| Summary |
Shu-Fang Yang, 2006.06
Graduate Institute of Transportation Science, Tamkang University
With vanpooling emerging as a viable potential policy for reducing private car in congested area, the vigorous study on the vanpooling problem is critical for its success. However, in the current practice, the supervisors of the vanpool program usually focus on the assignment of the patronages to the vans, but not on the routes taken. In practice, the routes are determined by driver’s experiences. With the rapid development in communication technology, we should be able to perform the van pooling program more effectively with these techniques.
To deal with the vanpooling program in this paper, we assume the customer appearance is uncertain, and the basic principle of the vanpooling’s operation is pick-up first drop-off second. With the constraints of the customer’s time and distance, and the maximum and minimum van capacity, the purpose of this study is to determine a set of m minimum travel time vehicle routes capable of accommodating as many users as possible, under these constraints.
A static solution procedure is developed in this study with preset demand as core, and some real time demand will be served too. A two-phase solution procedure is adopted in the study. In the first phase, route structure is determined based on the preset demand. A revised ant colony optimization (ACO) and the meta-heuristic of combining the threshold accepting and noising method (TA&NM) are used to construct the route structure. For the real time demands, based on their locations and the existing set of routes, a threshold accepting method (TA) is adopted to adjust the current set of routes to serve these demands in the second phase. With the incorporation of these meta-heuristics, the proposed solution procedure should provide a robust route structure and patronage assignment for the vanpooling program. .
A series of case studies with different characteristics such as demand density, demand size were used to test the solution capability of the proposed algorithms. Several versions of revised algorithms were developed in these studies. Based on the result of the case studies, an ACO was identified as the most suitable solution procedure for the problem addressed in this study. The proposed procedure can be used to develop a suitable vanpool problem with real-time demand information. |
