International Construction Specialty Conference of the Canadian Society for Civil Engineering (ICSC) (5th : 2015)

A routing algorithm to construct candidate workzones with distance constraints Eicher, Charel; Lethanh, Nam; Adey, Bryan T.


As highways deteriorate over time, it is necessary to execute preventive interventions to ensure that they continue to provide an adequate level of service. As the execution of interventions on highways almost invariably results in the interruption to traffic flow, it is often beneficial to group interventions. By grouping interventions into workzones, there is, for example, less lane changing required by vehicles traveling on the highway and, therefore, perhaps fewer accidents. The objects included in the optimal workzones depend on many factors, such as the condition/performance of the objects, the length of the workzone, the traffic configuration within the workzone, the length of time required to execute the interventions, and the budget available. Recent research by Hajdin & Lindenmann (2007) and Lethanh et al. (2014) has been focused on the development of optimization models to solve such problems. One difficulty with them though is the construction of the set of possible combinations, which was done manually. Once large networks are to be analyzed this is no longer possible. In this paper, a routing algorithm is presented that can be used together with these optimization models to automatically establish the combination matrix, taking into consideration constraints on the length of the workzone and the distance between workzones. The algorithm is developed in Matlab and empirically tested on a real world road network, with 671 km of roads and 567 objects including bridges, tunnels, and road sections. The state of each object is classified on a discrete scale of 5, with 1 being the best and 5 being the worst. Several scenarios based on setting constraints on maximum workzone length, and minimum distance between two adjacent workzones are used to verify the robustness of the algorithm. It is found that the algorithm is both efficient and fast for all scenarios investigated. The development potential, in particular with respect to integration in GISs, is discussed.

Item Citations and Data


Attribution-NonCommercial-NoDerivs 2.5 Canada