模拟退火粒子群与小波的地基沉降预测应用

针对粒子群优化算法易陷入局部极小值问题,改进学习因子使其自适应调整,并与具有良好全局搜索能力的模拟退火算法结合,充分利用两种算法各自的优点,同时结合小波分析去噪,优化神经网络参数,对地基累计沉降数据进行预测,并与标准粒子群优化算法做了对比,实验表明两种方法的结合具有良好的全局和局部搜索能力,预测精度高。

Forecast application of foundation settlement using simulated annealing particle swarm and wavelet algorithm

In dealing with the problem that particle swarm optimization (PSO) easily falls into local optima,the learn factor was adjusted adaptively firstly in the paper.Secondly the authors combined with the simulated annealing algorithm which has a better global searching ability to improve the global and local searching ability.Thirdly this composite optimization algorithm with wavelet was used to optimize neural network parameter to forecast the foundation settlement.Then a comparison with the standard particle swarm optimization algorithm was made.Finally an experiment indicated that this composite optimization algorithm had good ability to global and local search the optima with high predictive precision.