A neurodynamic optimization approach to robust pole assignment based on convex reformulation

Another neurodynamic optimization approach to robust pole assignment is presented for synthesizing linear control systems. The original pseudoconvex optimization problem for robust pole assignment is reformulated as a convex optimization problem. Three coupled recurrent neural networks operating in three different time scales are developed for solving the reformulated problem in real time. It is shown that robust parametric configuration and exact pole assignment of feedback control systems can be achieved. Two examples of the proposed approach are discussed in detail to demonstrate its effectiveness.