Modeling & Simulation of Complex Adaptive System——A Traffic Simulation System

Traffic system is a complex adaptive system (CAS). Intuited by CAS theory, we present a decentralized multi-agent based dynamic traffic simulation system with SWARM platform. We develop the traffic systems from bottom up, where vehicles and traffic signals are defined as agents. The interaction of these individual agents are simulated on a road map, thus leading to emergent patterns of traffic dynamics. Cellular Automaton is used to simulate the traffic flow and an evolutionary design is applied to self-organized control of the traffic signals. Furthermore, we compare the control approach if we join the forecast mechanism with the pure self-organized control method. From the experimental results, it can be concluded that the performance of the former approach is much superior to the latter one in that it can reduce the traffic jam.