摘要
为求解复杂函数优化问题,基于人类记忆原理和人际关系,提出了一种新型函数优化方法,即MP-IRO算法。在该算法中,将个体分为恋人、知己、敌人、小人、陌生人5种对象类型,对应于恋爱、聚集、攻击、排斥、防御等5种行为,并构造相应的演化算子。恋爱算子能优先选择拥有长时记忆的个体,分享其表征特性;聚集算子能使个体摆脱局部最优解的陷阱;攻击算子能使个体之间活跃度增强;排斥算子能让个体远离瞬时记忆试探解方向,扩大搜索范围;防御算子能增加随机性。测试结果表明,本算法对求解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 the complicated function optimization problems, a new optimization algorithm was constructed based on the memory principles (MP) and interpersonal relationship (IR), namely memory principles-based interpersonal relationship optimization (MP-IRO). There are five object types in interpersonal relationship system, including lovers, friends, enemies, villains and strangers which successively correspond to loving, gathering, attacking, rejecting, defending instincts and five operators; the loving operator gives preference to individuals which memory type is long memory(LM) and shares CP with other individuals; the gathering operator could make individuals avert local optima; the attacking operator enables an individual's vitality to increase; the rejecting operator can be far from instantaneous memory to expand the search coverage; the defending operator increases the randomness. Results show that the algorithm has characteristics of strong search capability and high adaptability for the complicated function optimization problems.
作者
苏佳
黄光球
Su Jia Huang Guangqiu(Management College, Xi'an University of Architecture & Technology, Xi'an 710055, China)
出处
《系统仿真学报》
CAS
CSCD
北大核心
2017年第2期309-318,共10页
Journal of System Simulation
基金
自然科学基础研究计划(2015JZ010)
西安市科协决策咨询课题(201517)
陕西省社科基金(2014P07)
关键词
函数优化
群智能优化计算
记忆原理
人际关系
MP-IRO算法
function optimization
population-based intelligent optimization computation
memory principles
interpersonal relationship
MP-IRO algorithm