A novel method of distribution power system reconfiguration using parallel cooperative meta-heuristics

摘要

This paper presents a new robust methodology for solving radial distribution system reconfiguration (DSR) problem based on the concept of cooperative multi-thread strategy and hybrid meta-heuristics. The parallel cooperative meta-heuristics (PCMH) method deploys multiple concurrent explorations of the solution space using genetic algorithm (GA), particle swarm optimization (PSO) and ant colony system (ACS) running in parallel on independent processors. The meta-heuristics are subjected to communicate and cooperate with each other synchronously at a specified interval in the course of iteration. Dynamically at each interval, the best solution so far found is exchanged among all the three processors. In addition, the worst individuals of GA are replaced by the best particles of PSO if the current optimum is delivered by PSO; otherwise the best ants from ACS replace these individuals. The proposed method is applied to the standard IEEE 33-bus distribution network with the objective of minimizing the overall power loss while at the same time accounting for the operational constraints within allowable limits. The results demonstrate the validity and the effectiveness of the proposed approach in terms of accuracy and robustness.