A coloured Petri Net based approach for flexible agent interactions

As one of the most important artificial intelligence techniques, multi-agent systems are widely applied in many fields. In a multi-agent system, agent interactions are established through exchanging messages by following some interaction protocols. However, as the application domains of MASs are getting more and more complex, some limitations of current multi-agent interaction mechanisms are rising up. To cover some limitations of current agent interaction mechanisms, in this paper, we propose a Coloured Petri Net based approach to enable flexible agent interactions in open environments. This approach adopts Coloured Petri Net techniques to model agent interaction protocols, and separates interaction protocols and agents into two different layers. Through this way, agents can be more independent from a particular multi-agent system. Also, the flexibility and extensibility of multi-agent systems are enhanced.