Non-smooth economic dispatch computation by fuzzy and self adaptive particle swarm optimization

Abstract

Economic dispatch (ED) problem is a nonlinear and non-smooth optimization problem when valve-point effects, multi-fuel effects and prohibited operating zones (POZs) have been considered. This paper presents an efficient evolutionary method for a constrained ED problem using the new adaptive particle swarm optimization (NAPSO) algorithm. The original PSO has difficulties in premature convergence, performance and the diversity loss in optimization process as well as appropriate tuning of its parameters. In the proposed algorithm, to improve the global searching capability and prevent the convergence to local minima, a new mutation is integrated with adaptive particle swarm optimization (APSO). In APSO, the inertia weight is tuned by using fuzzy IF/THEN rules and the cognitive and the social parameters are self-adaptively adjusted. The proposed NAPSO algorithm is validated on test systems consisting of 6, 10, 15, 40 and 80 generators with the objective functions possessing prohibited zones, multi-fuel effects and valve-point loading effects. The research results reveal the effectiveness and applicability of the proposed algorithm to the practical ED problem.