模拟退火法和遗传算法联合优化技术及在反演解释中的应用

结合遗传算法和模拟退火法逃逸局部极值，应用全局寻优的非线性优化算法，改进了传统基于梯度法的局部搜索算法的缺陷。同时，为了改进全局搜索法计算时间长的缺点而加快了收敛速度，在全局搜索法找到全局极值所在的山谷时，改用共轭梯度法快速到达谷底，运用笔者研究的方法对张集煤田的实际资料进行了反演处理，取得了比较好的结果。

Joint optimization technique combining simulated annealing with genetic algorithm and its application in inversion interpretation

In order to escape from local minima, genetic algorithm and simulated annealing are combined. Authors apply global optimization of non-linear algorithm and correct the defect of local search method that is based on traditional gradient algorithm. In the meantime, in order to decrease the compute time of global search and accelerate convergence, authors apply the conjugate gradient to move to the bottom of the valley when the global-seeking method finds the valley of global minimum. Finally, authors apply the studied global technique to inverse the actual data of coal-field and obtain good results.