Application of evolutionary computation techniques to multiobjective cascade control system design

Abstract

Cascade control systems are widely employed in most of the process newlineindustries involving multiple loops where inner loop disturbances are newlinepredominant Tuning of primary and secondary gains of controllers is an newlineimportant task in cascade control systems Single objective approaches fail to newlineyield better results when multiple conflicting objectives exist for a process newlineThe actual characteristics of the conflicting objectives are not preserved while newlinecombining more objectives into a single objective by weighted means This newlinethesis aims at developing a multiobjective cascade control system for tuning newlineboth the primary and secondary gains of cascade control systems based on newlineEvolutionary Multiobjective Optimization Algorithms such as NSGA II Non newlinedominated Sorting Genetic AlgorithmII and NSPSO Non dominated newlineSorting Particle Swarm Optimization taking into consideration the two newlineconflicting objectives overshoot and settling time The proposed scheme is newlineemployed for both regulatory and servo processes newline newline