility policy that should be implemented into the system to achieve the stated service policy. As part of this step, the maximum capacity of the system wi l l be deter-mined. Although very high arrival rates may not be encountered at the present time, they could be forecast for a few years hence. Thus, the result of this step could be an input for planning Air Canada's operations in the future. The methodology of this step was as follows. 1. Set X.. l 2. Vary the P^ such that the service policy is achieved. Simulate for 180 minutes. 3. Increase X_^  by 25 passengers and search for a new P.. l 4. Repeat steps 2 and 3 until the system can no longer achieve the service policy. It should be noted that when X. is increased, the number of l facil it ies open in the new P^ wi l l be equal to or greater than the previous policy. This in effect limits the search to only increases in P i . The results of the analysis appear in Table 3.10 and Table 3.11. FACILITY POLICY FOR X. l Revenue Standby Baggage (X ] [) Ticket ( X 2 ) Baggage (X^) Ticket ( X 4 ) EX. l 25 2 1 1 1 5 50 4 2 1 1 8 75 4 2 2 1 r i 9 100 5 3 2 1 11 125 6 4 2 1 13 150 7 5 2 1 15 175 8 5 2 i i 1 16 200 9 5 i i 2 1 17 225 , Service policy unattainable due to limited number of facil it ies Passenger Composition 85% Revenue Passengers 10% Use Curbside 15% Excess Baggage 15% Non-ticketed TABLE 3.11 FACILITY POLICIES AND SYSTEM STATISTICS Revenue Standby X. = X T. T. X l Baggage U% %>2.5 X2 Ticket U% %>2.5 T. T. X3 Baggage U% %>2.5 X4 Ticket U% %>2.5 25 2.8 2 30.0 6.5 1 20.9 0 2.1 1 16.6 13.5 1 10.2 9.0 50 2.4 4 30.7 3.2 2 30.0 4.1 3.1 1 32.4 23.0 1 18.6 16.7 75 2.9 4 50.3 9.8 2 43.8 4.2 3.3 2 25.6 4.0 1 42.1 15.0 100 2.6 5 57.5 5.2 3 48.2 6.4 2.9 2 31.2 6.0 1 49.9 30.0 125 2.8 6 62.4 4.7 4 53.3 11.9 3.2 2 51.0 17.3 1 55.7 32.4 150 2.7 7 64.5 6.5 5 48.9 3.5 3.3 2 58.2 22.3 1 56.1 32.8 175 2.9 9 57.4 4.3 5 58.0 4.4 3.6 2 62.5 20.8 1 73.1 36.8 200 3.0 9 63.6 6.6 5 63.7 10.5 5.5 2 74.4 37.2 1 76.5 49.0 T. T. = Transit Time X l = No. of Revenue Baggage Counters Open U% = Utilization of X. X X2 = No. of Revenue Ticket Counters Open %>2.5 = % of Passenger Units exceeding X3 = No. of Standby Baggage Counters Open 2.5 Minutes Waiting Time in Queues X4 = No. of Standby Ticket Counters Open oo o Table 3.10 gives the facil ity policy = x^/x^,x^ that should be implemented into the system i f the service standard is to be maintained. It is also representative of the capacity of the system. In each column there is a vertical line; this indicates the capacity of the system which results from the limited number of counters available. At X = 100, our first partial capacity is reached at the Standby ticket counter. When X = 175, the Standby Baggage counter just remains optimal but care must be used since passengers experience some delay at the ticket counter. The capacity of the Revenue counters is met at X = 200 per 15-minute period. The total agent requirements are also listed, and is computed as the E X . in the P. notation (P. = X n , X„ /X 0 , X . ) . l l l 1 2 3 4 The application of this step is very useful indeed. The result of Step 4 allows the use of independently determined facil ity policies to be used (15-minute horizon) for daily planning. Suppose the X_^  for the average 15-minute period t_^  was as follows: t l = xioo fc2 = X50 fc3 = X75 Then the P. for each t, would be as follows: l 1 t ^ P 1 = 5, 3/2, 1 Z X ± = 11 t 2 = P 2 = 4, 2/1,1 ZX± = 8 t 3 = P 3 = 4, 2/2, 1 ZX± = 9 The 15-minute faci l i ty policy that wi l l achieve the stated service policy is available from the model. The question then arises as to the hourly facil ity policy which is complicated by unknown contractual agreements. Although this is outside the scope of the thesis, the results of this step, as contained in Table 3.11 could be used as inputs for schedule determination. Step 3. The analysis that wi l l be undertaken in Step 3 is directed towards the operating policy of Air Canada. A few years ago, the operating policy was changed such that the baggage operations and ticket purchases of the Revenue passengers was separated from that of the Standby passengers. Revenue passengers paying excess charges were allowed to use the Standby ticket counter and Standby passengers under the same circumstances were allowed use of the Revenue ticket counter. In Part A of Step 3 the impact of complete separation of Revenue and Standby ticket operations is described. The implementation of this policy should produce a higher utilization of facilities and greater percentage of passenger units waiting longer than 2.5 minutes in queues. This is so because the Revenue passengers are limited to one fewer counters for service. The system, which was simulated for multiple 180-minute runs, produced higher system statistics than when the integrated policy was operational. (See Table 3.12). If the decision is made that complete separation is necessary, the facil ity polici es as stated in Table 3.10 should produce stability in the system for the Revenue system. REVENUE/STANDBY TICKET OPERATIONS IN FULL SEPARATION Revenue Transit Time per Passenger Unit Average Utilization % of Passenger Units Waiting Longer than 2.5 Minutes Integrated Separated 2.6 Min. 2.85 Min. 28.2% 49.9% 6.4% 13.8% Standby Transit Time Average Utilization % of Passenger Units Waiting Longer than 2.5 Minutes 2.9 Min. 49.2% 3^0% 3.75 Min. 62.1% k54.9% 25% + 5% would be acceptable but since there is only one faci l i ty, a decision must be made to either expand or revise the service policy at the Standby Ticket counter. In Part B of Step 3 we examine the operating policy at the Revenue Ticket counter more closely. Since the separation as in Part A, produced higher system statistics, the queueing discipline wi l l be changed to determine i f the statistics can be lowered while maintaining separated operations. Specifically, instead of using multiple queues or one queue for each agent, a single queue wi l l be used. A l l Revenue passengers purchasing tickets or paying excess charges line up in one queue. When they reached the front of the line, they wait until any one of the agents becomes available. Only then do they leave the line. This is depicted in Figure 3.11. The results of the simulations combined with the results of Part A for comparative purposes are in Table 3.13. The transit time per passenger unit is lower and is comparable to the integrated operation. The utilization is lower as well as the % of Revenue passenger units exceeding 2.5 minutes of waiting time than the integrated. Thus i f this particular method of queueing was adopted, separation of the Revenue/Standby counter could be achieved with viable results. In Part C of this step in the analysis we examine the method of checking in Standby passengers. Since there has been a large number of passengers exceeding the 2 1/2 minute of waiting time, the combination of ticket and baggage operations at one counter wi l l be analyzed. In effect, one agent would become both a ticket and baggage agent. When the passenger unit approaches the service counter, a single agent wi l l issue tickets, check the baggage and collect the excess charges. Since this operation QUEUE 1 QUEUE 2 QUEUE 3 V V AGENT 1 AGENT 2 AGENT 3 REVENUE TICKET COUNTER MULTIPLE QUEUES GOES TO ANY AGENT AGENT 1 AGENT 2 AGENT 3 MOVEABLE GUIDE RAILS REVENUE TICKET COUNTER SINGLE QUEUE Figure 3.11 - Multiple or Single Queues COMPARATIVE STATISTICS FOR REVENUE TICKET COUNTER UNDER VARIOUS OPERATING POLICIES Revenue Ticket Standby Excess Allowed Revenue Standby Separated One Line at Revenue Ticket Counter Transit Time per Passenger Unit Utilization % P. U. > 2.5' 2.6 Min. 48.2% 6.4% 2.85 Min. 49.9% 13.8% 2.75 Min. 42.9% 4.5% * Percentage of Revenue Passenger Units exceeding 2.5 Minutes Waiting Time multiple operations and is not at present in effect the correct dis-tribution of service times must be hypothesized. It would probably be approximated by an Erlang distribution with K = 4 or 5. In the simu-lation, the service times wi l l be drawn from the already existing dis-tributions for service and baggage check. The total service time vary from the present baggage service time only i f there are tickets to be purchased or excess charges to be paid. The estimated total service time w i l l probably be high which in this case is conservative since we are testing a completely new agent procedure. The flowchart of the operation for a single agent is depicted in Figure 3.12. The system statistics are given in Table 3.14. Thus i f this operational procedure is adopted, the transit time (average) of Standby passengers remains relatively the same as when the Revenue/Standby ticket operations were integrated. However, the percentage of passengers who exceed 2.5 minutes waiting time is reduced to 28.3%. Thus the service policy is maintained through the total system at X = 100 and the aforementioned parameter specifications. These are just a few of the experiments that can be performed on the model. The experiments presented hopefully have hit on the major areas of interest and demonstrated the vast power of simulation as an airport management tool. PURCHASE TICKET <— , CHECK-IN BAGGAGE - Denotes Service Time Figure 3.12 - 1 Standby Agent Performs Both Ticket and Baggage Operation TABLE 3.14 STANDBY OPERATIONS COMPARED Statistic Integrated 2 Baggage Agents 1 Ticket Agent Revenue/Standby Separated 2 Baggage Each Performing Baggage 1 Ticket and Ticket Operations Transit Time Baggage Average Utilization % Passengers Exceeding 2.5 Min. Waiting Time Ticket Average Utilization % Passengers Exceeding 2.5 Min. Waiting Time 2.9 Min. 31.2% 6.0% 49.9% 30.0% 3.75 Min. 30.3% 5.8% 62.1% 54.9% 3.0 Min. Combined Utilization 44.7% % < 2.5 Min.28.3% A = 100 passengers over 15-minute period CHAPTER IV CONCLUSIONS In presenting the conclusions of the thesis, Chapter IV wi l l be concerned with three areas: 1. The conclusions derived from the application of the 'systems approach' concept. As part of this area, the strengths of using simulation as an analytical technique wi l l be summarized. 2. The conclusions of the system analysis of Air Canada's Passenger Check-in System, Vancouver Airport, Richmond, British Columbia. 3. The proposal of future areas of research in the Airport system. THE SYSTEMS APPROACH CONCEPT The 'systems approach' concept has been applied to the Passenger Check-in System for the purpose of evaluating and formulating policies to ensure the attainment of the system's objectives. The relative strength of using this methodology is that i t forces the analyst to be pragmatic. The system must be defined explicitly, with logical and quantitative relationships carefully defined. Once this has been done, the sub-systems may be isolated and analyzed. The systems approach is readily applicable whenever a phenomena may be described as having an input element, a conversion or transformation process, and an output element. The l e v e l of d e t a i l of the systems an a l y s i s i s d i c t a t e d by the requirements of the ultimate user of the r e s u l t s and the s i z e of the system. The methodology contained i n the thesis i s a p p l i c a b l e to most Check-in systems of the major a i r l i n e s . The strengths of using simulation as an a n a l y t i c a l technique are summarized below. 1. The technique i s r e a d i l y a p p l i c a b l e to a 'flow system' permitting observation of the dynamic behaviour. 2. When there are no p r a c t i c a l a n a l y t i c a l approaches (queueing theory) a v a i l a b l e f o r complex systems, simulation can be u t i l i z e d to sim-p l i f y the a n a l y s i s . 3. A simulation model permits the evaluation of a p o l i c y e i t h e r before or a f t e r implementation since i t i s b a s i c a l l y a laboratory t o o l that can be used under c o n t r o l l e d conditions. 4. I t f a c i l i t a t e s the management of a system by providing a means of studying cause-effect r e l a t i o n s h i p s inherent i n the system. Thus the objec t i v e of t h i s study i s to provide management with a comprehensive t o o l that w i l l make po s s i b l e the t e s t i n g of a l t e r n a t i v e management p o l i c i e s , f u l f i l l e d by providing a simulation model of the Check-in system. CONCLUSIONS OF THE SYSTEMS ANALYSIS The systems a n a l y s i s , as.presented i n the l a t t e r part of Chapter I I I was di r e c t e d towards the problematic areas of system management. 1. The evaluation of a policy after i t has been implemented. 2. The determination of a facil ity policy such that the service policy is achieved. 3. The formulation and evaluation of an operating policy prior to implementation. These three analytical areas have provided a demonstration of the value of a simulation model and the merits of a systems analysis. The conclusions of each area are presented below. 1. The service policy that a maximum of 15% of the Revenue passenger units be allowed to exceed 2.5 minutes of waiting time has been formulated correctly. 2. The nature of the system is such that greater utilization of facil it ies wi l l not be achieved by a nominal increase in the allowable percentage of passenger units exceeding 2.5 minutes. 3. The facil ity policy and associated procedures have been formulated so that the objectives of the system wi l l be attained. 4. The maximum capacity of the system is exceeded when the arrival rate is in excess of 200 passengers per 15 minute period. 5. The use of a single queue at the Revenue ticket counter wi l l ensure greater attainment of the system objectives than the use of multiple queues. 6. The combination of the baggage and ticket operations at one counter is a viable alternative in the present system. AREAS OF FUTURE RESEARCH As a by-product of the Passenger Check-in System's analysis, there are four areas of possible research. They are: 1. A study carried out to determine the capacity of the baggage flow system. The ar r i v a l of passengers is the i n i t i a l i z a t i o n point of this system. Simulation could be an effective tool for this study. 2. A study of the arr i v a l rates as a function of the reser-vations at various times prior to f l i g h t departure. The information contained in Reservac II (Computerized Reservations System) would be most useful i n this study. 3. A similar study to the one completed in this thesis could be made of the Departure Lounge. Various operational policies could be tested to determine the most suitable departure gate policies for various t r a f f i c volumes and airc r a f t . 4. If the on-line reservation system is installed at the baggage counters, the service times w i l l probably change. With the use of the simulation model contained in this thesis, f a c i l i t y policies could be determined prior to this procedural change. Gensema, W. "Passenger Handling Models," AGIFORS Proceedings. New York: Printed at American Airlines, 1967. Hare, Van Court. Systems Analysis. New York: Harecourt Brace and World, 1967. IBM Application Program. "General Purpose Simulation System/360, Intro-ducing User's Manual," GH20-0304-4, IBM Corporations Technical Publications Department, 19 70. IBM Application Program. "General Purpose Simulation System/360, User's Manual," GH20-0326-3, IBM Corporation Technical Publications Department, 1970. Khanna, K. C. , Takamori, H. "Optimal Staffing at Airline Terminals," AGIFORS Proceedings. New York: Printed at American Airlines, 1969. Meir, Robert C , Newell, W. T. , Pazer, H. L. Simulation in Business and Economics. Englewood Cliffs , New Jersey: Prentice Hall Inc., 1969. Mountjoy, R. "Airport Simulation Models," AGIFORS Proceedings. New York: Printed at American Airlines, 1969. O'Broin, S. "Manpower Planning for Airport Operations," AGIFORS Proceedings. New York: Printed at American Airlines, 1968. Wiley, A. T. "Vancouver Passenger Market Survey." A private study, Air Canada, Montreal, 1967. Wozniuk, V. K. "Airport Manpower Planning Systems," AGIFORS Proceedings. New York: Printed at American Airlines, 1969. APPENDIX I AIR CANADA FACILITIES AIR CANADA FACILITIES VANCOUVER INTERNATIONAL AIRPORT 5 4 3 2 1 1 2 3 4 5 6 7 8 9 1 2 3 Standby Revenue Ticket Revenue Baggage Counters Standby Baggage ticket Offices To Departure Gates Airport Shops Enter System FLOWPATTERN OF PASSENGERS AT THE AIR CANADA FACILITIES Revenue Standby Ticket and Ex-cess payment Revenue Baggage Check-in Standby Baggage Check-in APPENDIX II FLOWCHARTS OF THE PASSENGER CHECK-IN SYSTEM FLOWCHARTS OF THE PASSENGER CHECK-IN SYSTEM Start or end A process - queue - service - set up characteristics 'Go to' or 'come from' depending on arrow. Section name and number given. Collect statistics SECTION 3.1 START OF SYSTEM GENERATE GROUPS i ASSIGN PASSENGER CHARACTERIS-TICS 0 SECTION 3.11 & 3.12 DETERMINATION OF GROUP SIZE & TYPE DEPENDENT FAMILY \ T Y P E >r > INDEPENDENT GROUPS SINGLE PASSENGER SECTION 3.2 REVENUE PASSENGER SYSTEM SECTION 3.21 ALLOCATION TO REVENUE SUBSYSTEMS SECTION 3.22 REVENUE PASSENGER BAGGAGE SYSTEM GROUP SELECTION OF QUEUE PASSENGER UNIT SELECTION OF QUEUE QUEUE AGENT YES NO CHECK-IN BAGGAGE ADJUST SERVICE TIME CHECK-IN BAGGAGE SECTION 3.23 REVENUE PASSENGER TICKET SYSTEM * SELECT QUEUE AS A GROUP FOR TICKET PURCHASES 1 SELECT QUEUE AS PASSENGER UNIT FOR EX-CESS PAYMENT QUEUE AGENT SECTION 3.31 STANDBY PASSENGER SYSTEM ALLOCATION TO STANDBY SUBSYSTEMS YES SECTION 3.32 STANDBY PASSENGER BAGGAGE SYSTEM © SPLIT OFF GROUP SELECT SHORTEST QUEUE QUEUE FOR AGENT CHECK IN BAGGAGE © ADJUST SERVICE TIME CHECK IN BAGGAGE G> QUEUE ONE LINE SPLIT GROUP > SELECT QUEUE FOR EXCESS PAYMENT 0 YES PAY EXCESS CHARGE QUEUE AGENT NO PURCHASE TICKET SEND (3.4) SPBB (3.32) SECTION 3.4 END OF CHECK IN END APPENDIX III PROGRAM LISTING AND SAMPLE OUTPUT $RUN * G P S 5 P A R = S I Z E = C ** S I M U L A T E ** #* ** S E C T I O N ONE - I N I T I L I Z A T I O N jj» 9|C ijs *|C 3jC J^C Sjj 5jC *J5 5}S *jc 5,C 5jC *jc ij? 5jc 5jS 3jC ?JC 5jC «JC «{C 1.1 F A C I L I T Y P O L I C Y : NO. OF COUNTER A V A I L A B L E FOR P A S S E N G E R P R O C E S S I N G . IN I T I A L I N I T I A L I N I T I A L XH1 ,4 X H 2 , 1 4 XH3 ,K17 1.2 ** A R R I V A L RATE PER I N I T I A L X H 7 , 5 0 1.3 ** PARAMETER V A L U E S XH1 = NO. OF COUNTERS OPEN XH2 = 16 - NO. OF COUNTERS OPEN XH3 = 16 + NO. OF COUNTERS OPEN 15 MINUTE P E R I O D . XH7 = NO. OF A R R I V A L S - P A R T S PER THOUSANDS ( R P B S ] ( R P T S ) ( S P B S ] I N I T I A L I N I T I A L I N I T I A L I N I T I A L I N I T I A L I N I T I A L X H 3 0 , 6 7 0 X H 3 1 , 1 5 0 X H 3 2 , 1 5 0 X H 3 3 , 1 1 8 X H 3 4 , 1 5 0 X H 3 5 , 1 5 0 R E VENUE GROUPS {%) REVENUE GROUPS WITHOUT T I C K E T S . STANDBY GROUPS WITHOUT T I C K E T S . REVENUE GROUPS U S I N G C U R B S I D E C H E C K - I N . REVENUE U N I T S WITH E X C E S S BAGGAGE. STANDBY U N I T S WITH E X C E S S BAGGAGE. 5j5 3|« S E C T I O N TWO -** •A, ^t. •JV vO 0< v*r* N** <JU> v**-~t** Jfr >j*. • V «Y> ** ** 2.1 ** F U N C T I O N S 1 FUNC TION R N l , C 2 4 0 0 .1 . 1 0 4 .2 .6 . 9 1 5 .7 1.2 .75 .90 2.3 .92 2.52 .94 . 98 3.9 .99 4.6 . 9 9 5 ** 2 F U N C T I O N R N 2 , 0 3 .1 6 10 .30 14 1.0 3 FUNC TION RN3,D5 . 42 2 .60 3 .81 4 F U N C T I O N RN4 ,C7 0 60 .12 120 .39 1.0 3 6 0 ** 5 FUNC TION RN5,C6 0 35 .30 6 0 .58 ** 6 FUNC TION P 4 0 , 0 5 2 6 0 3 90 4 N E G A T I V E E X P O N E N T I A L F U N C T I O N .222 .3 . 3 5 5 .4 . 5 0 9 .5 .69 1.38 .8 1.6 .84 1.83 .88 2.12 2 . 8 1 .95 2 . 9 9 .96 3.2 .97 3.5 5.3 .998 6.2 . 9 9 9 7.0 . 9 9 9 7 8.0 GROUP A R R I V A L C O M P O S I T I O N 2 0 GROUP S I Z E D I S T R I B U T I O N 4 .91 5 1.0 6 S E R V I C E T I M E TO PURCHASE T I C K E T 150 .58 180 .85 2 4 0 .92 3 0 0 S E R V I C E T I M E TO PAY E X C E S S CHARGE 9 0 .75 120 .85 150 1.0 180 BAGGAGE COUNTER S E R V I C E T I M E FOR GROUPS 125 5 160 6 180 ** ** ** ** ** 1 2 3 4 5 6 7 8 9 ** 1 2 3 4 5 6 7 8 9 ** 2.2 ** V A R I A B L E S FOR F A C I L I T Y S E L E C T I O N 2 . 2 1 * * REVENUE BAGGAGE COUNTERS ** 11 12 13 14 15 16 ** ** ** 17 18 17 18 ** 10 ** B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E B V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E F l F 2 F3 F 4 F5 F 6 F 7 F 8 F 9 BV1+Q1 BV2+Q2 BV3+Q3 BV4+Q4 BV5+Q5 B V 6 + 0 6 BV7+Q7 BV8+Q8 BV9+Q9 BOOLEAN V A R I A B L E S I F F A C I L I T Y F ( J ) I S I N USE , B V ( J ) W I L L =1, I F NOT B V ( J ) = 0 . A R I T H M E T I C V A R I A B L E S V A R I A B L E V ( J ) W I L L C A L C U L A T E THE LENG T H OF QUEUE 0 ( J ) FOR F A C I L I T Y S E L E C T I O N P R O C E S S 2.22 ** REVENUE T I C K E T COUNTERS 11 B V A R I A B L E F l l * 12 B V A R I A B L E F 1 2 • 13 B V A R I A B L E F 1 3 • 14 B V A R I A B L E F 1 4 • 15 B V A R I A B L E F 1 5 • 16 B V A R I A B L E F 1 6 * . S I M I L A R TO S E C T I O N 2 . 2 1 BOOLEAN, V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E V A R I A B L E B V 1 1 + 0 1 1 B V 1 2 + Q 1 2 B V 1 3 + Q 1 3 B V 1 4 + 0 1 4 B V 1 5 + Q 1 5 B V 1 6 + 0 1 6 S I M I L I A R TO S E C T I O N 2.21 A R I T H M E T I C . 2 . 2 3 ** STANDBY COUNTERS B V A R I A B L E B V A R I A B L E V A R I A B L E V A R I A B L E F 1 7 F 1 8 B V 1 7 + 0 1 7 B V 1 8 + 0 1 8 S I M I L I A R TO S E C T I O N S 2.21 + 2 . 2 2 . F V A R I A B L E ( 9 0 0 * 2 9 ) / ( X H 7 * 1 0 ) COMPUTES MEAN GROUP I N T E R A R R I V A L S E C T I O N THREE - THE C H E C K - I N SYSTEM ** GROUP A R R I V A L S - EACH P A S S E N G E R GROUP WILL HAVE AN OF C H A R A C T E R I S T I C S UPON A R R I V A L AT A I R P O R T ** SET ** THE** ** GENERATE V 1 0 , F N 1 f , , , 5 0 , H C R E A T I O N OF GROUPS MARK 10 TO COMPUTE TIM E I N S Y S T E M A S S I G N 3 0 , X H 3 0 * A S S I G N 3 1 , X H 3 1 • A S S I G N 3 2 , X H 3 2 • A S S I G N M E N T OF PARAMETER P ( J ) A S S I G N 3 3 , X H 3 3 • OF HALFWORD S A V E V A L U E X H ( J ) . A S S I G N 3 4 , X H 3 4 • A S S I G N 3 5 , X H 3 5 * VALUE 1 1 * * D E T E R M I N A T I O N OF GROUP S I Z E AND T Y P E TRANSFER FN2 FN2= D I S T R I B U T I O N OF GROUP TYPE A S S I G N A S S I G N A S S I G N TRANSFER A S S I G N A S S I G N A S S I G N A S S I G N A S S I G N TRANSFER A S S I G N ASS IGN A S S I G N TRANSFER 4 0 , 1 42 , 1 7 ,45 , AAA 41 ,FN3 4 2 , P 4 1 4 1 - , 1 1 4 5 AAA 4 0 , 7, 4 0 , F N 3 4 2 , P 4 0 7 ,FN6 , A AA P 4 0 = l P A S S E N G E R A R R I V E S ALONE NO. OF P A S S E N G E R S PER GROUP MEAN S E R V I C E T I M E AT BAGGAGE COUNTER TRANSFER T O • A L L O C A T I O N • S E C T I O N ( A A A ) GROUP A R R I V A L OF I N D E P E N D E N T P A S S E N G E R S P41= NO. IN GROUP P4 0 = S E R V E D I N D E P E N D E N T OF GROUP P7= MEAN S E R V I C E T I M E AT BAGGAGE COUNTEr TRANSFER T O ' A L L O C A T I ON• S E C T I O N ( A A A ) F A M I L Y GROUP A R R I V A L P40= S I Z E OF F A M I L Y FN6= MEAN S E R V I C E T I M E BY GROUP S I Z E 12 ** A L L O C A T I O N OF GROUPS TO S U B - S Y S T E M S ** S A V E V A L U E S A V E V A L U E T E S T E TRANSFER 20 + ,1 2 1 + , P 4 2 P 4 0 , l , R P S . * 3 0 , R P S ,SPS COUNT NO. OF GROUP A R R I V A L S -TOTAL COUNT NO. OF P A S S E N G E R S -TOTAL F A M I L I E S T RAVEL R E V E N U E ONLY I F P 4 0 = l , T R A N S A C T I O N I S NOT A F A M I L Y I D E N T I F I C A T I O N OF REVENUE PASSENGER GROUPS RPS= R E V E N U E P A S S E N G E R S Y S T E M SPS= STANDBY ** *# 3, ** ** 3, ** RPS CURB 2 ** REVENUE P A S S E N G E R S Y S T E M 21 ** A L L O C A T I O N TO REVENUE S U B S Y S T E M S ** ** ** ** 3.22 ** ** R P B S ** R P S S S A V E V A L U E 8 + , l COUNT = TOTAL REVENUE GROUPS S A V E V A L U E 9 + , P 4 2 COUNT = TOTAL R E V E N U E P A S S E N G E R S TRANSFER . * 3 1 , CURB , RP TS I D E N T I F I C A T I O N OF N O N - T I C K E T E D GROUPS RP T S = REVENUE P A S S . T I C K E T S Y S T E M S A V E V A L U E 24+,1 COUNT = TOTAL REVENUE GROUPS T I C K E T E D SAVE V A L U E 2 5 + , P 4 2 COUNT = TOTAL R E V E N U E P A S S . T I C K E T E D TRANSFER . * 3 3 ,RPBS ,CEND I D E N T I F I C A T I O N OF R E V E N U E GROUPS U S I N G C U R B S I D E C H E C K - I N . R P B S = REVENUE P A S S . BAGGAGE S Y S T E M CEND= END OF C H E C K - I N FOR CURBS IDE R P B S - REVENUE P A S S E N G E R BAGGAGE S Y S T E M RPBA RPBB RPBC RPBD S E L E C TMIN 1,1,XH1,,V MARK 1 21 S P L I T P 4 1 , R P B A TRANSFER , RPBA S E L E C TMIN 1,1,XH1 , ,V 1 MARK 1 21 QUEUE P I S E I Z E P I D E P A R T P I MARK 22 T A B U L A T E 1 A S S I G N 2 2 - , P 2 1 T E S T G P 2 2 , 1 5 0 , R P B B i S A V E V A L U E 2 1 + , P 4 0 , H i S A V E V A L U E 2 8 + , 1 l S A V E V A L U E 2 2 + , P 4 0 , H S A V E V A L U E 2 9 + , 1 i TRANSFER . * 3 4 , R P B C ,RPBD ADVANCE P7 ,FN1 R E L E A S E P I T RANSFER , END ASS IGN 7 + ,30 S A V E V A L U E 1 3 + , P 4 0 S A V E V A L U E 30 + , 1 ADVANCE P 7 , F N 1 R E L E A S E P I A S S I G N 3,1 TRANSFER ,RP TB S E L E C T I O N OF SHORTEST QUEUE. QUEUE NO. I S P L A C E D I N PARAMETER P1.XH1 = NO. OF COUNTERS OPEN TO COMPUTE TIM E I N QUEUE AT R P B S S E P A R A T E GROUPS OF I N D E P E N D E N T P A S S . S E L E C T I O N OF SHORTEST QUEUE FOR P A S S . FROM R P T S - T I C K E T PURCHASE TO COMPUTE T I M E I N QUEUES AT R P B S . WAIT IN QUEUE FOR S E R V I C E TIME P A S S . L E A V E S QUEUE FOR S E R V I C E TIME I N S Y S T E M C A L C U L A T E D WAS T I M E . G . 2 . 5 M I N . Y E S PROCEED. COUNT= NO. P A S S . LONGER THAN 2.5 M I N . COUNT= NO. P A S S . U N I T S LONGER THAN 2.5 COUNT= TOTAL P A S S E N G E R S COUNT = TOTAL P A S S . U N I T S E X C E S S B A G G A G E - R P B C ; NO E X C E S S - R P B D : E X C E S S S E R V I C E T I M E =P7 * F U N C T I O N 1 S E R V I C E I S C O M P L E T E D GO TO END S E C T I O N 3.34 A D J U S T S E R V I C E T I M E FOR E X C E S S BAGGAGE COUNT= NO. P A S S E N G E R S WITH E X C E S S . COUNT= NO. P A S S . U N I T S WITH E X C E S S S E R V I C E TIME =P7 * F U N C T I O N 1 S E R V I C E I S COMPLETE I D E N T I F Y P A S S . U N I T WITH E X C E S S : P 3 = l GO TO R E V . P A S S . T I C K E T 'B« ** 3 . 2 3 ** #* RPTS R P TA S P L I T SEL EC TM IN RPTB R P TE RP TC RP TD ** ** 3. ** 3. ** S P S TRANSFER S E L EC TMIN MARK QUEUE S E I Z E D E P A R T MARK TABULATE A S S I G N TEST G S A V E V A L U E SAVE VALUE S A V E V A L U E S A V E V A L U E TEST E ADVANCE R E L E A S E TRANSFER ADVANCE R E L E A S E A S S I G N TRANSFER P 4 1 , R P T A 2 , X H 2 , 1 5 , , V »RP TE 2 ,XH2,16 , ,V 2 4 P2 P2 P2 2 5 2 2 5 - , P 2 4 P2 5 , 1 5 0 , R P TC 2 4 + , P 4 0 , H 31 + , 1 25 + ,P40 ,H 3 2 + , 1 P 3 , K 1 , R P TD 1 ,FN5 P2 , END 1 ,FN4 P2 6 + , 1 ,RPSS D I V I D E INTO P A S S E N G E R U N I T S S E L E C T SHORTEST Q U E U E ( X H 2 - 1 5 ) P L A C E NO. I N PARAMETER 2.NON-TICK ETED PASS, GO TO RPTE S E L E C T I O N OF SHORTEST QUEUE FOR P A S S , WITH E X C E S S BAGGAGE ONLY TO COMPUTE T I M E I N QUEUES AT R P T S * . WAIT IN QUEUE FOR S E R V I C E * TIME P A S S . L E A V E S QUEUE FOR S E R V I C E TIME I N S Y S T E M C A L C U L A T E D WAS TIME .G. 2.5 M I N . YES PROCEED COUNT = P A S S . W A I T I N G LONGER THAN 2. COUNT = P A S S . U N I T S LONGER THAN 2. COUNT = TOTAL P A S S . I N R P T S COUNT = TOTAL P A S S . U N I T S I N RPTS I F P A S S . HAS E X C E S S P3= l , PROCEED . S E R V I C E FOR E X C E S S PAYMENT.GO TO • * GO TO 'END' S E C . 3.4 END' * S E R V I C E FOR PURCHASE OF T I C K E T I D E N T I F Y P A S S . WITH T I C K E T P 6 = l GO TO REVENUE P A S S . B A G G A G E ' S 1 , S E C . 3 . 2 2 ** STANDBY P A S S E N G E R S Y S T E M 31 ** A L L O C A T I O N TO S U B - S Y S T E M S A S S I G N 5 + , l I D E N T I F Y STANDBY P A S S E N G E R P 5 = l S A V E V A L U E 3 9 + , P 4 2 COUNT = TOTAL STANDBY P A S S E N G E R S SAVE V A L U E 4 0 + , 1 COUNT = TOTAL STANDBY GROUPS TRANSFER . # 3 2 , S P B S , S P T S I N D E N T I F Y STANDBY GROUPS N O N - T I C K E T E D ** ** ** S P B S S P B B ** 3.32 ** S P B S - STANDBY P A S S E N G E R BAGGAGE S Y S T E M S P L I T S EL EC TMIN MARK QUEUE S E I Z E DEPAR T MARK P4 1 ,SPBB 4 , 1 7 , X H 3 , , V 27 P4 P4 P 4 28 S E P A R A T E GROUPS OF I N D E P E N D E N T PASSENGER S E L E C T I O N OF SHORTEST QUEUE. QUEUE NO. I S P L A C E D IN PARAMETER P4 . X H 3 - 1 6 =NU. OF COUNTERS O P E N . TO COMPUTE T I M E I N QUEUES AT SPS-BAGGAGb * . W A I T I N G IN QUEUE FOR S E R V I C E TIME P A S S E N G E R L E A V E S QUEUE ** SPBC ** ** SPBD S P B E ** 3 ** S P T S SPTA ** S P T B ** S P T C SPTD S P T E TABULA TE 3 A S S I G N 2 8 - , P 2 7 TIME IN S Y S T E M C A L C U L A T E D T E S T G P 2 8 , 1 5 0 , SPBC WAS TIME G R E A T E R THAN 2.5 M I N . - S P S C ( N G ) SAVE V A L U E 2 7 + , P 4 0 , H COUNT = TOTAL P A S S E N G E R OVER 2.5 MIN S A V E V A L U E 41 + , 1 COUNT = TOTAL P A S S . U N I T S OVER 2.5 MIN S A V E V A L U E 2 8 + , P 4 0 , H COUNT = TOTAL STANDBY P A S S . AT BAGGAGE S A V E V A L U E 42 + , 1 COUNT = TOTAL STANDBY P A S S . U N I T S BAGG. TRANSFER . * 3 5 , S P B D , S P B E I D E N T I F Y P A S S . WITH E X C E S S - S P B E ADVANCE P 7 , F N 1 S E R V I C E TIME TO C H E C K - I N R E L E A S E P4 L E A V E COUNTER TRANSFER , S END GO TO 'END' S E C . 3.4 A S S I G N 7+,30 A D J U S T MEAN F O R ' E X C E S S • T I M E S A V E V A L U E 1 7 + , P 4 0 COUNT = NO. P A S S E N G E R S WITH E X C E S S S A V E V A L U E 43 + , 1 COUNT = NO. P A S S E N G E R U N I T S E X C E S S ADVANCE P 7 , F N 1 S E R V I C E TIME TO C H E C K - I N WITH E X C E S S R E L E A S E P4 L E A V E COUNTER A S S I G N 3,1 I D E N T I F Y P A S S E N G E R S WITH E X C E S S TRANSFER ,SPTB GO TO STANDBY P A S S . T I C K E T »B' S E C . 3.^ 1 3 SPTS - STANDBY P A S S E N G E R T I C K E T S Y S T E M * A S S I G N 2+,16 P A S S . QUEUE AT #16 TO BUY A T I C K E T S A V E V A L U E 3 3 + , P 4 2 COUNT = NO. P A S S . N O N - T I C K E T E D S A V E V A L U E 34 + , 1 COUNT = NO. P A S S . U N I T S N O N - T I C K E T E D S P L I T P 4 1 ,SPTA S E P A R A T E GROUPS OF INDEPENDENT P A S S . TRANSFER ,SPTC GO T O ' S P T C S E L E C T M I N 2 ,XH2 ,16 » t V S E L E C T SHORTEST QUEUE P L A C E N O . I N P2 MARK 24 TO COMPUTE T I M E I N QUEUES QUEUE P2 * S E I Z E P2 . WAIT I N QUEUES FOR S E R V I C E D E P A R T P2 * MARK 25 TIME P A S S . L E A V E S QUEUE TABULA TE 4 A S S I G N 2 5 - , P 2 4 C A L C U L A T E T I M E I N QUEUE T E S T G P 2 5 , 1 5 0 , SPTD WAS TIME GREATER THAN 2.5 M I N . - S P T D S A V E V A L U E 3 5 + , P 4 0 COUNT = NO. P A S S . GREATER THAN 2.5 M I N . S A V E V A L U E 36 + , 1 COUNT = NO. P A S S . U N I T S GRT TH 2.5 M I N . S A V E V A L U E 3 7 + , P 4 0 COUNT = TOTAL P A S S E N G E R S AT S P T S S A V E V A L U E 38 + , 1 COUNT = TOTAL P A S S . U N I T S AT S P T S TEST E P 3 , 1 , S P T E I F FOR E X C E S S PAYMENT - C O N T I N U E ADVANCE 1 t FN5 * S E R V I C E TO PAY E X C E S S R E L E A S E P2 TRANSFER , S END GO TO 'SEND' S E C . 3.4 ADVANCE 1 ,FN4 S E R V I C E TO PURCHASE T I C K E T R E L E A S E P2 TRANSFER ,SPBB GO TO STANDBY P A S S . BAGGAGE • B 1 SEC.3.< ** ** ** END CEND SEND SAVEVALUE SAVEVALUE MARK SAVEVALUE SAVEVALUE TERM INA TE SAVEVALUE SAVEVALUE TERM INA TE SAVEVALUE SAVEVALUE MARK SAVEVALUE SAVE VALUE TERMINATE 1+,P40 44+, 1 11 2+,V51 45+,V38 11+,P40 48 + ,1 3+,P40 46 + , 1 11 4+,V51 47+,V38 COUNT = NO. REV. PASSENGERS CHECKED-IN COUNT = NO. REV. PASS. UNITS CHKD IN COMPUTE TIME LEAVE SYSTEM CALCULATE=TIME * NO.PASSENGERS =AVG.T.1 CALCULATE = TI ME FOR AVERAGE TRANSIT TH END COUNT = NO. CURBSIDE PASSENGERS COUNT = NO. CURBSIDE PASSENGER UNITS SAME AS 'END',ONLY THIS IS FOR STANDE PASSENGER CALCULATIONS 4.0 ** SECTION FOUR - CALCULATIONS AND CONTROL 4.1 ** VARIABLES FOR STATISTICAL CALCULATIONS CALCULATIONS ARE ASSIGNED TO SAVEVALUES IN SECTION 4.2* CORRESPONDING SAVEVALUES APPEAR ON THE RIGHT. 30 FVARIABLE 31 FVARIABLE 32 FVARIABLE 34 FVARIABLE 35 FVARIABLE 36 FVARIABLE 37 FVARIABLE 38 VARIABLE 39 FVARIABLE 40 VARIABLE 41 VARIABLE 42 FVARIABLE 43 VARIABLE 44 VARIABLE 45 VARIABLE 46 VARIABLE 47 VARIABLE 48 VARIABLE 49 FVARIABLE 50 FVARIABLE 51 VARIABLE 52 FVARIABLE 53 FVARIABLE 54 FVARIABLE (X28*1000)/X29 (X31*1000)/X32 (X41*1000)/X42 (X36*1000)/X38 (X8*1000 J/X20 1000-V35 (X9*1000)/X21 P11-P10 X8-X24 16-XH2 XH3-16 (V39*1000)/X8 RPBS-P.U. .G.2.5 (X55) RPTS-P.U. .G.2.5 (X56) SPSB-P.U. .G.2.5 (X57) SPST-P.U. .G.2.5 (X59) % REV. GROUP ARRIVALS (XH42) % STB. GROUP ARRIVALS (X60) % REV. PASSENGERS TO TOTAL (X49) PASS. UNIT TRANSIT TIME REV. GROUPS NON-TICKETED (X50) NO. REV. TICKET COUNTERS (XH40) NO. STB. BAGG. COUNTERS (XH41) % REV. GROUPS NON-TICKETED (X51) FR1+FR2+FR3 + FR4+FR5+FR6 + FR7+FR8 + FR9 REV.BAG.UT ILIZ, V43/XH1 AVG. UTILIZATION-RPBS FR11+FR12+FR13+FR14+FR15 V45/V40 FR18+FR17 V47/V41 X9-X25 {V49*1000)/X9 (P11-P10 )=.P40 (X48*1000)/X8 (Xll*1000)/X9 (X39*1000)/X21 AVG. UTILIZATION-RPTS AVG. UTILIZATION-SPSB NO. PASS.UNITS NON-TICKETED % PASS. UNITS NON-TICKETED TRANSIT TIME- REV.PASSENGERS % GROUPS CURBSIDE CHECK-IN % PASS. CURBSIDE CHECK-IN (XH44) (XH46) (XH48) ( X5 ) (X6) (X7) ( X10) % STANDBY PASSENGERS OF T0TAL(X12) 55 56 57 58 59 60 61 62 63 64 65 ** ** ** ** *# FVARIABLE (X34*1000)/X40 STB. FVARIABLE (X33* 1000)/X39 % STB FVARIABLE (X30*1000 )/X29 % REV FVARIABLE (X13*1000)/XH22 % REV FVARIABLE (X43*1000)/X42 % STB FVARIABLE (X17*1000)/XH28 % STB FVARIABLE (X45/X44./6 AVG. FVARIABLE (X2/XD/6 AVG. FVARIABLE (X47/X46)/6 AVG. FVARIABLE (X4/X3)/6 AVG. VARIABLE FR16 STB. GROUPS NON-TICKETED-^ (X14) . PASS. NON-TICKETED (X16) .PASS.UNITS EXCESS (X18) .PASS. EXCESS (X19) . PASS.UNITS EXCESS (X53) . PASS.EXCESS (X54) TRANSIT TIME - REV.PU. (X61) TRANSIT TIME - REV P. (X62) TRANSIT TIME - STB PU. ( X63) TRANSIT TIME - STB P. (X64) UTILIZATION -SPST (X65) 4.2 ASSIGNMENT OF VARIABLE VALUES FOR OUTPUT VALUES OF SECTION 4.1 ASSIGNED TO SAVEVALUES IN 4.2 FOR PRINTING IN REQUESTED FURMAT OF OUTPUT GENERA TE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE SAVEVALUE TERM I NA TE 300 40,V40,H 41 ,V41,H 42,V35,H 44,V44 ,H 46,V46,H 48, V48,H 5 ,V49 6, V50 7, V52 10,V53 12,V54 14,V55 16,V56 18 ,V57 19,V58 49, V37 50, V39 51, V42 53, V59 54, V60 55, V30 56, V31 57, V32 59, V34 60, V36 61, V61 62, V62 63, V63 64, V64 65, V65 ** ** 1 2 3 4 ** ** ** T A B L E M P 2 1 , 0 , 3 0 , 1 4 TABLE M P 2 4 , 0 , 3 0 , 1 4 TABLE M P 2 7 , 0 , 3 0 , 1 4 TABLE M P 2 4 , 0 , 3 0 , 1 4 P A S S . U N I T W A I T I N G I N QUEUES. TIME C A L C U L A T E D I N I N T E R V A L S OF 3 0 S E C TABLE NO. R E F E R S TO ' T A B U L A T E 1 CARD I N MAIN PROGRAM S E C T I O N 3. 4.4 CONTROL OF S I M U L A T E D TIME GENERA TE S A V E V A L U E T E R M I N A T E S T A R T 60 9 + , l , H 1 120 G E N E R A T E EVERY 60 SECONDS COUNT = NO. OF S I M U L A T E D T I M E :MINUTES END = ONE MINUTE S TAR T= AMOUNT OF T I M E TO S I M U L A T E S E C T I O N F I V E - R EQUESTED OUTPUT $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ R E P O R T E J E C T S P A C E 12 TEXT 31 TEXT S P A C E 10 TEXT S P A C E 12 TEXT 12 TEXT 12 T E X T S P A C E 12 TEX T 12 TEXT 12- T EXT 12 TEXT SSENGER UN I T 12 T E X T , 2 / 1 L X X X ,X#Sg S P A C E 12 T E X T 12 TEXT 12 T E X T COUNTERS 1 1 - 1 5 ONLY) THE MEAN A R R I V A L RATE WAS # XH7,2/XXXX# P A S S E N G E R S PER 15 MINUTE P E R I O D . 1 REVENUE P A S S E N G E R S Y S T E M 1 T R A N S I T TIME IN THE S Y S T E M AVERAGE TIME PER P A S S E N G E R U N I T # X 6 1 , 2 / 1 L X X X . X # M I N . A V E R A G E TIME PER P A S S E N G E R # X 6 2 , 2 / 1 L X X X . X # M I N . 1 R E VENUE P A S S E N G E R BAGGAGE S Y S T E M NUMBER OF COUNTERS OPEN # X H 1 , 2 / X X X # . A V E R A G E U T I L I Z A T I O N OF F A C I L I T I E S # X H 4 4 , 2 / 1 L X X X . X # % AVERAGE W A I T I N G TIME I N QUEUES # T 1 , 3 / X X X X # S E C . PER P£ % OF P A S S E N G E R U N I T S I N QUEUES OVER 2.5 M I N . WAS #X55 REVENUE P A S S E N G E R T I C K E T S Y S T E M NUMBER OF COUNTERS OPEN # X H 4 0 , 2 / X X X X # . A V E R A G E U T I L I Z A T I O N OF F A C I L I T I E S # X H 4 6 , 2 / 1 L X X X . X # % ( 12 T E X T AVERAGE W A I T I N G TIME I N QUEUES # T 2 , 3 / X X X X # S E C . PER P/! SSENGER U N I T ( 1 1 - 1 5 ) . T E X T 2.5 M I N . WAS # X 5 6 , 2 / 1 L X X X . X # % S P A C E T E X T S P A C E TEXT TEXT TEXT S P A C E T E X T TEXT TEXT TEXT 12 VER 10 12 12 12 12 12 12 12 SSENGER U N I T . 12 T E X T l L X X X . X # 5 g S P A C E 12 TEXT 12 TEXT 12 TEXT A S S E N G E R U N I T . 12 TEXT 1 L X X X . X # % ** % OF P A S S E N G E R U N I T S IN A L L T I C K E T OR E X C E S S QUEUES C # X 6 3 , 2 / l L X X X . X # M I N . # X 6 4 , 2 / 1 L X X X . X # M I N . STANDBY P A S S E N G E R SYSTEM 1 T R A N S I T TIME I N THE S Y S T E M AVERAGE TIME PER P A S S E N G E R U N I T AVERAGE TIME PER P A S S E N G E R 1 STANDBY P A S S E N G E R BAGGAGE S Y S T E M NUMBER UF COUNTERS OPEN # X H 4 1 , 2 / X X X X # AVERAGE U T I L I Z A T I O N OF F A C I L I T I E S # XH48, 2/ 1 L X X X . X#% AVERAGE W A I T I N G TIME I N QUEUES # T 3 , 3 / X X X X # S E C . PER Pt % P A S S E N G E R U N I T S IN QUEUES OVER 2.5 M I N . WAS # X 5 7 , 2 / STANDBY P A S S E N G E R T I C K E T SYSTEM AVERAGE U T I L I Z A T I O N OF F A C I L I T I E S # X 6 5 , 2 / 1 L X X X . X # % AVERAGE W A I T I N G TIME I N QUEUES # T 4 , 3 / X X X X # S E C . PER . * P A S S E N G E R U N I T S IN QUEUES OVER 2.5 M I N . WAS # X 5 9 , 2 / APPENDIX I I I SAMPLE OUTPUT REVENUE PASSENGER SYSTEM TRANSIT TIME IN THE SYSTEM AVERAGE TIME PER PASSENGER UNIT 2.8 MIN. AVERAGE TIME PER PASSENGER 3.1 MIN. REVENUE PASSENGER BAGGAGE SYSTEM NUMBER OF COUNTERS OPEN 5. AVERAGE UTILIZATION OF FACILITIES 59.6% AVERAGE WAITING TIME IN QUEUES 21 SEC. PER PASSENGER UNIT % OF PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 5.5% REVENUE PASSENGER TICKET SYSTEM NUMBER OF COUNTERS OPEN 3. AVERAGE UTILIZATION OF FACILITIES 41.8% (COUNTERS 11-15 ONLY) AVERAGE WAITING TIME IN QUEUES 32 SEC. PER PASSENGER UNIT (11-15). % OF PASSENGER UNITS IN ALL TICKET OR EXCESS QUEUES OVER 2.5 MIN. WAS 9.1% STANDBY PASSENGER SYSTEM TRANSIT TIME IN THE SYSTEM AVERAGE TIME PER PASSENGER UNIT 2.3 MIN. AVERAGE TIME PER PASSENGER 2.3 MIN. STANDBY PASSENGER BAGGAGE SYSTEM NUMBER OF COUNTERS OPEN 2 AVERAGE UTILIZATION OF FACILITIES 34.1% AVERAGE WAITING TIME IN QUEUES 41 SEC. PER PASSENGER UNIT % PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 10.7% STANDBY PASSENGER TICKET SYSTEM AVERAGE UTILIZATION OF FACILITIES 45.5% AVERAGE WAITING TIME IN QUEUES 104 SEC. PER PASSENGER UNIT. % PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 27.0% APPENDIX IV USER'S MANUAL The user's manual for the simulation model (Appendix III) wil l be presented in five sections: I. Reference Material. Basic support material wi l l be provided concerning the language and computer time used for simulating the system. II. Identification of the Simulation Model. Working closely with the program l ist ing, the basic sections of the model wi l l be identified. No programming knowledge is expected in this section. III. A Program for Analysis. A step by step approach wil l be presented to enable the user to carry out the first two steps of the systems analysis as presented in Chapter III. Only a very basic knowledge of computer use is expected. ( IV. The Verification Steps. The adjustment of the main program for the verification steps contained in Chapter III wi l l be presented. V. Analytical Steps. Similar to Section IV, the adjust-ments made for the analytical steps wil l be presented. on the GPSS VERSION V compiler. Most computer companies have s i m i l a r languages. Operating Equipment: The U n i v e r s i t y of B r i t i s h Columbia has an IBM360/67 Duplex computer with a v a r i e t y of software. Time Usage: The computer time usages (assembly and execu-tion) f o r a simulation run of 180 minutes i s approximately .7 minutes. IBM Manuals: 1. General Purpose Simulation System/360 User's Manual Reference Number GH20-0326-3. 2. GPSS/360 Introductory User's Manual Reference Number H20-0304-0. I I . IDENTIFICATION OF THE SIMULATION MODEL The program l i s t i n g i n Appendix I I I i s subdivided into f i v e sections. The user should follow along with the l i s t i n g to f a m i l i a r i z e himself with the computer 'deck' make-up. Introductory Statements 1. Each l i n e i n the l i s t i n g represents 1 computer card. Each card consists of 80 spaces or columns. 2. Each alphabetic or numerical character or blank represents 1 space or column. 3. Column '1' i s on the l e f t hand side . 4. An ' * ' asterik in column '1' represents a comment card or l ine. It may be used to comment on the operation of the deck or as a 'space' card. 5. A l l other cards are operational. The two cards prior to Section 1 are control cards and must start each deck. The first card is : Column Controls 1 $ 2-4 RUN 6 * 7-10 GPS5 13-22 PAR=SIZE=C The second card is : Column 8-15 SIMULATE The program has been called and activated. Section 1. Section one is the initialization of the system or the setting of the parameters. It consists of three sub-sections: Section 1.1. Facility Policy. The number of counters that are available are stated here. Section 1.2. Arrival Rate. The number of passengers that arrive in a 15-minute period is stated in this section. Section 1.3. The percentage of passengers having spe-cif ic characteristics (non-ticketed, etc.) is stated here. The above three sub-sections wi l l be explained under Program for Analysis. The essence of this section is to set the in i t i a l conditions of the system Section 2. Section two contains two sub-sections. The first is the specifications of the functions used to describe the service times, and type and group composition. The second sub-section contains the variables for the facility selection process described in Chapter III. Section 3. Section three contains the program of the passenger flow through the various systems. The subsections are: 3.10 Arrival of groups 3.11 Determination of Group size and type 3.12 Allocation of groups to sub-systems (the Arrival Stream) 3.2 Revenue Passenger System 3.21 Allocation to Revenue Subsystems 3.22 Revenue Baggage System 3.23 Revenue Ticket System 3.3 Standby Passenger System 3.31 Allocation to Subsystems 3.32 Standby Baggage System 3.33 Standby Ticket System 3.4 End of Flow for Passengers On the right hand side starting in Column 33 there is a narrative of the process performed by each line. This thesis suggests that an attempt be made to read through the program. Section 4. This section is for calculations and control of the simulation. 4.1 Calculation of system statistics. 4.2 Transfer of statistics to 'save values' for reference. 4 .3 Table identification for calculation of transit times. 4.4 Section 4.4 is very important as i t controls the length of simulated time. Both this section (4.4) and Section 1 wi l l be discussed in Program for Analysis. Section 5. This section is the request for the output to be printed in a certain manner. If this section is omitted, GPSS/360 has an automatic output. After Section 5, there is another control card: Columns 8-10 END If Section 5 is omitted, the card is entered after Section 4. I l l PROGRAM FOR ANALYSIS This section wi l l enable the user to carry out an analysis similar to Analysis Step I or Step II as cited in Chapter III. Step 1. (a) Select the facility policy that is in operation—the number of counters open for the Revenue baggage and ticket counter and the Standby baggage counter. (b) Remove the three cards in Section 1.1. (c) Revenue Baggage counters open. Columns 8-14 INITIAL Columns 19-22 XHI, Column 23 print the number of counters open. For example, i f 5 were open the card would read: INITIAL XHI, 5 (d) Revenue Ticket Counters open. Columns 8-14 INITIAL Columns 19-22 XH2 Columns 23-24 print the value of 16—no. of counters open For example, i f 3 ticket counters were open, the card would read: INITIAL XH2, 13 (e) Standby Baggage Counters Columns 8-14 INITIAL Columns 19-22 XH3 Columns 23-24 print the value of 16 + number of Standby counters open. For example, i f two baggage Standby counters were open, the card would read INITIAL XH3, 18 (f) Replace Section 1.1 with the three new cards. our example, Section 1.1 would read: INITIAL XHI, 5 INITIAL XH2, 13 INITIAL XH3, 18 Step 2. (a) The a r r i v a l rate of passengers per 15 minute period. Remove the card i n Section 1.2 that i s marked INITIAL. (b) Select the a r r i v a l rate f o r a 15 minute period (c) Columns 8-14 INITIAL Columns 19-22 XH7 Columns 23, 24 & 25 the a r r i v a l rate. For example, i f the a r r i v a l rate was 100 passengers per 15 minute period, the card would read: INITIAL XH7, 100 (d) Replace the new card i n Section 1.2 Step 3. (a) Section. 1.3 contains 6 cards marked INITIAL ( s t a r t i n g i n column '8') and specif y the percentage of passengers that a r r i v e at the a i r p o r t with c e r t a i n c h a r a c t e r i s t i c s . Start i n column '19' there i s an 'XH—,' The two numbers a f t e r the 'XH' reference c e r t a i n passengers c h a r a c t e r i s t i c s . 'XH30,' = the percentage of Standby groups to t o t a l inde-pendent groups. Remember that 70% of group a r r i v a l s are f a m i l i e s and therefore dependent. 'XH31,' = the percentage of Revenue groups without t i c k e t s 'XH32,' = the percentage of Standby groups without t i c k e t s . 'XH33,' = the percentage of revenue groups t o t a l allowed ( i . e . minus non-ticketed groups) that use curb-side s e r v i c e . 'XH34,' = the percentage of Revenue passenger units with excess baggage. 'XH35,1 = the percentage of Revenue passenger units with excess baggage. The percentages must be specified in parts per thousand. Thus i f 15% of Standby passenger units had excess baggage, the card would read: INITIAL XH35, 150 (b) Remove cards in Section 1.3 (c) Column 8-14 INITIAL Column 19-20 XH Column 21-22 the reference number Column '23' print ' , ' Column '24'-'26' the percentage (in parts per thousand) of the specification (d) Replace new cards in Section 1.3. There must be six cards. Step 4. The step has to do with the length of simulated time. (a) In section 4.4 has four cards. The fourth card has START printed in Columns 8 - 1 2 . Remove this card. (b) Then Column '8'-'12' START Column '19'-'21' the desired simulated time in minutes. For example, i f the simulated time was to be 180 minutes, the card would read: START 180 Step 5. Compile the Complete Deck. Step 6. Place deck in card reader with appropriate command cards surrounding the deck. Step 7. The output wi l l be similar to that on the last page of Appendix III. This section wi l l contain the program statements used in the verification steps, Chapter III. A l l the statements or words on the left start in column '8 ' , and the statements on the right start in column '19.' One card per line. Only change the sections indicated. Verification Step 1. Section 1.1 should read: INITIAL XH1, 5 INITIAL XH2, 13 INITIAL XH3, 18 Section 1.2 should read: INITIAL XH7, 100 Section 1.3 should read: INITIAL XH30, 670 INITIAL XH31, 150 INITIAL XH32, 150 INITIAL XH33, 118 INITIAL XH34, 150 INITIAL XH35, 150 Section 1.4 should read: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 120 Verification Step 2. Section 1.1 should read: INITIAL XH1, 5 INITIAL XH2, 13 INITIAL XH3, 18 Section 1.2 should read: INITIAL XH7, 100 Section 1.3 should read the same as Verification Step 1. Section 4.4 should read: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 240,60 CLEAR XH1-XH3, XH7, XH30-XH35 START 240,60 CLEAR XH1-XH3, XH7, XH30-XH35 START 240,60 CLEAR XH30-XH35 INITIAL XHI, 4 INITIAL XH3, 17 INITIAL XH7, 50 START 240,60 CLEAR XH1-XH3, XH7, XH30-XH35 START 240,60 CLEAR XH1-XH3, XH7, XH30-XH35 START 240,60 Verification Step 3. Section 1.1 should read: INITIAL XHI, 4 INITIAL XH2, 13 INITIAL XH3, 17 Section 1.2 should read: INITIAL XH7, 64 Section 1.3 should read: INITIAL XH30, 890 INITIAL XH31, 120 INITIAL XH32, 120 INITIAL XH33, 118 INITIAL XH34, 130 INITIAL XH35, 100 Section 4.4 should read: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 180 CLEAR XH1-XH3, XH7, XH30-XH35 START 180 Section 1.1 should read for = 100 INITIAL XH1, 5 INITIAL XH2, 13 INITIAL XH3, 18 Section 1.2 should read: INITIAL XH7, 100 Section 1.3 should read the same as in Verification Step 1 and 2. Section 4.4 should read for a l l three runs: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 180 CLEAR XH1-XH3, XH7, XH30-XH35 START 180 Part (b): Sections 1.1 and 1.2 wi l l have the facility policy and asso-ciated arrival rate as determined in Verification Step 4. V. ANALYTICAL STEPS Step 1. Analysis of Service Policy Section 1.1 should read: INITIAL XH1, 4 INITIAL XH2, 14 INITIAL XH3, 17 Section 1.2 should read: INITIAL XH7, 40 Section 4.4 should read: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 180 CLEAR XH1-XH3, XH7, XH30-XH35 INITIAL XH7, 60 START 180 CLEAR XH1-XH3, XH7, XH30--XH35 START 180 CLEAR XH1-XH3, XH30, XH35 INITIAL XH7, 80 START 180 CLEAR XH1-XH3, XH7, XH30--XH35 START 180 CLEAR XH1-XH3, XH30-XH35 INITIAL XH7, 100 START 180 CLEAR XH1-XH3, XH7, XH30--XH35 START 180 CLEAR XH1-XH3, XH30-XH35 INITIAL XH7, 120 START 180 CLEAR XH1-XH3, XH7, XH30--XH35 START 180 CLEAR XH1-XH3, XH30-XH35 INITIAL XH7, 140 START 180 CLEAR XH1-XH3, XH7, XH30--XH35 START 180 CLEAR XH1-XH3, XH30-XH35 INITIAL XH7, 160 START 180 CLEAR XH1-XH3, XH7, XH30-•XH35 START 180 determine facility policy Sections 1.1, 1.2 and 1.3 must be specified by the user in the same format that has been presented up to now. The essence of Step 2 was to set an arrival rate \^ and search for a facility policy P^. Thus, set section 1.2 and vary section 1.1 until the service policy is achieved. Section 4.4 should read: GENERATE 60 SAVEVALUE 9+, 1, H TERMINATE 1 START 180 CLEAR XH1-XH3, XH7, XH30-XH35 START 180 Section 1.1 should read: INITIAL XH1, 5 INITIAL XH2, 13 INITIAL XH3, 18 Section 1.2 should read INITIAL XH7, 100 Section 1.23 - RPTS has a minor adjustment RPTB SELECTMIN 2, XH2, 16, V Columns '25 and '26 should have a '15' RPTB SELECTMIN 2, XH2, 15, V Section 3.33 - SPTS The card marked SPTB SELECTMIN 2, XH2, 16, V should be changed to: Column 2-5, SPTB Column 8-13 ASSIGN Column 19-22 2, 16 Section 4.4 should read the same as Step 1. Step 3. Part (b) Sections 1.1, 1.2, 1.3, 4.4 should be the same as Step 1. Section 3.23 - RPTS = REVENUE PASSENGER TICKET SYSTEM uses a single queue for a l l ticket counters, when an agent becomes free, a passenger unit wi l l leave the line and go to the agent that is free. RPTS Split P41, RPTA Mark 25 Substitute with following: RPTS Split P41, RPTB Transfer RPTB RPTB Mark 24 Queue 10 TEST L V19, V40 SELECTMIN 2, XH2, 15, F SEIZE P2 DEPART 10 MARK 25 Add in Section 2.22 19 Variable BV11 + BV12 + BV13 + BV14 + BV15 40 Variable 16 - XH2 Part (c) Sections 1.1, 1.2, 1.3 and 4.4 same as Part A. Remove Sections 3.31, 3.32 and 3.33. Substitute with following. (Starting columns are 2, 8 and 19). SPS SAVEVALUE 39+,P42 SAVEVALUE 40+,1 TRANSFER .*32,SPSA,SPSB SPSB ASSIGN 5,1 SAVEVALUE 33+,P42 SAVEVALUE 34+,l SPSA SELECTMIN 4,16,XH3,,V SPLIT P41,SPSC SPSC MARK 27 MARK 24 QUEUE P4 SEIZE P4 DEPART P4 MARK 28 MARK 25 TABULATE 3 TABULATE ASSIGN TEST G SAVEVALUE SAVEVALUE SPSD SAVEVALUE SAVEVALUE TEST E ASSIGN TEST G SAVEVALUE SAVEVALUE SPSF SAVEVALUE SAVEVALUE ADVANCE SPSE TRANSFER SPSH ASSIGN ADVANCE SAVEVALUE SAVEVALUE TEST G SAVEVALUE SAVEVALUE SPSI SAVEVALUE SAVEVALUE ADVANCE RELEASE TRANSFER SPSG ADVANCE RELEASE TRANSFER 3 28-,P27 28,150,SPSD 27+,P40,H 41+, 1 28+,P40,H 42+, 1 P5,1,SPSE 25-,P24 P25,150,SPSF 35+.P40 36+,1 37+.P40 38+, 1 1,FN4 •*35,SPSG,SPSH 7+,30 P7,FN1 17+,P40 43+,1 P25,150,SPSI 35+.P40 36+, 1 37+.P40 38+,1 1,FN5 P4 ,SEND P7,FN1 P4 , SEND APPENDIX V REPORT TO AIR CANADA The primary o b j e c t i v e of t h i s study i s to provide a compre-hensive management t o o l that w i l l a i d i n p o l i c y f o r m u l a t i o n and e v a l u a t i o n f o r the Passenger Check-in System. The A i r Canada Check-in System at the va r i o u s a i r p o r t s can be de f i n e d by the a r r i v a l of passengers and t h e i r c h a r a c t e r i s t i c s , and, the f a c i l i t y and s e r v i c e p o l i c i e s i n op e r a t i o n . L i s t e d below are the three performance o b j e c t i v e s of the Check-in System. 1. To o b t a i n a high u t i l i z a t i o n of check-in f a c i l i -t i e s and thus reduce the cose of s e r v i c e per passenger. 2. To achieve a minimum passenger w a i t i n g time i n queues and thus enhance customer r e l a t i o n s h i p s . 3 . To check-in the passengers i n accordance w i t h p r o c e d u r a l p o l i c y . In order that these three o b j e c t i v e s are f u l f i l l e d , s e v e r a l p o l i c i e s have been implemented i n the system. They are: 1. Separate the Standby passenger s e r v i c e (baggage check-in and t i c k e t purchase) from that of the Revenue Passenger. Since Revenue passengers comprise the l a r g e s t percentage of t r a v e l l e r s and pay full-f a r e for their tickets, they are the most important. If they become dissatis-fied with service, hopefully i t w i l l not stem from waiting in queues while a Standby passen-ger is being served. 2. Staff the f a c i l i t i e s such that 85% of the Revenue and 75% of the Standby passenger units (a passenger unit i s equivalent to a single person travelling in a group or a family unit) do not exceed 2^ minutes waiting time i n any one queue. 3. Use a 15-minute staff horizon to effectively and eff i c i e n t l y allocate manpower in the system. This policy has been implemented because the demand for service (the arrival rate) fluctuates great-ly during the day and the mobility of agents to other functions in the airport. The effectiveness of a policy may be measured by: 1. The average u t i l i z a t i o n of the various check-in f a c i l i t i e s (baggage and ticket counters). 2. The percentage of passenger units who exceed 2% minutes of waiting time in queues. 3. The average transit time per passenger unit (the total time spent in waiting in queues and being checked-in at the counters). The aforementioned tool that wi l l aid in policy formulation and evaluation is a computer simulation model. The Check-in system at the Vancouver International Airport was studied as to its sta-t is t ical and operational characteristics, then modelled. The tool was verified that i t did in fact simulate the real system. In order to evaluate a policy, the user of the model simply specifies certain starting conditions, they are: 1. The number of passengers arriving in the system per 15 minute period. 2. The Passenger Composition: Revenue Passengers: % of total passengers % using Curbside Check-in % arriving non-ticketed % having excess baggage Standby Passengers: % arriving non-ticketed % having excess baggage 3. The Facility Policy in Operation The number of Revenue baggage and ticket counters open. The number of Standby baggage and ticket counters open. What the model wi l l do is define the beahviour of the system using the three previously mentioned statistics for the specified conditions. To demonstrate the capabilities of this model, an analysis was undertaken in three areas of policy management. The areas are: 1. To determine the implications of the behaviour of the Passenger Check-in System on policy for-mulation. This analysis, w i l l , in effect, de-termine i f the present service policy is formu-lated correctly. 2. To determine the facility policies that should be implemented in order to achieve the present service policy. As part of this analysis the maximum capacity of the system wil l be deter-mined. 3. To formulate alternative operating policies and test for viability prior to implementation. The results of the analysis are presented sequentially. 1. Service Policy Evaluation The present service policy for the Revenue passengers is as follows: to allow no more than 15% of the Revenue passenger units to wait in any one queue longer than 2% minutes. The main elements which determine the state pf the system are the arrival rate and the facility policy. If the facil ity policy (4 Revenue baggage counters and 2 Revenue ticket counters) is held constant, then the effect of the arrival rate on the subsequent behaviour of the system may be determined. Figure 1 depicts the effect of the arrival rate on the average transit time per Revenue passenger unit. When the stated facility policy is in operation and arrival rate is in the range of 40 to 90 passengers per 15 minutes, the average transit time is between 2 and 3 minutes. The system is thus relatively stable or insensitive to the number of passengers in the system. When the arrival rate is greater than 90 passengers the curve becomes unstable and is sensitive to the number of passengers in the system. For example, when the arrival rate of 100 goes to 120 passen-gers per 15 minute period, the average transit time goes from 3.3 to 6.5 minutes per Revenue passenger unit. This is a 20 per cent in-crease in volume and results in 100 per cent increase in the average transit time per passenger unit. Thus, i f a curve was derived for each facil ity policy at the various arrival rates, the vertical portion of the curve would indicate the range of arrival rates such that the facility policy produces stability in the system. ARRIVAL RATE 160 PASSENGERS PER 15 MIN. EFFECT OF VARIOUS ARRIVAL RATES ON THE AVERAGE TRANSIT TIME NO. OF FACILITIES OPEN HELD CONSTANT INSEN-SITIVE RANGE 140 120 100 80 4 REVENUE BAGGAGE COUNTERS 60 ] 40 — i i .i 2 4 6 INSENSITIVE RANGE 8 10 12 AVERAGE TRANSIT TIME IN MINUTES PER REVENUE PASSENGER UNIT Figure 2 is a graph of the average utilization of the Revenue baggage counters as the arrival rate varies. Note that the curve is almost linear in the range of 40 to 120 passengers per 15 minutes. When the arrival rate increases past 120, the curve becomes vertical in nature. At a very high arrival rate, average utilization of the Revenue Baggage Counters would approach 100%. Also from the previous figure, there would be an extremely long transit time. Thus, the average utilization for the Revenue baggage counters wi l l be in the range of 30% to 65% when the arrival rate is in the range of 40 to 90 passengers and 4 counters open. Figure 3 is a plot of the arrival rate versus the percen-tage of Revenue passenger units waiting longer than 2.5 minutes at the baggage counter. This curve has a vertical and horizon-tal section. When the arrival rate is in the range of 40 to 100 passengers, the Revenue baggage counter is able to process the passengers at such a rate that the 15 per cent rule is not vio-lated. In the range above 100 passengers the curve becomes horizontal in nature. The changeover occurs when approximately 13% to 15% of the passenger units exceed 2^  minutes of waiting time. ARRIVAL RATE (PASSEN-GERS PER 15 MIN.) 160 140 120 100 80 60 EFFECT OF VARIOUS ARRIVAL RATES ON THE REVENUE BAGGAGE COUNTER UTILIZATION FACILITY POLICY CONSTANT 40 10 20 30 40 50 60 70 80 90 100 % AVERAGE UTILIZATION REVENUE BAGGAGE COUNTERS ANNUAL 160 RATE (PASSEN-EFFECT OF VARIOUS ARRIVAL RATES ON THE % OF REVENUE PASSENGER UNITS EXCEEDING 2% MINUTES WAITING TIME IN A QUEUE GERS PER 15 MIN.) 140 • 120 -100 . 80 -60 PRESENT SERVICE POLICY 15% 40 1 i i 0 10 15 20 30 40 50 60 70 80 % REVENUE PASSENGER UNITS EXCEEDING 2h MINUTES WAITING TIME IN QUEUES This curve then justifies the current service policy. If the percentage is increased, the system becomes very sensitive to the amount of passengers in the system for the facility policy (at arrival rate of 120 passengers, 49% of Revenue passenger units exceed 2.5 minutes). The vertical portion of the curve represents the range within which the facility policy produces stability in the system. These results are consistent with the other graphs. In Figure 4 the utilization and corresponding percentage of passenger units exceeding 2.5 minutes waiting time has been graphed (Figure 2 and Figure 3). For policy formulation, the effect of increasing the per-centage of passenger units allowed to exceed 2.5 minutes waiting time on the facility utilization may be determined from the graph. Alternatively, this gives the trade-off of opposing system objectives. If the percentage was increased to 25%, the average utilization that could be expected would be 75%. 2. Facility Policy Determination The second area of the analysis is to determine how the model is able to formulate an optimal facil ity policy for a given set of conditions. The given set was: AVERAGE UTILIZATION 100%. 90 ' 80 AVERAGE UTILIZATION AT THE REVENUE BAGGAGE COUNTERS VERSUS THE % OF REVENUE PASSENGER UNITS EXCEEDING 2h MINUTES WAITING TIME IN QUEUES FACILITY POLICY HELD CONSTANT 70 60 50 40 30 20 10 SERVICE POLICY 15% REVENUE PASSENGER UNITS ALLOWED TO EXCEED 2\ MINUTES 50 60 70 80 % REVENUE PASSENGER UNITS EXCEEDING 2% MINUTES WAITING IN QUEUES 2. The a r r i v a l rates are i n the range of 25 passengers to the yet undetermined maximum capacity of the system. 3. The Passenger composition was Revenue passengers 85% of t o t a l passenger a r r i v a l s 15% of a r r i v e d non-ticketed (group) 10% used curbside check-in (group) 15% had excess baggage (passenger units) Standby passengers 15% a r r i v e d non-ticketed (groups) 15% had excess baggage (passenger units) The r e s u l t of t h i s analysis appear i n Table 1 which r e f l e c t s the f a c i l i t y p o l i c y for various a r r i v a l rates. The maximum capacity of the i n d i v i d u a l counters i s also i n d i c a t e d . 3. Operating P o l i c y Formulation and Evaluation The analysis undertaken here i s di r e c t e d towards the operating p o l i c i e s of A i r Canada f o r the t i c k e t counters, both Revenue and Standby A few years ago, the operating p o l i c y was changed such that baggage operations and t i c k e t purchases of the Revenue passengers was separated from that of the Standby passengers. Revenue passengers FACILITY POLICY FOR THE VARIOUS ARRIVAL RATES ARRIVAL RATE TOTAL PASSENGERS PER 15 MIN. PERIOD REVENUE COUNTERS BAGGAGE TICKET STANDBY BAGGAGE , TICKET MANPOWER REQUIREMENTS TOTAL 25 50 75 100 125 150 175 200 2 4 4 5 6 7 8 9 CAPACITY REACHED 1 2 2 3 4 5 5 5 CAPACITY. REACHED CAPACITY REACHED 1 1 2 2 2 2 2 PASSENGER COMPOSITION 85% OF TOTAL ARE REVENUE 10% OF REVENUE USED CURBSIDE 15% HAD EXCESS BAGGAGE 15% WERE NON-TICKETED ON ARRIVAL CAPACITY REACHED 1 1 1 1 1 1 1 1 9 11 13 15 16 17 paying excess charges were allowed to use the Standby ticket counter and Standby passengers under the same circumstances were allowed to use the Revenue ticket counters. Part A of this analysis, focuses on the impact of com-plete separation of Revenue and Standby ticket operations. The implementation of this policy should produce a higher uti l iza-tion of facilities and greater percentage of passenger units waiting longer than 2.5 minutes in queues. This is so because the Revenue passengers are limited to one fewer counters for service. The simulation model produced higher system statistics than the integrated policy was operational (see Table 2). For example the average utilization of the Revenue ticket facil it ies increased from 28.3% to 45.5%. If the decision is made that com-plete separation is necessary, the facility policies as stated in Table 1 should produce stability in the system for the Revenue system. In Part B of this analysis we examine the operating policy at the Revenue ticket counter more closely. Since the separation as in Part A produced higher system statistics, the queuing disci-pline wi l l be changed to determine i f the percentage of Passenger units exceeding 2\ minutes waiting time can be lowered while main-taining separated operations. REVENUE Tr a n s i t Time per passenger Unit Average U t i l i z a t i o n % of Passenger Units waiting longer than 2.5 minutes INTEGRATED 2.6 Min. 28.2% 6.4% SEPARATED 2.85 Min. 49.9% 13.8% STANDBY T r a n s i t Time Average u t i l i z a t i o n % of Passenger Units waiting longer than 2.5 minutes 2.9 Min. 49.2% 30%* 3.75 Min. 62.1% 54.9%* *25% + 5% would be acceptable but since there i s only one Standby t i c k e t f a c i l i t y , a de c i s i o n must be made to e i t h e r expand or rev i s e the service p o l i c y at the Standby t i c k e t counter. S p e c i f i c a l l y , instead of using multiple queues or one queue f o r each agent, a s i n g l e queue w i l l be used. A l l Revenue passengers purchasing t i c k e t s or paying excess charges l i n e up i n one queue. When they reach the front of the l i n e , they wait u n t i l any one of the agents becomes a v a i l a b l e . Only then do they leave the l i n e . This i s depicted i n Figure 5. The r e s u l t s of the simulations com-bined with the r e s u l t s of Part A for comparative purposes are i n Table 3. COMPARATIVE STATISTICS FOR REVENUE TICKET COUNTER UNDER VARIOUS OPERATING POLICIES STATISTIC STANDBY EXCESS REVENUE STANDBY ALLOWED SEPARATED (INTEGRATED) MULTI-QUEUES SINGLE QUEUES AVERAGE TRANSIT TIME PER REVENUE PASSENGER UNIT AVERAGE UTILIZATION (%) % PASSENGER UNITS EXCEEDING 2h MINS. WAITING TIME IN QUEUES OF TICKET COUNTER 2.6 MIN. 48.2% 6.4% 2.85 MIN. 49.9% 13.8% 2.75 MIN. 42.9% 4.5% FIGURE 5 SINGLE OR MULTIPLE QUEUES AT THE REVENUE TICKET COUNTER CM CO . <^QUEUE . ^/QUEUE • <^QUEUE AGENT 1 AGENT 2 AGENT 3 REVENUE TICKET COUNTER MULTIPLE QUEUE OPERATION SINGLE QUEUE OPERATION GOES TO ANY AGENT \ V MOVEABLE GUIDE RAILS AGENT AGENT AGENT 1 2 3 REVENUE TICKET COUNTER The transit time per passenger unit is lower and is com-parable to the integrated operation. The utilization is lower as well as the percentage of Revenue passenger units exceeding 2.5 minutes of waiting time than the integrated. Thus i f this par-ticular method of queuing was adopted, separation of the Revenue/ Standby counter could be achieved with viable results. In Part C of this step in the analysis we examine the method of checking in Standby passengers. Since there has been a large number of passengers exceeding the 2% minutes of waiting time, the operating policy of the ticket and baggage operations at one counter wi l l be analyzed. In effect, one agent would become both a ticket and baggage agent. When the passenger unit approaches the service counter, a single agent wil l issue tickets, check the baggage and collect the excess charges. Since this question in-volves multiple operations and is not at present in effect the cor-rect distribution of service times must be hypothesized. The esti-mated total service time wil l probably be high which in this case is conservative since we are testing a completely new agent proce-dure and want to be very careful. The flowchart of the operation for a single agent is depicted in Figure 6. The system statistics are given in Table 4. YES NO PURCHASE TICKET CHECK IN BAGGAGE PAY EXCESS CHARGE ONE STANDBY AGENT PERFORMS BOTH BAGGAGE AND ' TICKET OPERATIONS TABLE 4 VARIOUS OPERATING POLICIES AT STANDBY COUNTERS STATISTIC INTEGRATED REVENUE/STANDBY SEPARATED TWO BAGGAGE AGENTS TWO BAGGAGE AGENTS THREE AGENTS EACH ONE TICKET AGENT ONE TICKET AGENT PERFORMS BOTH TICKET AND BAGGAGE OPERATION TRANSIT TIME PER STANDBY PASSENGER (MIN. 2.9 Min. 3.75 Min. 3.0 Min. BAGGAGE COUNTER AVERAGE UTILIZATION (%) 31.2% 30.3% ) % PASSENGERS EXCEEDING ) 2h MIN. WAITING TIME IN ( COMBINED UTILI-QUEUES (%) 6.0% 5.8% ) ZATION 44.7% ( % PASSENGERS TICKET COUNTER ) EXCEEDING 2% ( MIN. 28.3% AVERAGE UTILIZATION (%) 49.9% 62.1% ) % PASSENGERS EXCEEDING ) 2h MIN. WAITING TIME IN ( QUEUES 30.0% 54.9% ) If this operational procedure is adopted, the transit time (average) of Standby passengers remains relatively the same as when the Revenue/Standby ticket operations were integrated. However, the percentage of passengers who exceed 2.5 minutes waiting time is reduced to 28.3%. Thus the service policy is maintained through the total system at an arrival rate of 100 passengers and the aforementioned parameter specifications. These are just a few of the experiments that can be per-formed on the model. The experiments presented hopefully have hit on the major areas of interest and demonstrated the vast power of simulation as an airport management tool. A sample of the output from the computer model is given in Figure 7. This may be varied depending upon the requirements of the user. FIGURE 7 OUTPUT FROM COMPUTER THE MEAN ARRIVAL RATE WAS 100 PASSENGERS PER 15 MINUTE PERIOD. REVENUE PASSENGER SYSTEM TRANSIT TIME IN THE SYSTEM AVERAGE TIME PER PASSENGER UNIT 2.7 MIN. AVERAGE TIME PER PASSENGER 2.9 MIN. REVENUE PASSENGER BAGGAGE SYSTEM NUMBER OF COUNTERS OPEN 5. AVERAGE UTILIZATION OF FACILITIES 60.1% AVERAGE WAITING TIME IN QUEUES 25 SEC. PER PASSENGER UNIT % OF PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 5.6% REVENUE PASSENGER TICKET SYSTEM NUMBER OF COUNTERS OPEN 3. AVERAGE UTILIZATION OF FACILITIES 50.3% (COUNTERS 11-15 ONLY) AVERAGE WAITING TIME IN QUEUES 31 SEC. PER PASSENGER UNIT (11 % OF PASSENGER UNITS IN ALL TICKET OR EXCESS QUEUES OVER 2.5 MIN. WAS 9.5% STANDBY PASSENGER SYSTEM TRANSIT TIME IN THE SYSTEM AVERAGE TIME PER PASSENGER UNIT 3.9 MIN. AVERAGE TIME PER PASSENGER STANDBY PASSENGER BAGGAGE SYSTEM NUMBER OF COUNTERS OPEN 2 AVERAGE UTILIZATION OF FACILITIES 38.6% AVERAGE WAITING TIME IN QUEUES 24 SEC. PER PASSENGER UNIT % PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 6.8% STANDBY PASSENGER TICKET SYSTEM AVERAGE UTILIZATION OF FACILITIES 70.8% AVERAGE WAITING TIME IN QUEUES 307 SEC. PER PASSENGER UNIT % PASSENGER UNITS IN QUEUES OVER 2.5 MIN. WAS 61.1% 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101735/manifest

Comment

Related Items