Optimization of the Distributed Systems by Autonomous Cooperation —Distributed Maximum Principle—

This paper proposes a methodology for optimizing the whole distributed systems by their autonomous activities by means of the cooperative work among them. Although the proposed method is developed for the design of a large-scale complex system such as chemical process plants, its wide applicability is expected to realize the optimum collaborative control of the distributed autonomous robotic systems. The method intorduces market mechanism to provide optimizing prices by which individual subsystems make decisions to maximize their profits autonomously so as to achieve the optimization of the total system. The optimizing prices are obtained through the cooperation among subsystems. The completely decentraized and distributed framework of the method provides the concurrent, robust solution for a large scale problem as well as distributed systems.