Abstract: This paper aims at solving the low efficiency when the large-scale traffic network is simulated by single CPU. We built a traffic simulation cluster system with MPI/OOMPI parallel library. In the cluster, each node processes a sub-network, which is obtained by the partition of a large-scale traffic network, depending on orthogonal recursive bisection method. In this way, the computational loads are transferred to each node effectively. From the result of experiment it is concluded that the excellent speed-up of large-scale network simulation can be achieved.

Key words: cluster, traffic, MPI/OOMPI, parallel, simulation