Proceedings of the
The Nineteenth International Conference on Computational Intelligence and Security (CIS 2023)
December 1 – 4, 2023, Haikou, China
A Two-Stage Hybrid Multi-Objective Competitive Particle Swarm Optimization Algorithm with Assisted Search Techniques
1School of Computer Science and Technology, Qinghai Normal University, Xining, China.
2School of Mathematics and Statistics, Qinghai Normal University, Xining, China.
ABSTRACT
The balance between diversity and convergence has always been a hot topic in multi-objective optimization problems (MOPs). In this paper, a two-stage hybrid multi-objective competitive particle swarm optimizer with assisted search techniques (HMOPSO-AST) is proposed. Firstly, during the early stages of evolution, to enhance the information exchange among individuals, the multi-objective competitive particle swarm algorithm serves as the main algorithm, with the model-based distributed estimation algorithm employed as an auxiliary approach. Subsequently, in the later stages of evolution, the main algorithm persists in its evolution, while genetic operators are utilized as auxiliary search methods to expedite convergence rate and enhance the uniform distribution of solutions, thus effectively steering the algorithm's search direction. Furthmore, the performance of the proposed algorithm is verified on two test suites. The experimental results demonstrate that HMOPSO-AST excels in achieving better balance between diversity and convergence, thereby enhancing the precision of the solutions.
Keywords: Multi-objective optimization, Competitive PSO algorithm, Model-based MOEA, Genetic operators, Assisted search techniques.
Download PDF
1School of Computer Science and Technology, Qinghai Normal University, Xining, China.
2School of Mathematics and Statistics, Qinghai Normal University, Xining, China.
ABSTRACT
The balance between diversity and convergence has always been a hot topic in multi-objective optimization problems (MOPs). In this paper, a two-stage hybrid multi-objective competitive particle swarm optimizer with assisted search techniques (HMOPSO-AST) is proposed. Firstly, during the early stages of evolution, to enhance the information exchange among individuals, the multi-objective competitive particle swarm algorithm serves as the main algorithm, with the model-based distributed estimation algorithm employed as an auxiliary approach. Subsequently, in the later stages of evolution, the main algorithm persists in its evolution, while genetic operators are utilized as auxiliary search methods to expedite convergence rate and enhance the uniform distribution of solutions, thus effectively steering the algorithm's search direction. Furthmore, the performance of the proposed algorithm is verified on two test suites. The experimental results demonstrate that HMOPSO-AST excels in achieving better balance between diversity and convergence, thereby enhancing the precision of the solutions.
Keywords: Multi-objective optimization, Competitive PSO algorithm, Model-based MOEA, Genetic operators, Assisted search techniques.
Download PDF