Abstract
Nature-inspired population-based swarm intelligence (SI) algorithms can be used to obtain moderate results for optimization problems. However, by creating connections between the nature-inspired algorithms, further improved results can be achieved. Hybrid algorithms can play a noticeable role in improving the search capability of algorithms. In this paper, two swarm intelligence algorithms have been hybridized, which are—a modified version of the Artificial Bee Colony algorithm (ABC-T) and the Firefly Algorithm (FA). The proposed hybrid algorithm, named as Artificial Bee Colony-Inspired Firefly Algorithm (ABC-IFA), gave comparatively better results on a number of continuous function optimization problems. The hybridization is performed by implementing an island model where both ABC-T algorithm and FA are executed in an alternative manner, sequentially running one after another on two separate populations of candidate solutions. Then after every generation, some individuals from one population are migrated to the other to hybridize the properties of the different algorithms. As the migration policy, five different strategies are proposed, which are compared with each other. The hybrid approach is evaluated on 11 different benchmark functions, and the results are compared among the ABC-T, FA and the proposed hybrid ABC-IFA. The hybrid approach is observed to perform better on some of the benchmark functions compared to the ABC-T algorithm and FA.KeywordsHybrid algorithmMetaheuristic algorithmSwarm intelligence algorithmArtificial Bee Colony AlgorithmFirefly AlgorithmMigration
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call