论文部分内容阅读
为了解决脉冲整形实验中经常碰到的遗传算法收敛速度慢,早熟等问题,我们对传统的遗传算法进行了几点改进,例如:将两个个体间的欧几里得距离作为判断是否进行交叉操作的判据之一,而不再仅仅依靠个体的适应度值(fitness),这样能有效地保持种群的基因多样性,提高交叉算子的效率;第二,引入多个交叉算子共同作用于种群.由于算子的组合效应,共同作用产生的子代适应度值要优于任何一个算子单独作用时产生的子代适应度值.因而可以产生更大的探索范围,防止算法收敛在某个局部最优解;第三,为了提高收敛速度,我们提出一种新的插值方式:非线性插值,即依据频谱的强度大小决定插值点的密度.我们初步将此改进算法应用到飞秒整形光路输出光的相位补偿实验中,得到了比较令人满意的结果.
In order to solve the problems of slow convergence and prematurity of genetic algorithms often encountered in pulse shaping experiments, we made some improvements to the traditional genetic algorithms. For example, we consider the Euclidean distance between two individuals as a criterion for determining whether to cross Instead of relying solely on the individual’s fitness, which can effectively maintain the genetic diversity of the population and improve the efficiency of the crossover operator. Second, the introduction of multiple crossover operators Due to the combined effect of operators, the fitness value of the offspring produced by the interaction is better than that of any offspring generated when any operator acts alone, so that a larger exploration range can be generated to prevent the algorithm from converging at We propose a new interpolation method: nonlinear interpolation, which determines the density of interpolation points according to the intensity of spectrum.We initially apply this improved algorithm to femtosecond In the experiment of phase compensation of the shaping light output light, satisfactory results were obtained.