PETRI NETS BASED MODELING AND SIMULATION OF COMPLEX ADAPTIVE SYSTEMS

摘要

The world is replete with all kinds of complex systems, and complex adaptive system models provide a new way for understanding the complex. In this paper, we review the past research approaches, point out the inefficiency of the conventional modeling methods in describing complex adaptive systems(CAS), and emphasize the needs for a new approach. Then we highlight five class models for studying CAS from the literatures, including network model, Evolutionary computing, Individual-based model, agent-based model and Echo model, and discuss their different characters. This paper introduces a new advanced Petri Nets model, which takes advantage of the theory of complex adaptive system and Petri Nets method. In this method, the combination of classifier systems and Petri Nets offers a distinctly useful approach to modeling complex adaptive systems. Later we try to model a kind of complex adaptive system, namely supply chain networks, using adaptive agents that represent individual entities each with its own rules and describe how it should interact with its environment. Finally through simulating with Swarm software platform, we simulate the simplified supply chain model, and analyze the results of simulation.