Boosting Particle Swarm Optimization by Backtracking Search Algorithm for Optimization Problems
Sukanta Nama,Apu Kumar Saha,Sanjoy Chakraborty,Amir H. Gandomi,Laith Abualigah
DOI: https://doi.org/10.1016/j.swevo.2023.101304
IF: 10.267
2023-03-28
Swarm and Evolutionary Computation
Abstract:Adjusting the search behaviors of swarm-based algorithms during their execution is a fundamental errand for addressing real-world global optimizing challenges. Along this line, scholars are actively investigating the unvisited areas of a problem domain rationally. Particle Swarm Optimization (PSO), a popular swarm-based optimization algorithm, is broadly applied to resolve different real-world problems because of its more robust searching capacity. However, in some situations, due to an unbalanced trade-off between exploitation and exploration, PSO gets stuck in a suboptimal solution. To overcome this problem, this study proposes a new ensemble algorithm called e-mPSOBSA with the aid of the reformed Backtracking Search Algorithm (BSA) and PSO. The proposed technique first integrates PSO's operational potential and then introduces BSA's exploration capability to help boost global exploration, local exploitation, and an acceptable balance during the quest process. The IEEE CEC 2014 and CEC 2017 test function suite was considered for evaluation. The outcomes were contrasted with 26 state-of-the-art algorithms, including popular PSO and BSA variants. The convergence analysis, diversity analysis, and statistical test were also executed. In addition, the projected e-mPSOBSA was employed to evaluate four unconstrained and seven constrained engineering design problems, and performances were equated with various algorithms. All these analyses endorse the better performance of the suggested e-mPSOBSA for global optimization tasks, search performance, solution accuracy, and convergence rate.
computer science, artificial intelligence, theory & methods