Multiobjective fuzzy genetic algorithmoptimisationapproach to nonlinear control system design
Ashitey Trebi‐Ollennu, Brian A. White · IEE Proceedings - Control Theory and Applications · 1997
Owing to the large number of free control parameters for modern nonlinear robust controllers, it is almost impossible to heuristically tune these parameters. The multiobjective fuzzy genetic algorithm optimisation is shown to provide an effective, efficient and intuitive framework for selecting these parameters. The control structure and specifications are assumed to be given. Using the concept of fuzzy sets and convex fuzzy decision making, a multiobjective fuzzy optimisation problem is formulated and solved using a genetic algorithm. The relative importance of the objective functions is assessed by using a new membership weighting strategy. The technique is applied to the selection of free control parameters for an input-output linearising controller with sliding mode control, in a remotely-operated underwater vehicle depth control system.