确定河流纵向离散系数的快速SA法

The Fast Simulated Annealing Algorithm to Determine the Longitude Dispersion Parameters of River

将模拟退火算法应用求解分析河流水团示踪试验数据,确定河流纵向离散系数的函数优化问题。针对一般模拟退火算法收敛性能差的缺陷,采取了改变模型扰动、接受概率、降温方式以及增加附加约束条件和增加记忆功能等综合措施对算法进行了改进。实例分析结果表明,改进后的模拟退火算法确定河流纵向离散系数的优化问题不仅可行,而且求解精度较高,更重要的是收敛速度比其它算法快的多,这在一些高维复杂的优化问题中很值得推广。

The function optimization problem for analyzing the data of river water mass tracing test, to determine the longitudinal dispersion parameters of river, is to be solved by simulated annealing algorithm. In view of the limitation of the traditional simulated annealing algorithm, adopted comprehensive measures such as changing the model disturbance, accept probability, temperature lowering way, adding supplement-restraint conditions to the variables and adding memory in inner loops were taken to improve the convergence of algorithm. Application of this method to the published data shows the later method is not only feasible for confirming the dispersion parameters, but also has higher accuracy. The most importance is that much faster than other algorithm methods. So it is worth to popularize for resolving some multidimensional optimization problems.