Study on train operation adjustment system based on multiagent

The traditional systems for train operation adjustment are based on expert system, fuzzy logic and Petri-network. In this paper, these systems were introduced and their advantages were discussed. In order to implement the real time processing of unexpected events in train operation, a system based on multi-agent approach is put forward. This distributed system is composed of three types of agent: train agent, station agent and control center agent. The structure and parts of the system are described including their advantages.