Open Collections

Inventory management at Finning Ltd.

University of British Columbia

This thesis deals with two separate but related projects completed by the author, with the assistance of those acknowledged later, for Finning Ltd. The goals of these projects were to improve customer service and minimize distribution costs in Finning's Caterpillar parts distribution activities by creating more effective and efficient methods of inventory control and demand forecasting. The first project, a parts inventory & distribution model, consists of two components; a simulation and an analytical model which work together to determine target inventory levels for each part at each Finning branch. The thesis deals specifically with a sub-component of the simulation - a transportation subroutine which minimizes the time required to transport a part to a branch which stocks out of the part from a branch where it is available. The algorithm developed for the subroutine resembles a classic Shortest Path algorithm, with some important modifications. The most important of these is that the transportation algorithm is entirely dependant on time; it seeks to minimize total elapsed time rather than total travel time. To account for this difference, the problem was approached through the use of a time-oriented graph rather than a physical graph. With this approach in mind, the graph was constructed and an algorithm developed which ensures that the optimal transportation route is selected. The purpose of the second project was to improve the accuracy of demand forecasts for Caterpillar parts at the various Finning branches. Finning's forecasting requirements indicated the need for an automatic time series forecasting package. It was determined that AUTOBOX 3.0, a statistical software package which automatically generates time series forecasts using the ARIMA method, could potentially meet Finning's needs in this area. An evaluation of AUTOBOX's forecasting techniques was undertaken, with emphasis on several specific options: Remodelling; Robust Autocorrelation; Holt-Winters Identification; Outlier Detection; and Convert Pulse-to-Step. Following the evaluation of these options, recommendations are made regarding the appropriate use of AUTOBOX by Finning, including a caution against relying blindly on forecasts made by any automatic procedure, without considering the peculiarities of particular datasets. At date of writing, these two projects have received varying levels of attention from Finning's Operations Research team. The primary commitment of Finning is to the development of the simulation and the implementation of its recommendations. Finning estimates its savings through use of the simulation to date at several hundred thousand dollars. It is estimated that this saving will increase to more than one million dollars when all planned changes are implemented. The demand forecasting evaluation, in contrast, remains to be used at some future date. The completion of these projects required that specific quantitative data be provided to the author by Finning with respect to its customer demand and operations. Due to the confidentiality of this information, specific quantitative data is not included in this thesis.

Thesis/Dissertation

application/pdf

2009-01-12

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/3592

Business Administration

University of British Columbia

UBCV

DSpace