OPTIMAL SWITCHING CONTROL OF MESSAGE TRANSMISSION IN COMPUTER NETWORKS

. In this paper, we consider the router buffer modeling and the optimal message transmission problem in computer networks. For the routing networks with multiple source inputs and multiple link outputs, we propose a switched system, in which the switchings depending on both state and control, to describe the change of queue length in router buffers. Taking the weighted sum of the received data at the destination routers and the congestion time of routers as the performance index, we then present an optimal control problem involving the switched system and subject to continuous state inequality constraints. To solve this optimal control problem, we transform the above switched system to the one with time-dependent switchings. Furthermore, by using the control parametrization, time-scaling transformation and constraint transcription techniques, we approximate the resulting optimal control problem by a series of parameter optimization problems, and develop an improved particle swarm optimization algorithm for solving the parameter optimization problems. Finally, we demonstrate the effectiveness of the algorithm by solving three test problems and two example problems.