Multi-objective optimal reactive power dispatch using modified game theory

Abstract

Multi-objective optimal reactive dispatch (OPRD) problem can be formulated as to maximize reactive power reserve and minimize bus voltage deviation while complying with all the power system network constraints. State-of-the-art methods for multi-objective optimization are learning-based evolutionary algorithms that attempt to generate Pareto solutions as dispersed as possible at the expense of huge computational complexities ranging from O(MN2) to O(MN3) for single generation, where M is the number of objective and N is the population size. They also require storage size that can be as large as O(N2). In this paper, we propose a modified cooperative game theory based approach to handle the multi-objective OPRD problem. Our modified game theory based approach can generate excellent compromise solution without creating the entire Pareto front, and thus have significantly less computation cost and minimal storage requirement. The proposed algorithm has a computational complexity of O(2N) and storage requirement of O(N) at each iteration. We evaluated the effectiveness of the proposed approach using the IEEE39 bus system, and very good results are reported.