Abstract: In order to solve the problems existing in the current transit system simulation model, such as complexity to build an analytical model and difficulty to data collection, a simulation model of bus rapid transit system based on agent paradigm is proposed. The passengers, buses and signal lights are abstracted as different agents. Vehicle scheduling agent and station management agent are also introduced. The situation of bus rapid transit system is simulated through the interaction among agents. In order to solve soaring passengers’ problem in some stations, this paper proposes a vehicle scheduling algorithm based on dynamic adjustment to reduce the waiting time of passengers. The experiment and analysis show that the system not only can simulate the situation of bus rapid transit system actually, but also can supply references to utilize bus resources reasonably.

Key words: agent based simulation, bus rapid transit system, vehicle scheduling algorithm

摘要： 针对目前公交系统仿真模型中存在的建模复杂、数据搜集困难等问题, 构建了一个基于agent的快速公交系统仿真模型。将乘客、公交车、信号灯抽象为不同的agent, 并引入车辆调度agent和站台管理agent, 通过各agent之间的交互来仿真快速公交系统的运行。为了解决个别站点可能出现的“涌现”现象, 模型中给出了一种基于动态调整的车辆调度算法, 从而减少了乘客的等待时间。实验和分析表明, 该模型建模简单, 易于理解, 不仅可以真实地模拟快速公交系统运行现象, 而且对如何充分合理利用公交车资源有一定的借鉴作用。

关键词: 智能代理仿真, 快速公交系统, 车辆调度算法