Scholarlink

This paper attempts to enhance the modelling results of the radial basis function-based neural network (RbN) through bionic evolutionary algorithms (EAs). Then, the artificial bee colony (ABC) and particle swarm optimization (PSO) algorithms are applied to modulate RbN. The proposed integrated of ABC and PSO principles (IABPS) algorithm combines principles from these bionic EAs, consolidates the variation and stability features for optimization solving. The property of population alteration has demonstrated significant potential in having the global optimal to remove local optimal being constrained and this algorithm is commonly used in solving three continuous nonlinear function experiments. The investigational results have shown that the hybrid of ABC and PSO principles is an outstanding method and thus the IABPS algorithm is proposed, aiming to achieve optimal accuracy in modelling compared to associated algorithms in this paper. The algorithm then assesses effects from three experimental functions, which reveals that the proposed IABPS algorithm outperforms relevant algorithms in precision for function approximation.

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

Consolidation of Biomimicry and Swarm-Based Metaheuristic Algorithms with Supervised Learning Neural Network for Function Approximation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zhen-Yao Chen,
  • Ming-Te Cheng,
  • Yun-You Tsai

摘要

This paper attempts to enhance the modelling results of the radial basis function-based neural network (RbN) through bionic evolutionary algorithms (EAs). Then, the artificial bee colony (ABC) and particle swarm optimization (PSO) algorithms are applied to modulate RbN. The proposed integrated of ABC and PSO principles (IABPS) algorithm combines principles from these bionic EAs, consolidates the variation and stability features for optimization solving. The property of population alteration has demonstrated significant potential in having the global optimal to remove local optimal being constrained and this algorithm is commonly used in solving three continuous nonlinear function experiments. The investigational results have shown that the hybrid of ABC and PSO principles is an outstanding method and thus the IABPS algorithm is proposed, aiming to achieve optimal accuracy in modelling compared to associated algorithms in this paper. The algorithm then assesses effects from three experimental functions, which reveals that the proposed IABPS algorithm outperforms relevant algorithms in precision for function approximation.