Scholarlink

<p>The search process with metaheuristic algorithms is mostly performed using one operator. The most important problem of using only one operator in the algorithm is that the success of the algorithm is determined by the success of the operator used. If the selected operator fails, it can be said that it is very difficult for the algorithm to be successful. To improve the algorithm’s performance, the number of operators can also be increased. Particle Swarm Optimization is a nature-inspired meta-heuristic optimization algorithm in which particles move through a search space, updating their position and velocity to search for the best solution. The well-known Particle Swarm Optimization algorithm also achieves better results using the operator. This paper is motivated to develop a more flexible and efficient optimization approach by addressing the limitations of the search processes in metaheuristic algorithms, which are usually performed using a single operator. The performance of search strategies based on a single operator is limited by the efficiency of the selected operator, and the overall performance of the algorithm is severely degraded when the operator fails. In order to overcome this problem, we propose the Particle Swarm Optimization Probability Matching (PSOPM) algorithm, which allows the use of three different operators. The proposed method is based on determining the optimum parameters through parameter tuning processes and then selecting the most suitable operator using Probability Matching. The algorithm is tested on CEC 2021 problems (1 unimodal, 6 multimodal, 3 combination functions), CEC 2013 problems (5 unimodal, 15 multimodal, 8 combination functions) and engineering design problems Speed Reducer Design (SRD) and Welded Beam Design (WBD). The main contributions of the study are that PSOPM outperforms existing algorithms on SRD and WBD problems, provides the best results on CEC 2021 problems in all dimensions, and provides the best results on CEC 2013 problems, especially in dimensions 30 and 50. Moreover, the competitive performance of the algorithm on the 100-dimensional CEC 2013 problems demonstrates the flexibility and efficiency of the proposed approach. These findings show that the use of multiple operators and the Probability Matching method make important scientific contributions to optimization problems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An effective approach for adaptive operator selection and comparison for PSO algorithm

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月16日

查看全文
  • Ahmet Akkaya,
  • Cemil Közkurt

摘要

The search process with metaheuristic algorithms is mostly performed using one operator. The most important problem of using only one operator in the algorithm is that the success of the algorithm is determined by the success of the operator used. If the selected operator fails, it can be said that it is very difficult for the algorithm to be successful. To improve the algorithm’s performance, the number of operators can also be increased. Particle Swarm Optimization is a nature-inspired meta-heuristic optimization algorithm in which particles move through a search space, updating their position and velocity to search for the best solution. The well-known Particle Swarm Optimization algorithm also achieves better results using the operator. This paper is motivated to develop a more flexible and efficient optimization approach by addressing the limitations of the search processes in metaheuristic algorithms, which are usually performed using a single operator. The performance of search strategies based on a single operator is limited by the efficiency of the selected operator, and the overall performance of the algorithm is severely degraded when the operator fails. In order to overcome this problem, we propose the Particle Swarm Optimization Probability Matching (PSOPM) algorithm, which allows the use of three different operators. The proposed method is based on determining the optimum parameters through parameter tuning processes and then selecting the most suitable operator using Probability Matching. The algorithm is tested on CEC 2021 problems (1 unimodal, 6 multimodal, 3 combination functions), CEC 2013 problems (5 unimodal, 15 multimodal, 8 combination functions) and engineering design problems Speed Reducer Design (SRD) and Welded Beam Design (WBD). The main contributions of the study are that PSOPM outperforms existing algorithms on SRD and WBD problems, provides the best results on CEC 2021 problems in all dimensions, and provides the best results on CEC 2013 problems, especially in dimensions 30 and 50. Moreover, the competitive performance of the algorithm on the 100-dimensional CEC 2013 problems demonstrates the flexibility and efficiency of the proposed approach. These findings show that the use of multiple operators and the Probability Matching method make important scientific contributions to optimization problems.