Open AccessAdaptive multiple evolutionary algorithms search for multi‐objective optimal reactive power dispatch
Author(s) -
Hongxin Li,
Yinhong Li,
Jinfu Chen
Publication year - 2014
Publication title -
international transactions on electrical energy systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.428
H-Index - 42
ISSN - 2050-7038
DOI - 10.1002/etep.1730
Subject(s) - mathematical optimization , pareto principle , evolutionary algorithm , computer science , complementarity (molecular biology) , convergence (economics) , multi objective optimization , pareto optimal , process (computing) , consistency (knowledge bases) , algorithm , mathematics , artificial intelligence , genetics , economics , biology , operating system , economic growth
SUMMARY Multi‐objective evolutionary algorithm (MOEA) based on Pareto optimality has emerged as an important approach for optimal reactive power dispatch (ORPD) problem. However, because of the deficiency of adopting evolution operators with single search characteristics, existing MOEA for multi‐objective ORPD (MORPD) often fails to maintain universal and robust performance in different phases of optimization process. On the basis of running multiple algorithms simultaneously and adaptive selection strategy, this paper proposes a new optimal method for MORPD. Firstly, an algorithm candidate pool containing four different algorithms is constructed through the analysis of characteristics of state‐of‐the‐art MOEA while considering the rules of consistency and complementarity. Then, during the optimizing search, the quantity of offspring individuals generated by each candidate algorithm at different stages of optimization process is determined adaptively by learning from its previous experience in generating promising solutions. The proposed method is tested on the IEEE 30‐bus system; its computing performance is compared with existing popular MOEAs from the point of view of Pareto fronts, outer solutions and C measure. Experimental results show that the new method can obtain better performance of convergence during the entire optimization process. Copyright © 2013 John Wiley & Sons, Ltd.