正交遗传算法

Orthogonal Genetic Algorithm

本文提出正交遗传算法 .该算法把正交设计与遗传算法结合起来 ,用正交设计确定区间水平 ,在每个区间水平按遗传算法进行搜索 ,正交设计能够发现包含最好解的区间水平 ,再在这个区间水平使用遗传算法进行搜索 .反复使用上述算法 ,当前区间水平变小 .最后 ,当前区间水平变为一点了 ,它就是这些区间水平的最优点 .借助正交设计 ,本文证明了正交遗传算法是收敛的和进行了计算机模拟 .模拟结果显示该算法是有效的和可行的 .

The paper proposes an Orthogonal Genetic Algorithm(OGA) that combines Genetic Algorithm(GA) with Orthogonal Design.Orthogonal Design determines interval level,GA is performed in the interval level.The Orthogonal Design can find an interval that contains the best solution,GA is performed again in the interval level,The Orthogonal Design can again find an interval that contains the best solution,GA is performed again in the interval level,the interval level gets smaller and smaller.At last,it gets one point.The point is expected optimal point of interval levels.The paper has proved that GA is able to converge to the global optimum.Computer simulation is performed,the result shows that OGA is efficient and feasible.