Title: Adaptive distributed modified extremal optimisation for maximising contact map overlap and its performance evaluation
Authors: Keiichi Tamura; Hajime Kitakami; Tatsuhiro Sakai
Addresses: Graduate School of Information Sciences, Hiroshima City University, 3-4-1, Ozuka-Higashi, Asa-Minami-Ku, Hiroshima 731-3194, Japan ' Department of Information Systems and Management, Faculty of Applied Information Science, Hiroshima Institute of Technology, 2-1-1 Miyake, Saeki-ku, Hiroshima 731-5193, Japan ' Graduate School of Information Sciences, Hiroshima City University, 3-4-1, Ozuka-Higashi, Asa-Minami-Ku, Hiroshima 731-3194, Japan
Abstract: Maximising the contact map overlap (CMO) problem is one of the simplest yet most robust techniques for finding optimal protein structure alignment. This optimisation is known as the CMO problem, and is also known as NP-hard. We have been developing bio-inspired heuristics using distributed modified extremal optimisation (DMEO) for the CMO problem. DMEO is a hybrid of population-based modified extremal optimisation (PMEO) and the island model. In our previous work, we proposed a DMEO-based bio-inspired heuristic, i.e., DMEO with different evolutionary strategies (DMEODES) to maintain the population diversity of evolution. DMEODES efficiently maintains population diversity; however, once the population falls into local optimal solutions, there is no mechanism for getting out of them. In this paper, we propose a novel heuristic model to improve the DMEO's ability to prevent evolution stagnation. The new model integrates an adaptive generation alternation mechanism in DMEO called ADMEO. The experimental results show that ADMEO outperforms DMEODES.
Keywords: contact map maximisation problem; extremal optimisation; distributed extremal optimisation; bio-inspired heuristic; island model.
International Journal of Computational Intelligence Studies, 2017 Vol.6 No.4, pp.288 - 310
Received: 01 Jun 2017
Accepted: 10 Aug 2017
Published online: 23 Jan 2018 *