Gravitational search algorithm combined with modified differential evolution learning for planarization in graph drawing

Abstract

Gravitational search algorithm (GSA) is one of the powerful population based meta-heuristics. It has achieved many successes in various applications derived from optimization, data mining, information security, etc. However, it still suffers from the local optima trapping problem and cannot obtain promising solutions especially for practical problems. Graph planarization arises from many practical applications of VLSI circuit design, automatic graph drawing, etc, and is proved to be NP-hard. To solve this problem, this study proposes a hybrid GSA by combined with a differential evolution operator. The proposed method GSADE is used to acquire optimal planar subgraphs for a given graph. Experimental results based on thirty graph instances show that GSADE is a very competitive method in comparison with previous state-of-the-art methods.