Scholarlink

<p>Constrained optimization problems are widely used in science and engineering. The traditional genetic optimization methods are easy to fall into local optimal solutions and the convergence speed can be increased. In this paper, we propose a double variant evolutionary algorithm (DMGESNP) based on extended spike neural membrane system and improved genetic algorithm, which incorporates genetic operators into the evolution process of membrane system, groups, crosses and mutates population individuals, and acts as the input population of the next P system. This paper aimed to make the evolutionary process more in line with the characteristics of biological development and improve the speed of evolutionary convergence. In the experiment, sensitivity analysis was conducted on three key parameters, followed by testing the optimization performance of DMGESNP using four complex benchmark functions. Finally, the classical 0–1 knapsack problem (KP) was used for constraint testing, with a total of 12 KP of different scales used to verify the performance of DMGESNP, and population diversity was analyzed. The comparative experimental results show that DMGESNP outperforms the comparative algorithms in terms of optimization performance, average knapsack value, and corresponding standard deviation. It also maintains good population diversity during the evolution process, accelerates convergence, and has stronger stability, allowing individuals in the population to surround the optimal solution.</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

A dual mutation strategy based on an improved genetic algorithm and a spike neural membrane system to solve binary problems

  • 期刊论文

  • 发表日期：

    2025年9月30日

查看全文
  • Jiachang Xu,
  • Dongliang Yu,
  • Ruichong Fang,
  • Hongjin Li,
  • Shuzhi Su

摘要

Constrained optimization problems are widely used in science and engineering. The traditional genetic optimization methods are easy to fall into local optimal solutions and the convergence speed can be increased. In this paper, we propose a double variant evolutionary algorithm (DMGESNP) based on extended spike neural membrane system and improved genetic algorithm, which incorporates genetic operators into the evolution process of membrane system, groups, crosses and mutates population individuals, and acts as the input population of the next P system. This paper aimed to make the evolutionary process more in line with the characteristics of biological development and improve the speed of evolutionary convergence. In the experiment, sensitivity analysis was conducted on three key parameters, followed by testing the optimization performance of DMGESNP using four complex benchmark functions. Finally, the classical 0–1 knapsack problem (KP) was used for constraint testing, with a total of 12 KP of different scales used to verify the performance of DMGESNP, and population diversity was analyzed. The comparative experimental results show that DMGESNP outperforms the comparative algorithms in terms of optimization performance, average knapsack value, and corresponding standard deviation. It also maintains good population diversity during the evolution process, accelerates convergence, and has stronger stability, allowing individuals in the population to surround the optimal solution.