[ 1 ] [ 2 ] [ 3 ] [ 4 ] [ 5 ] [ 6 ] [ 7 ] [ 8 ] [ 9 ] [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ]

Journal of Information Science and Engineering, Vol. 36 No. 6, pp. 1293-1314

Broadening Selection Competitive Constraint Handling Algorithm for Faster Convergence

1Faculty of Engineering Sciences and Technology
Hamdard University
Karachi, 74600 Pakistan
E-mail: ftayyab.ahmed; dr.sajjad; rizwan.tanweerg@hamdard.edu.pk

2Faculty of Science and Information Technology
University Technology Petronas
32610 Seri Iskandar, Perak, Malaysia
E-mail: manzoor.hashmani@utp.edu.my

In this paper, a new algorithm incorporating broadening selection strategy in competitive constraint handling paradigm for finding the optimum solution in constrained problems has been proposed, referred as Broadening Selection Competitive Constraint Handling (BSCCH). Although, competitive constraint handling approaches have proved to be very efficient, but they lack faster convergence due to offspring generation from random individuals. By incorporating selection strategy such as broadening selection in the competitive approach, better results are obtained and convergence rate is improved significantly. Incorporating said strategy, the BSCCH algorithm has been proposed which is generic in nature and can be coupled with various evolutionary algorithms. In this study, the BSCCH algorithm has been coupled with Differential Evolution algorithm as a proof of concept because it is found to be an efficient algorithm in the literature for constrained optimization problems. The proposed algorithm has been evaluated using 24 benchmark functions. The mean closure performance of the BSCCH algorithm is compared against seven selected state-of-the-art algorithms, namely Differential Evolution with Adaptive Trial Vector Generation Strategy and Cluster-replacement-based Feasibility Rule (CACDE), Improved Teaching Learning Based Optimization (ITLBO), Modified Global Best Artificial Bee Colony (MGABC), Stochastic Ranking Differential Evolution (SRDE), Novel Differential Evolution (NDE), Partical Swarm Optimization for solving engineering problems – a new constraint handling mechanism (CVI-PSO) and Ensemble of Constraint Handling Techniques (ECHT). The median convergence traces have been compared with two different algorithms based on differential evolution, i:e: Ensemble of Constraint Handling Techniques (ECHT) and Stochastic Ranking Differential Evolution (SRDE). ECHT is considered to be a flagship ensemble technique till date for constrained optimization problems, whereas SRDE employs a parent selection mechanism for constrained optimization. The proposed algorithm is found to provide better solutions and achieve significantly faster convergence in most of the problems. 

Keywords: constraint handling techniques, competitive approach, selection strategy, differential evolution, ranking methodology

  Retrieve PDF document (JISE_202006_11.pdf)