A PSO Methahuristic for Origin-Destination Matrix Estimation Problem

Information on the origin-destination (OD) demands through a transportation system is an important input in many areas of operational analysis and planning. The high cost of manpower associated with surveying methods of obtaining OD matrices have motivated the development of models and methods that can estimate these matrices from easily available traffic flows. In the last years the OD matrix estimation problem (ODMEP) based on traffic counts has been formulated as bilevel programming models. Although the conventional methods of optimization can theoretically reach to the exact solution of the problem, they have some computational difficulties that prevent them from being applicable in real-case problems with too many O-D pairs. In this paper, to cope with the limits of the conventional methods, a PSO metaheuristic is employed to solve the ODMEP. Computational results on a sample networks are presented to evaluate the PSO method against the widely used optimization method of gradient

Keywords:

Transportation , O-D Matrix Estimation , Metaheuristics , Particle Swarm