A hierarchical optimization algorithm on distributed computer control system

This paper describes a hierarchical optimization algorithm for distributed computer control system based on the theory of decomposition-coordination on large scale systems. With the algorithm, the control in each local agent can be divided two parts. The one is the local feedback by the subsystem, the other comes from the coordination. The feedback is only related to the structure and states of subsystem but the coordination is concerned with whole systems. By regulating the slack matrix, the iterations are reduced and coordination strategy converges very fast. The carefully designed experiments show that the method is suited to DCCS.