Multi-objective Cascade Control of Regulatory Process with Two Conflicting Objectives

Abstract

This paper describes a system for tuning PI controller of cascade control system. The tuning approach is based on Evolutionary Multiobjective Optimization Algorithm (EMOA). The EMOA used are NSGA-II (Non Dominated Sorting Genetic Algorithm-II) and NSPSO (Non Dominated Sorting Particle Swarm Optimization). In the proposed method, steady state analysis and disturbance rejection are the objectives considered. Therefore Multiobjective optimization technique is used. PI controllers are used for controlling both level and flow process. PI controllers are preferred for a fast process like flow, the reason being its faster than other conventional controllers. The multiobjective optimization results in approximation of Pareto fronts. The final optimal solution is selected based on the objectives. Simulation results obtained using NSGA-II and NSPSO are compared.