A hybrid Mountain Gazelle particle swarm-based algorithm for constrained optimization problems

• Published in: Evolving systems Vol. 16; no. 1; p. 35
• Main Authors: Rani, Rekha, Garg, Vanita, Jain, Sarika, Garg, Harish
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2025; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Benchmarks; Complex Systems; Complexity; Convexity; Design optimization; Engineering; Evolutionary computation; Exploitation; Genetic algorithms; Heuristic; Hybridization; Integer programming; Methods; Mountains; Optimization algorithms; Optimization techniques; Original Paper; Particle swarm optimization; Performance enhancement; Problem solving; Statistical analysis; Structural design; Benchmark functions; CEC functions; Hybrid algorithms; Design problems; Particle swarm optimization; Mountain Gazelle optimization
• ISSN: 1868-6478; 1868-6486
• DOI: 10.1007/s12530-024-09654-w

Nature inspired techniques have been essential in addressing complex, non-linear and non-convex optimization problems. Combining two or more techniques is useful to improve accuracy, adaptability and efficiency in solving complex problems to handle the uncertainties and non-linearity effectively. In this paper, we introduce a hybrid approach called Mountain Gazelle Particle Swarm Optimization (MGPSO) which merges Particle Swarm Optimization (PSO) and Mountain Gazelle optimization (MGO) to improve problem-solving abilities. PSO has certain limitations in exploration but strong exploitation capabilities, while MGO has a strong exploration by enabling a more diverse and extensive search space. Keeping this in mind, we have integrated PSO with the MGO algorithm to strength of both algorithm and hence to improve the performance in solving complex optimization problems. The performance of the proposed hybrid algorithm has been tested against 23 benchmark functions of varying complexity, structural design problems and 12 IEEE CEC22 (IEEE Congress on Evolutionary Computation 2022) functions. The results computed are compared with several of the existing state-of-art literature. Experimental results shows that proposed MGPSO performs significantly better than other existing algorithms. Among the 23 benchmark functions, MGPSO archives superior results in 16 and for 12 CEC22 function, proposed algorithm excels in 10. Additionally, the MGPSO results are validated through the statistical analysis test.