中文摘要：

关键交通基础设施（Critical Transportation Infrastmcture，CTI）动态仿真中，车辆Agent的行为是涌现CTI宏观特性的关键；而车辆Agent的实时路径搜索与选择算法是车辆行为的核心和难点之一。为解决CTI中车辆Agent的实时路径搜索与选择算法问题，分析了主要的最佳路径搜索算法。从降低算法的复杂度入手，提出了改进的Floyd算法，进行了算法的复杂度分析，仿真试验结果表明了算法的实用性和高效性。

英文摘要：

In Critical Transportation Infrastructure （CTI） dynamic simulation, vehicle Agent behavior is a key factor of emerging macro-feature in CTI. Real-time best path search algorithm is one of the nuts and difficulties of vehicle Agent＇s behavior. In order to solve real-time vehicle Agent best path search algorithm, major best path search algorithms such as Floyd, Dijkstra, and A＊ were briefly analyzed. An improved Floyd algorithm was put forward through reducing the complexity of algorithm. The complexity of improved Floyd algorithm was also analyzed. Simulation results show the effectiveness and practicality of the algorithm.