Test and implement of a parallel shortest path calculation system for traffic network

Abstract

This paper describes the parallel implementation of a traffic network model. A two-queue parallel shortest path algorithm is employed using recursive spectral bisection decomposition approach where each processor runs the same program but acts on a different subset of road network. The objective is to reduce the execution time of shortest path computing in dynamic traffic assignment. The model is parallelized and tested by 1, 2, 4, 8, 16 and 32 processors. The performances of the parallel model are discussed, and we think that two-queue parallel shortest path algorithm and recursevive spectral bisection approach are useful in solving shortest path problem of traffic network. © Springer-Verlag 2010.