Linking events-based communication mechanism in multi-agent systems

Multi-agent systems (MAS) consist of several independent agents with decision-making ability. Each agent has the ability of communication and perception. It can adjust the running state through messages interaction for achieving global optimization. The accuracy and efficiency of messages transfer among agents has a decisive impact on the performance of the MAS due to the complexity between problems. In this paper, the transfer of the messages in large distributed MAS is studied and the method of messages transfer through linking events is proposed. Firstly, the characteristics of messages in MAS are analyzed and the information model based on MAS is defined. Then, the messages are formatted and classified and the relationship between different events is established. After that, the messages can be transmitted according to the relationship between events. The proposed messages transfer method through linking events is validated by applying this method in an industry information system. The results show a efficient transmission of messages between agents. Furthermore, it can extract associated events from mess messages, and provides feedback of the running state of MAS.