Scholarlink

<p>Traditional metaheuristic methods often rely on random exploration and exploitation mechanisms, which can lead to inefficient search processes because of a lack of guidance. This paper introduces a novel metaheuristic algorithm that overcomes these limitations using two specialized mechanisms for exploration and exploitation. The population is divided into explorer and exploiter agents, each using distinct strategies. Explorer agents that resemble a swarm follow a trajectory generated through Latin hypercube sampling to efficiently explore the search space. Exploiters use evolutionary game theory, where weaker agents adopt the strategies of stronger ones, ensuring efficient exploitation. This integration of global information enhances diversity, improves the solution quality, and reduces computational costs. Validated against benchmark functions, the proposed algorithm delivered superior results, achieving faster convergence and higher-quality solutions.</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 improved swarm optimization algorithm using exploration and evolutionary game theory for efficient exploitation

  • 期刊论文

  • 发表日期：

    2025年3月4日

查看全文
  • Nahum Aguirre,
  • Erik Cuevas,
  • Alberto Luque-Chang,
  • Hector Escobar-Cuevas

摘要

Traditional metaheuristic methods often rely on random exploration and exploitation mechanisms, which can lead to inefficient search processes because of a lack of guidance. This paper introduces a novel metaheuristic algorithm that overcomes these limitations using two specialized mechanisms for exploration and exploitation. The population is divided into explorer and exploiter agents, each using distinct strategies. Explorer agents that resemble a swarm follow a trajectory generated through Latin hypercube sampling to efficiently explore the search space. Exploiters use evolutionary game theory, where weaker agents adopt the strategies of stronger ones, ensuring efficient exploitation. This integration of global information enhances diversity, improves the solution quality, and reduces computational costs. Validated against benchmark functions, the proposed algorithm delivered superior results, achieving faster convergence and higher-quality solutions.