Distribution network reconfiguration using a genetic algorithm with varying population size

Abstract

Genetic algorithm (GA) has been shown to be an effective way to solve the complex combinatorial, and constrained non-linear mixed integer optimization problem of distribution network reconfiguration. Extensive work has been done in the literature to efficiently apply GA to the reconfiguration problem. Nonetheless, to date, all the previous work in this area have presupposed that the GA population size should remain constant throughout the evolution of the search process. In this paper, we propose the application of a genetic algorithm with variable population size (GAVAPS) to the reconfiguration problem. We demonstrate that allowing the population size to adaptively grow and shrink according to the status of the GA search can allow for a more efficient solution, compared to standard genetic algorithm.