摘要
引入一种控制参数少、寻优机制强的人工蜂群算法(ABC),以红松人工林为例基于净现值(NPV)最大为目标优化林分经营措施,并同Hooke&Jeeves直接搜索算法、差分进化算法(DE)、进化策略算法(ES)和粒子群优化算法(PSO)进行对比评估,探讨ABC算法参数配置及各算法特性。通过模拟器推演标准红松人工林的生长及经营过程,以NPV为经营目标,遍历ABC算法参数组合,确定最优参数。结果表明:根据ABC算法参数寻优结果显示,随着蜂群规模大小增加NPV呈上升趋势,当蜂群大小为90时NPV均高于385500元/hm^(2)。NPV均值的排序为PSO>ABC>DE>ES>HJ,变异系数的排序为DE<PSO<ABC<ES<HJ;当仅将群体大小减少到5,而其他参数保持最优时,NPV均值排序为ABC>DE>PSO>ES,变异系数排序为DE<ABC<PSO<ES。本研究系统评估了5种林分经营优化算法在最优参数配置下的性能,整体上,DE、PSO和ABC算法均表现优异且能维持候选解的多样性,在处理复杂优化问题时,ABC算法的执行效率颇具优势。通过对比,本研究评估了5种算法优化经营措施的可行性,为ABC算法在林分经营优化中的应用提供了参考。
An artificial bee colony algorithm(ABC)with few control parameters and strong optimization mechanism was introduced to optimize stand management measures based on the maximum net present value(NPV),the parameters configuration and characteristics of ABC algorithm are also discussed by comparing with Hooke&Jeeves direct search algorithm,differential evolution algorithm(DE),evolutionary strategy algorithm(ES)and particle swarm optimization algorithm(PSO).Through simulating the growth and management processes of an initial Pinus koraiensis plantation and optimizing its management schedule,focusing on NPV as the objective,the parameters of the ABC algorithm were systematically explored to determine their optimal configuration.The results showed that NPV increases with the increase of the swarm size.When swarm size was 90,NPV consistently exceeded 385,500 yuan/ha.Comparative results for different algorithms revealed that,according to the mean NPV of repeated optimization runs,the ranking of the algorithms was PSO>ABC>DE>ES>HJ.According to the coefficient of variation,the ranking was DE(best)<PSO<ABC<ES<HJ.When the population size(swarm size)was reduced to 5 while the other parameters were kept in near-optimal values,the ranking of the population-based algorithms in terms of mean NPV was ABC>DE>PSO>ES.In terms of the coefficient of variation,the ranking was DE<ABC<PSO<ES.Overall,DE,PSO,and ABC algorithms exhibited good performance and robustness as well as the ability to maintain diversity among the population of the solutions.However,in handling complex optimization problems,the ABC algorithm demonstrates superior execution efficiency.By comparison,this study evaluated the feasibility of five algorithms to optimize management measures,and provided scientific support for the application of ABC algorithm in stand management optimization.
作者
罗隽泳
金星姬
Timo Pukkala
郝元朔
Luo Junyong;Jin Xingji;Timo Pukkala;Hao Yuanshuo(Key Laboratory of Sustainable Forest Ecosystem Management-Ministry of Education,College of Forestry,Northeast Forestry University,Harbin Heilongjiang 150040,China;School of Forest Sciences,University of Eastern Finland,Joensuu 80101,Finland)
出处
《西南林业大学学报(自然科学)》
北大核心
2025年第3期184-193,共10页
Journal of Southwest Forestry University:Natural Sciences
基金
国家自然科学基金面上项目(32071758)资助
中央高校基本科研业务费专项资金项目(2572020BA01)资助。
关键词
红松
人工林
经营优化
人工蜂群算法
直接搜索算法
Pinus koraiensis
plantation
management optimization
artificial bee colony algorithm
direct search algorithm
