分布式并行计算在交通网络仿真中的应用*

根据交通网络仿真的并行特征采用域分解方法设计交通并行仿真系统的框架,把交通网络分为几个子网,集群系统的每个节点机分别负责其中的一个子网,提出基于车辆数负载的网络分割算法来平衡各子网的负载量,并分析子网之间的通信机理.同时,在基于MPI 的并行计算平台上实现设计的并行仿真系统.通过实例表明,提出的并行算法能大大提高交通网络仿真的速度和效率.

Application of distributed and parallel computing in traffic network simulation

The paper designed framework of parallel traffic simulation system using domain decomposition method according to parallel characteristics of traffic network simulation, which meant that traffic network was partitioned to several sub-network, and each processor of clusters was responsible for a different sub-network of the simulated region. Used a network-partition al- gorithm based on the number of vehicle to optimize load balancing. Described information exchange mechanism between subnetworks, the developed parallel simulation system in parallel computing platform based on message passing interface. From the results of numerical example, the proposed parallel simulation algorithm can improve speed and efficiency of traffic network simulation.