基于起讫点的均衡交通分配改进算法

AnthonyChen (2002)提出的基于起讫点的Frank-Wolfe (ODBFW)算法, 对同一起点的所有OD对搜索最短路树, 采用one-OD-at-a-time流量更新策略, 是一种收敛性能较好的均衡交通分配算法. 主要对ODBFW算法的初始化、线性搜索及最短路搜索算法进行改进, 提出ODBFW改进算法. 将ODBFW改进算法应用于国际上同类研究常用的SiouxFalls网络及我国山东省诸城市实际道路网, 比较标准FW算法、ODBFW算法及ODBFW改进算法的可行性、效率及稳定性. 结果显示, 改进算法效率较标准FW算法有显著提高, 稳定性较好. 在小型测试路网上改进算法较ODBFW算法达到收敛的时间减少近15%, 在大型测试路网上减少近5%.

Improved origin-destination-based algorithm for equilibrium traffic assignment

Based on the origin-destination-based Frank-Wolfe algorithm proposed by Anthony Chen (2002) which means the shortest-path of all OD pairs with the same origin was searched in the algorithm at one time and the link flow was updated by using one-OD-at-a-time flow update strategy,the initialization, linear search and shortest path search of ODBFW algorithm was improved in this paper,and the improved ODBFW algorithm was proposed.The improved ODBFW algorithm was used for Sioux Falls network which was common used in international similar researches and the road network of Zhucheng City,Shandong Province,China to compare the feasibility,efficiency and stability of the standard FW algorithm,ODBFW algorithm and improved ODBFW algorithm.The results indicated that the improved ODBFW algorithm has more efficiency and stability than standard FW algorithm.The time used for the improved ODBFW algorithm to reach convergence is nearly 15%less than the time used for the original algorithm in small road network,and is nearly 5%less than that in large road network.