一种改进的遗传算法及其在剩余静校正中的应用

本文综合利用模拟退火和遗传算法的特点，给出一种改进的遗传算法，即根据模拟退火的概率分布函数提供群体的初值，然后使用遗传算法的选择算子和杂交算子进行搜索，在此基础上，再降低温度提供新的初值，进行反复迭代，从而搜索到全局最优解，通过估计剩余静校正的模拟试验表明，该方法只需要较小的群体规律，与常规的遗传算法相比，没有其它更多的控制参数，且全局收敛的效率有了很大的提高。

An ameliorated genetic algorithm and its application to residual static correction

An ameliorated genetic algorithm has been developed by combining the advan-tages of simulated annealing and usual genetic algorithm. It works in the steps:Using the probability distribution function of simulated annealing to offer ini-tial group values, followed by the search that is done with the aid of selection andhybrid operators of the genetic algorithm..Lowering temperature to produce new initial values, followed by a new itera-tion operation to search out global optimal solution.The model experiment for estimating residual static correction values says thatthe method only needs quite small group scale, and that it has less control parame-ters and brings faster convergence than usual genetic algorithms.