Automatic weight selection for H/sub /spl infin// control design

The paper is concerned with the development of a procedure to calculate the parameters of the weighting functions used in H/sub /spl infin// controller designs in order to achieve a desired system performance. A genetic algorithm is employed to search for suitable solutions. To cope with the imprecision and vagueness that arises in the description of objective functions and problem constraints, concepts of the fuzzy logic are incorporated in the solution. A multi objective fuzzy optimisation is stated and fuzzy convex decision making is established. To evaluate the potentiality of the proposed method, the tuning of an autopilot of a ship is investigated and compared to results obtained after a traditional design stage. Validation includes nonlinear simulations of the ship dynamics.