柔性多面体在遗传算法中的应用研究(I)  

Study on Application of Flexible Polyhedron to Genetic Algorithm(Ⅰ)

作　　者：肖宏峰[1] 梁丰[2] 谭冠政[2] 

机构地区：[1]湖南师范大学计算机教学部,长沙410081 [2]中南大学信息科学与工程学院,长沙410083

出　　处：《计算机工程与应用》2003年第4期83-85,111,共4页Computer Engineering and Applications

基　　金：中国科学院机器人学开放研究实验室资助(编号:RL200002)

摘　　要：文章提出了一种柔性多面体的方向进化算子,并在基本遗传算法中嵌入柔性多面体搜索算法,从而构成了一种基于柔性多面体的新的混合遗传算法(flexiblepolyhedronhybridgeneticalgorithm,FP_HGA)。方向进化算子紧跟基本遗传算法的变异操作之后,其作用是使适应度较低的个体向适应度较高的个体进化;柔性多面体局部搜索算法作用是对当前代所有新个体在进入到下一代之前,使它移动到局部最优点。并用FP_HGA来求解Rosenbrock测试函数的最小值,FP_HGA算法和SGA(SimpleGeneticAlgorithm,SGA)算法的计算结果表明该混合遗传算法在收敛速度和精度方面均得到很大提高。In this paper,a directional evolution operator based on flexible polyhedron and a new hybrid genetic algo-rithm based on the flexible polyhedron search algorithms and the directional evolution operator are proposed,the new hybrid genetic algorithms is called FP_HGA(flexible polyhedron hybrid genetic algorithm,FP_HGA).The function of the directional evolution operator is to evolve the worse individuals to the better individuals,and the function of the flexible polyhedron search algorithms is to move the new individuals obtained from crossover operator to local optimum before they enter the next generation.The FP_HGA is used to obtain the minimum of experiment function Rosenbrock,the re sults from FP_HGA and SGA(Simple Genetic Algorithm,SGA)indicate that the convergence speed and accurate of FP_HGA are enhanced more greatly than those of SGA.

关 键 词：柔性多面体 遗传算法 随机搜索算法 进化算子 

分 类 号：O242.23[理学—计算数学]