基于PSO的Kriging相关模型参数优化

The Optimization of Parameters of Kriging Correlation Model Based on Particle Swarm Optimization

Kriging插值计算过程中的相关模型参数确定是构造回归模型的关键,常用的模式搜索方法求解相关模型参数时存在计算精度依赖搜索起始点的缺点,从而导致最优解的不稳定.利用一种改进的二进制编码微粒群算法(Genetic Particle Swarm Optimization,GPSO)来求解相关函数的参数,该方法采用动态选择和调整变异算子概率的策略,克服了参数优化过程中对初始点设置的依赖问题,函数测试的性能比较表明该方法具有良好的收敛速度和稳定性.

In Kriging interpolation algorithm, the determination of parameters of a correlation model is a key issue to con- struct the regression model, the computing accuracy rely on the initialization when using the conventional numerical opti- mization methods such as pattern search method when computing the parameters, which result in the instability of the op- timization solution. In this paper, a modified binary particle swarm optimization （GPSO）, independent of the initializa- tion of parameters, is presented to compute the global optimums o[ correlation model parameters. In the GPSO, the prob- abilities of combination and mutation are dynamically adjusted result in faster convergence, the experimental results show that the proposed GPSO significantly improves the search efficacy and stability compared to the exiting optimization algo- rithm.