Analysis of takeover time and convergence rate for harmony search with novel selection methods. (1st January 2013)

• Record Type: Journal Article
• Authors: Doush, Iyad Abu; Al-Betar, Mohammed Azmi; Khader, Ahamad Tajudin; Awadallah, Mohammed A.; Mohammed, Ashraf Bany
• Is Part Of: International journal of mathematical modelling and numerical optimisation. Volume 4:Number 4(2013)
• Page Start: 305
• Page End: 322
• Publication Date: 2013-01-01
• Subjects: harmony search algorithm -- HSA -- evolutionary algorithms -- EA -- selection mechanisms -- meta-heuristic algorithm; Mathematical models -- Periodicals; Mathematical optimization -- Periodicals; 519.605
• Languages: English
• ISSNs: 2040-3607

Abstract

Recently, common selection schemes used in harmony search algorithm (HSA) are altered in memory consideration operation to imitate the natural selection principle of survival of the fittest. The selection schemes adopted include: random, proportional, tournament, and linear rank. In this paper, these selection schemes are analysed in order to evaluate their effect on the performance of HSA. The analysis considers takeover time and convergence rate to measure the effectiveness of each selection scheme. Furthermore, a scaled proportional selection scheme is proposed to replace the proportional selection scheme to overcome its shortcoming with negative fitness values. To study the effect of these different selection schemes we use eight global optimisation functions with different characteristics. An experimental evaluation show that linear rank selection provides the highest convergence speed and highest takeover time. On the other hand, scaled proportional selection provides the slowest convergence speed and slowest takeover time. This indicates the effect of the type of the selection method used in memory consideration in takeover time and convergence rate.