基于多智能体博弈的路径选择策略仿真研究  

Route Guidance Strategy Simulation Based on Multi-agent Game Approach

作　　者：安实[1] 崔娜[1] 李静[1] 

机构地区：[1]哈尔滨工业大学,哈尔滨150090

出　　处：《交通信息与安全》2009年第3期1-5,共5页Journal of Transport Information and Safety

基　　金：国家自然科学基金项目(批准号:70673016;70503008);哈尔滨工业大学优秀青年教师培养计划项目(批准号:HIT;2006;19)资助

摘　　要：系统最优与用户最优代表了交通分配中路网管理者与出行者两种不同的利益出发点。在综合考虑两者在路径选择过程中动态交互特点的基础上,引入博弈论的思想协调两者的利益冲突,建立了路网系统管理者与出行者之间的路径选择博弈模型。为验证模型的有效性,结合多智能体技术进行了相应的仿真实验,并利用多智能体仿真软件Starlogo进行模拟。通过对无信息无博弈(随机)出行、用户最优出行、有信息有博弈出行以及系统最优出行等4种不同仿真方案的比较分析,验证了系统与出行个体之间协调的博弈模型性能满足了驾驶员出行需求,提高了路网整体效率,为建立实用的诱导策略提供了参考。During the process of traffic assignment, user equilibrium （UE） and system optimum （SO） are on behalf of the different benefits of network manager and the travelers. On the basis of comprehensively considering the dynamic interactive characteristics of UE and SO, game theory was applied to the coordination of the two incompatible ideas, and a route choice game model between the network manager and the traveler was presented. In order to validate the feasibility of the model, the multi-agent technology was combined to design corresponding simulation experiments. By using Starlogo, which is a famous multi-agent simulation software, four different simulation scenarios were simulated and compared including non-information without game travel, UE travel, UE-SO coordination travel, and SO travel. Through the analysis of the experiment results, the effectiveness and superiority of the coordination model are demonstrated, which provides reference for the practical guidance strategy.

关 键 词：交通流诱导 路径选择 博弈论 多智能体 

分 类 号：U491.54[交通运输工程—交通运输规划与管理]