摘要
为了解决目标函数中含有sin、cos等周期函数的优化问题,基于生态系统循环食物链思想提出了一种新型函数优化算法,即AFC-ASO算法。在该算法中,假设在生态系统中的某个循环食物链系统中生活有多种不同类型的动物,这些不同类型的动物采取循环食物链的方式维持该生态系统的生态平衡。进食的方法是采用攫取食饵动物部分器官或吸取其体内物质的方式,但不会危及食饵动物的生命;同类型的动物分雌、雄两种性别。每种类型的动物在该生态系统中活动时,具有捕食、交配、集群、逃逸、游弋五种行为,依据这五种行为构造出了相关的演化算子。其中,捕食算子能够使得个体器官间交换信息;交配算子能使强壮个体将其优良信息传给虚弱个体;集群算子能使个体摆脱局部最优解陷阱;避险算子能增强个体之间的分散度;闲逛算子可以增加当前个体的活跃度;生长算子能确保该算法具有全局收敛性。结果表明,算法对求解某些类型的复杂函数优化问题,特别是目标函数中含有sin、cos等周期函数的一类复杂函数优化问题,具有较高的适应性和收敛速度。
In order to solve complicated function optimization problems whose objective function includes some periodic func- tions like sine, cosine and so on, this paper constructed a new optimization algorithm based on autonomic behaviours and inter- actions among animals in a cyclic food chain ecosystem, namely artificial food chain-based animal swarm optimization （AFC- ASO）. The algorithm supposed that many different classes of animals existed in an ecosystem, these animals took a way of cy- clic food chain to maintain ecosystem balance. For a predator animal, the method of taking food was to bite its prey animal＇ s organ or absorb liquid materials from its prey animal＇ s body, but it didn＇ t injure the prey animal＇ s life, animals within the same class were differentiated by male or female sex. When an animal cruised in the ecosystem, it manifested the preying, mating, gathering, escaping and sauntering instincts, these instincts of animals were constructed into the preying, mating, gathering, escaping and sauntering operator in AFC-ASO. The preying operator enabled animals to exchange information a- mong different organs, the mating operator enabled a strong animal to transfer its good features to weak animals, the gathering operator enabled an animal to escape from local optima, the escaping operator could enhance the dispersion among animals, the sauntering operator enabled an animal＇ s vitality to increase, the growth operator ensured the algorithm to globally con- verge. Results show that the algorithm has characteristics of strong search capability and high adaptability for the complicated functions optimization problems whose objective function includes some periodic functions like sine, cosine and so on, espe- ciallv for some high dimensional ontimization problems.
出处
《计算机应用研究》
CSCD
北大核心
2014年第9期2673-2680,共8页
Application Research of Computers
基金
陕西省科学技术研究发展计划资助项目(2013K11-17)
陕西省重点学科建设专项资金资助项目(E08001)
关键词
函数优化
群智能优化计算
食物链
人工动物群
function optimization
population-based intelligent optimization computation
food chain
artificial animal swarm
