基于多目标人类学习算法的电力系统环境经济调度研究

程传良

基于多目标人类学习算法的电力系统环境经济调度研究

投稿时间：2026-04-12 修订日期：2026-04-28 点此下载全文

引用本文：

摘要点击次数: 73

全文下载次数: 0

作者	单位	邮编
程传良^*	Chaohu University	238024

基金项目:国家自然科学基金项目（青年项目）

中文摘要:环境经济调度 (EED) 的目的是在满足供需平衡和生产指标的前提下，合理分配发电机组的发电量，同时最大限度地降低火力发电的经济成本和环境污染。在解决此类问题的多目标优化算法中，冲突的目标要求算法提供多个最优解，这导致算法难以确保解集的多样性，且容易陷入局部最优。因此，本文提出了一种多目标人类学习优化算法 (MOHLO)，通过结合平均距离和拥挤距离指标来保持帕累托最优前沿的多样性。此外，引入了一种基于帕累托近似中点的次优解消除机制，可以有效识别和删除弱帕累托解。最后，该算法被应用于电力系统EED模型中，并在收敛性和多样性方面显示出其有效的结果。

中文关键词:环境经济调度多目标人类学习优化平均距离拥挤距离次优解消除机制

Research on Environmental Economic Dispatch of Power System Based on Multi-Objective Human Learning Algorithm

Abstract:The purpose of environmental-economic dispatch (EED) is to reasonably allocate the electricity generation of power units while ensuring the balance of supply and demand and meeting production targets, and to minimize the economic costs and environmental pollution of power generation to the greatest extent. In multi-objective optimization algorithms for solving such problems, conflicting objectives require the algorithm to provide multiple optimal solutions, making it difficult for the algorithm to ensure the diversity of the solution set and prone to falling into local optima. Therefore, this paper proposes a multi-objective human learning optimization (MOHLO) algorithm that maintains the diversity of the Pareto optimal front by combining average distance and crowding distance indicators. In addition, a sub-optimal solution elimination mechanism based on the Pareto approximation midpoint is introduced, which can effectively identify and discard weak Pareto solutions. Finally, the algorithm is applied to the EED model of the power system, demonstrating superior results in terms of convergence and diversity.

keywords:environmental-economic dispatch multi-objective human learning optimization average distance crowding distance sub-optimal solution elimination mechanism

查看全文 查看/发表评论 下载pdf阅读器