Decentralized control of discrete-time large-scale systems by dynamic programming

A decentralized control algorithm for discrete-time large-scale systems with dynamic interconnections and external disturbances is presented. The method is based upon minimization of a discrete linear regulator performance measure and application of the principle of optimality. A suboptimal control law may be found by applying an approximation to the design. An example of a three-reach river pollution control problem is presented to illustrate the algorithm.