Since their introduction, bio-inspired algorithms, especially the ones based on the social behaviour of the animals that live in colonies have demonstrated great potential in finding near-optimal solutions for both unconstrained and constrained hard optimization problems. In this research, a parallel version of the popular Artificial Bee Colony (ABC) algorithm for optimization of constrained problems, has been introduced. An island-based model, in which the whole population is divided into subpopulations, is used. Subpopulations execute the serial version of the original algorithm and occasionally exchange the obtained results. The proposed algorithm has been tested based on a set of well-known constraint benchmark functions and five real-world engineering design problems. The results demonstrate clear improvements compared with those obtained with the original ABC algorithm.
*This research is the continuation of our manuscript “Parallelized Multiple Swarm Artificial Bee Colony Algorithm (MS-ABC) for Global Optimization” published in Studies in Informatics and Control, vol. 23(1), 2014. In the previous research modified parallelized version of the original Artificial Bee Colony (ABC) algorithm has been applied to the optimization of unconstrained functions. The current paper presents parallelized version of ABC variant modified for the optimization of constrained functions. In the original paper the name of the algorithm was Multiple Swarm Artificial Bee Colony (MS-ABC) while in this paper the name Parallelized Multiple Swarm Artificial Bee Colony (PMS-ABC) is used.
Artificial bee colony, Optimization metaheuristics, Swarm intelligence, Parallelized algorithms, Nature inspired algorithms, Constraint optimization.
Miloš SUBOTIC, Aleksandar MANASIJEVIC, Aleksandar KUPUSINAC, "Parallelized Multiple Swarm Artificial Bee Colony (PMS-ABC) Algorithm for Constrained Optimization Problems*", Studies in Informatics and Control, ISSN 1220-1766, vol. 29(1), pp. 77-86, 2020. https://doi.org/10.24846/v29i1y202008
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).