遗传算法的单纯形交叉算子

Simplex Crossover Operator for Genetic Algorithm

提出一种新型的遗传算法交叉算子，即单纯形交叉算子。这种算子实现了遗传算法与单纯形算法之间的结合，它能加快遗传算法的寻优速度，提高遗传算法定位最优解的精度。通过动态调整该算子的调用概率，可以方便地控制它的使用。本文还提出了一种所谓的“淘汰选择”，这种操作使得操作算子生成的新解不再是简单地取代其亲生父代个体，而是取代父代群体中的劣质个体。测试的算例表明该算子效果显著。

A novel crossover operator, simplex crossover operator,is proposed for genetic algorithm. By integrating the conventional simplex method into genetic algorithm, the presented operator speeds up the genetic algorithm and improves its accuracy of locating the optimal solution. The probability of activating the operator provides flexibility in calling the conventional method. The paper also proposes an operation called 'Killing Selection'which enables genetic operators to substitute the children for the selected 'bad' solutions rather than the operand parent (s ).Test example are given to illustrate the performance of the present method.