MOEA/D with many-stage dynamical resource allocation strategy to solution of many-objective OPF problems

As people's electricity demand and environmental awareness increase, single-objective optimization of power systems can no longer meet the requirements of modern power system operation and management. More and more optimization objectives will be considered in the mathematical modeling of power systems optimization, and solution techniques for many-objective optimal power flow problem (objectives usually more than three) are needed. This paper firstly formulates the OPF problem as a many-objective OPF (Ma-OPF) problem with consideration of minimizing many objectives and multiple complicated constraints. Then the MOEA/D with manystage dynamical resource allocation strategy is proposed to solve the established model. In the proposed approach, three improvements are proposed: Firstly, a novel polymerization method is proposed when allocating computing resource; Secondly, the MOEA/D-MRA divides evolutionary process into two kinds of alternate stages throughout the evolution process: convergence stage and diversity enhancement one; Thirdly, to avoid the recurrent selections, in each generation, subproblems selected in previous evolution will be removed. At last, the famous benchmark test power systems, i.e., IEEE 30-bus, IEEE 57-bus system and 118-bus system are employed to test the MOEA/D-MRA. The obtained results and statistical analysis demonstrate the competitiveness and effectiveness of the proposed MOEA/D-MRA for Ma-OPF problems.