Improved Routh-Padé Approximants Using Vector Evaluated Genetic Algorithm to Unstable Systems

Abstract

This note describes a novel approach to Routh-Padé approximation problem relating to the construction of reduced-order approximants for continuous-time unstable systems. In this method, stability and the first r timemoments/Markov-parameters are preserved as well as the errors between a set of subsequent timemoments/Markov-parameters of the system and those of the model are minimized. For the solution of this problem a method using the concept of Pareto-optimality is proposed. Pareto-optimal curve is the solution of Multi-objective Optimization problem. Evolutionary Algorithm such as real parameter Genetic Algorithm is used to get Paretooptimal curve. The search area for GA is very wide and it usually converges to a point near global optima