Multi-objective reconfiguration of the distribution systems by using NSGA-II and local improvement

Abstract

Many heuristics for network reconfiguration rely on the systematic applying of the branch-exchange technique. In this work, two novel genetic operators for crossover and mutation have been developed that are based on the referred technique. The chromosome's codification to use these operators is straightforward and is not required any additional knowledge of graph theory to achieve the feasibility of individuals. As one of their main novelties, the methodology shows how can be employed a local improvement step, used commonly in the single-objective optimization, in the multi-objective optimization. This step increases the convergence of the optimization with populations of much reduced size. The proposed methodology is tested by solving several examples of the literature, including or not the local improvement step. The comparison of the results with the best solutions published for these examples shows the effectiveness of the method.