Research on agent-based G-Net train group operation model

In order to satisfy the requirements of the railway transportation heading for intelligence, we research the modeling method of agent-based G-Net, construct the model of train group operation, analyze the conceptions and realization techniques of the model and give the structures and cooperative relations of the main agents. Compared with the existed model of train group operation system with object-oriented Petri subnet (TGOSOPS), the intelligence of the system is emphasized on many layers. Compared with the model of train operation Petri net with agents (TOPNA), we further improve the synchronization and asynchrony among the agents, divide the agents more definitely and give prominence to the principal agents of railway intelligent transportation system: trains and stations. Because the model has such characteristics as owning clear structure, fine extensibility and designing distributed program easily, the simulation system can be designed more reliably and effectively.