A new improved hybrid meta-heuristics method for unit commitment with nonlinear fuel cost function

Abstract

In this paper, a new improved hybrid meta-heuristic method is proposed to solve the unit commitment problem effectively. The objective is to minimize operation cost while satisfying the power balance constraints and so on. It may be formulated as a nonlinear mixed-integer problem. In other words, the unit commitment problem is hard to solve. Therefore, this paper makes use of a hybrid meta-heuristic method with two layers. Layer 1 determines the on/off conditions of generators with tabu search (TS) while Layer 2 evaluates output of generators with evolutionary particle swarm optimization (EPSO). The construction phase of Greedy Randomized Adaptive Search Procedure (GRASP) is used to create initial feasible solutions efficiently. Three kinds of meta-heuristic methods such as TS, EPSO and GRASP are combined to solve the problem. In addition a parallel scheme of EPSO is developed to improve the computational efficient as well as the accuracy. The effectiveness of the proposed method is tested in sample systems. © 2009 The Institute of Electrical Engineers of Japan.