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

Journal of Information Science and Engineering, Vol. 34 No. 2, pp. 353-369

Multi-Objective Optimization using Co-variance Guided Artificial Bee Colony

Department of Computer Science and Engineering
Motilal Nehru National Institute of Technology Allahabad
Allahabad, 211004 India
E-mail: {divyak; kkm}@mnnit.ac.in 

    Use of statistics and co-variance principles in the domains of soft computing is an emerging and intriguing research area. Co-variance based optimization algorithms have faster convergence rate and remarkable efficiency as they possess fair information about the gradient of objective function. Artificial Bee Colony (ABC) is one of the highly researched and utilized optimization technique, based on the foraging behavior of honey bee swarms. In this article, we have proposed a novel ABC based algorithm for multi- objective optimization. The proposed Multi-objective Co-variance based ABC (M-CABC), is the first algorithm in its class which works on the coalition of statistical co-variance and ABC. The performance of the proposed algorithm has been gauged on both unconstrained and constrained multi-objective benchmark functions on the basis of error rate, generational distance and spacing metrics. The results have shown that M-CABC has eminently converged close to optimal pareto front while successfully maintaining high diversity in the solutions. The article concludes with the observatory remarks on the performance of M-CABC, which is consistent in solving low as well as high dimension multi-objective benchmark problems when it has been compared with other traditionally established multi-objective optimization algorithms. 

Keywords: meta-heuristics, swarm intelligence, artificial bee colony, C-variance, multi- objective optimization, non-dominated sorting

  Retrieve PDF document (JISE_201802_03.pdf)