人工免疫算法优化双支持向量机在拱坝变形预测中的应用

为了能够通过监测数据直观反映出坝体是否处于稳定运行状态,采用人工免疫算法优化的双支持向量机方法,对高拱坝变形数据进行了拟合预测分析,双支持向量机与标准支持向量机相比极大地提高了计算速度,在进行批量重复计算中计算效率明显提升。针对双支持向量机计算结果受参数影响较大且参数多的问题,引入人工免疫算法搜寻双支持向量机参数,人工免疫算法在遗传算法的基础上保留了一定数量的较优解,提升了算法的搜索效率。工程算例分析表明,参数对双支持向量机结果影响较大,通过人工免疫算法搜寻最优参数后,双支持向量机能够较好地拟合拱坝坝体变形数据,预测结果符合工程精度要求,最大误差仅为1 mm左右。

Predicting Arch Dam Displacement Using Twin Support Vector Machine Optimized by Artificial Immune Algorithm

In this paper the twin support vector machine optimized by immune algorithm is proposed to analyze the displacement of arch dam and predict the deformation of the dam. Compared with support vector machine, twin support vector machine greatly improves the calculation speed and the calculation efficiency in batch repetitive calculation. Considering the effect of parameters on the fitting results, the artificial immune algorithm is incorporated in the twin support vector machine to optimize the parameters. Artificial immune algorithm retains a certain number of better solutions based on the genetic algorithm and improves the search efficiency of the algorithm. Engineering example analysis shows that parameters have great influence on the results of the twin support vector machine. After searching the optimal parameters by using artificial immune algorithm, the twin support vector machine can better fit the deformation data of the dam. The prediction result meets requirements, and the maximum error is merely 1 mm.