OSA: Orthogonal Simulated Annealing Algorithm and Its Application to Designing Mixed<tex>$rm H_2 /rm H_infty$</tex>Optimal Controllers

Abstract

This paper proposes a novel orthogonal simulated annealing (OSA) algorithm for solving intractable large-scale engineering problems and its application to designing mixed H/sub 2//H/sub /spl infin// optimal structure-specified controllers with robust stability and disturbance attenuation. High performance of OSA arises mainly from an intelligent generation mechanism (IGM), which applies orthogonal experimental design to speed up the search. IGM can efficiently generate a good candidate solution for next move of OSA by using a systematic reasoning method. It is difficult for existing H/sub /spl infin//- and genetic algorithm (GA)-based methods to economically obtain an accurate solution to the design problem of multiple-input, multiple-output (MIMO) optimal control systems. The high performance and validity of OSA are demonstrated by parametric optimization functions and a MIMO super maneuverable F18/HARV fighter aircraft system with a proportional-integral-derivative (PID)-type controller. It is shown empirically that OSA performs well for parametric optimization functions and the performance of the OSA-based method without prior domain knowledge is superior to those of existing H/sub /spl infin//- and GA-based methods for designing MIMO optimal controllers.