A hybrid evolutionary algorithm with guided mutation for minimum weight dominating set

Abstract

This paper presents a hybrid evolutionary algorithm with guided mutation (EA/G) to solve the minimum weight dominating set problem (MWDS) which is NP-hard in nature not only for general graphs, but also for unit disk graphs (UDG). MWDS finds practical applications in diverse domains such as clustering in wireless networks, intrusion detection in adhoc networks, multi-document summarization in information retrieval, query selection in web databases etc. EA/G is a recently proposed evolutionary algorithm that tries to overcome the shortcomings of genetic algorithms (GAs) and estimation of distribution algorithms (EDAs) both, and that can be considered as a cross between the two. The solution obtained through EA/G algorithm is further improved through an improvement operator. We have compared the performance of our hybrid evolutionary approach with the state-of-the-art approaches on general graphs as well as on UDG. Computational results show the superiority of our approach in terms of solution quality as well as execution time.