快速模拟退火算法在含水层参数识别中的应用

针对一般模拟退火算法收敛速度慢的缺陷,本文提出了含水层参数识别的快速模拟退火算法.它以广义Boltzmann-Gibbs统计理论为基础,采用依赖于温度的似Cauchy分布产生新的扰动模型,通过改变模型扰动、接收概率和降温方式来加快模拟退火算法的收敛速度.实例的计算结果表明,快速模拟退火算法不仅可行,而且求解精度高,收敛速度比其他算法快得多,是一种值得在实际中推广应用的高效的含水层参数识别方法.

Application of fast simulated annealing algorithm in aquifer parameter identification

With the hmit of the shortcomings of the slow convergence of the common simulated annealing algorithm, the fast simulated annealing algorithm is proposed, which is based on the generalized Boltzmann-Gibbs statistics and the turbulent model derived from temperature dependent pseudo Cauchy distribution. It speeds up the convergence rate of simulated annealing algorithm by changing the model disturbance, acceptance probability and temperature lowering way. Apphcation of this method to the pubhshed data shows that the result is high accuracy, and the speed is much higher than any other methods. Hence, the fast simulated annealing algorithm is a high efficient method which is worth to be apphed to the identification of aquifer parameter.