Improving the responsiveness of NSGA-II in dynamic environments using an adaptive mutation operator - A case study

Abstract

This paper presents a comparative analysis of the results obtained with two different implementations of the NSGA-II genetic algorithm in the framework of load management activities in electric power systems. The multiobjective real-world problem deals with the identification and the selection of suitable control strategies to be applied to groups of electric loads aimed at reducing maximum power demand, maximize profits and minimize user discomfort. It is shown that the algorithm performance is improved when the NSGA-II mutation operator is adaptively changed to incorporate information about the results of the search process and transfer this "knowledge" to the population. © 2008 Springer-Verlag Berlin Heidelberg.