摘要
受拟态物理学方法的启发,就物理个体与理想粒子的特征异同问题,通过建立拟态物理学方法与基于种群优化算法的映射关系,设计出一种面向全局优化函数的拟态物理学算法框架.这是一种基于群体的随机优化算法,每个样本解被看作一个具有质量、速度和位置属性的物理个体,个体质量是用户定义的有关其目标适应值的函数,个体的适应值越好质量就越大,则个体间的虚拟作用力就越大.利用牛顿万有引力定律定义了个体之间的虚拟作用力,制定了个体之间的引?斥力规则,使得适应值较好个体吸引适应值较差个体,适应值较差个体排斥适应值较好个体,最好个体则不受其他个体的吸引或排斥.该方法利用这种引?斥力规则使得整个种群向更好的搜索区域移动.实验结果表明该算法的有效性.
Inspired by artificial physics（AP） approach,a framework of artificial physics optimization（APO） algorithm is presented to solve global optimization problem.Comparing the similarities and differences of physical individual and ideal particle,we construct a mapping between AP approach and a population-based optimization algorithm.APO algorithm is a population-based stochastic search method.In the framework,each sample point can be treated as a physical individual with the properties of mass,velocity and position.The mass of each individual corresponds to a user-defined function of the value of an objective function to be optimized.The better the objective function value,the bigger the mass,and then the higher the magnitude of attraction.The virtual forces among individuals are defined by Newton＇s gravity law and an attraction-repulsion rule is established among them,which makes the individual attract others with the worse fitness and repel others with the better fitness,and the individual with the best fitness attracts all the others,whereas it is never repelled or attracted by others.The attractive-repulsive rule can be treated as the search strategy in optimization algorithm which will lead the population to search the better fitness region of the problem.The simulation results indicate the validity of the approach.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2011年第5期848-854,共7页
Journal of Computer Research and Development
基金
山西省自然科学基金项目(2008011027-2)
关键词
群体智能
拟态物理学
全局优化算法
牛顿第二定律
虚拟力
swarm intelligence
physicomimetics
global optimization
Newton＇s second law
virtual force
