Multi-Agents Model and Simulation for the Evolution of Industrial Clusters

By using multi-agents system model simulation method, this paper tries to reveal the key driving force for the evolution of industrial clusters under the dynamic development of the external environment and internal innovation. Also, how important role of policy recommendations is proved for the regional industrial development. For the purpose of studying the evolution of industrial clusters, a multi-agent system model is constructed and the model's learning algorithm addressed on genetic algorithm. First, industrial clusters are formed as a conceptual system which corresponds to a virtual multi-agent system and the basic genetic algorithm is employed as an agent's intelligent learning algorithm. Then, simulation results are carried out by conducting the learning algorithm on Matlab7.0 to simulate the evolving behaviors of the multi-agent system. By mapping the corresponding simulation results back to the conceptual system, the evolving rules of the industrial clusters are revealed thereafter. The study by this method shows that the evolution of industrial clusters comes from the complex interaction of inner agents by themselves. The leading actions of initiative enterprises are the fundamental factors in the process of evolution of industrial clusters. Finally, the evolution trajectories of the agents are presented graphically that visibly verify and obviously describe the dynamic evolution process.