Scholarlink

This paper investigates several well-known meta-heuristic algorithms and conducts experiments and analyses on their performance using the universal test suites on evolutionary computation. The experimental results reveal the following findings: (1) The performance of the Particle Swarm Optimization (PSO) algorithm is usually inferior to Differential Evolution (DE) and the QUasi-Affine TRansformation Evolution (QUATRE) algorithm in terms of both optimization accuracy and convergence speed. (2) Genetic Algorithms (GA), DE, and QUATRE belong to the same class of algorithms. DE originated from the genetic annealing algorithm, which is a hybrid genetic algorithms and simulated annealing, thereby inheriting the search framework of genetic algorithms. (3) The QUATRE algorithm introduces an evolution matrix that replaces the crossover operation in DE, while still utilizing the same search manner of a high-dimensional cube, thereby achieving unbiased and efficient search. In summary, the QUATRE algorithm exhibits fast convergence speed and strong global search capability, making it suitable for solving high-dimensional and complex problems in comparison with DE and PSO.

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

Performance Analysis of Several Famous Meta-Heuristic Algorithms Under the Congress on Evolutionary Computation Test Suites

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Xiubin Chen,
  • Dewang Chen,
  • Zhenyu Meng

摘要

This paper investigates several well-known meta-heuristic algorithms and conducts experiments and analyses on their performance using the universal test suites on evolutionary computation. The experimental results reveal the following findings: (1) The performance of the Particle Swarm Optimization (PSO) algorithm is usually inferior to Differential Evolution (DE) and the QUasi-Affine TRansformation Evolution (QUATRE) algorithm in terms of both optimization accuracy and convergence speed. (2) Genetic Algorithms (GA), DE, and QUATRE belong to the same class of algorithms. DE originated from the genetic annealing algorithm, which is a hybrid genetic algorithms and simulated annealing, thereby inheriting the search framework of genetic algorithms. (3) The QUATRE algorithm introduces an evolution matrix that replaces the crossover operation in DE, while still utilizing the same search manner of a high-dimensional cube, thereby achieving unbiased and efficient search. In summary, the QUATRE algorithm exhibits fast convergence speed and strong global search capability, making it suitable for solving high-dimensional and complex problems in comparison with DE and PSO.