NSGA-II with local search for a multi-objective reactive power compensation problem

Abstract

This paper presents the results of an Elitist Non-dominated Sorting Genetic Algorithm (NSGA II) enhanced with local search for computing solutions to a multi-objective reactive power compensation model. NSGA II has revealed a good performance in comparison with other multi-objective evolutionary algorithms also tested to tackle this problem. This performance is still improved using a local search scheme within NSGA II, which is specially tailored to the problem characteristics. The use of local search within the NSGA II operational framework in this complex combinatorial problem improves convergence towards the non-dominated front and ensures that the solutions attained are well spread over it. A comparative study is presented between the results obtained using a standard NSGA II based approach and the enhanced NSGA II approach with local search, to provide decision support in the VAR planning problem in radial electric distribution systems. The model explicitly considers two objective functions concerning economical and operational evaluation aspects.