基于免疫SVM的机械产品寿命预测  

Life prediction for mechanical products based on immune SVM

作　　者：林立忠[1] 刘智国[1] 张雅明[2] 

机构地区：[1]石家庄学院,石家庄050035 [2]石家庄信息工程职业学院,石家庄050035

出　　处：《现代制造工程》2012年第6期104-107,共4页Modern Manufacturing Engineering

摘　　要：传统机械产品寿命预测方法成本较高,精度较低,难以满足机械产品寿命预测要求。为了改变这种状况,提出基于人工免疫算法优化支持向量机(免疫SVM)回归器的机械产品寿命预测技术,采用人工免疫算法进行支持向量机回归器参数选取。人工免疫算法具有良好的全局搜索能力,能较好地选择合理的支持向量机回归器参数。采用普通SVM预测方法与免疫SVM预测模型进行比较,试验结果表明,相比于普通SVM,免疫SVM具有更高的机械产品预测性能。Generally, the high cost and the low prediction accuracies of traditional prediction methods for mechanical products are difficult to meet the requirement of life prediction for mechanical products. In order to change the situation, Support Vector Ma- chine（SVM） optimized by artificial immune algorithm is presented to life prediction for mechanical products ,where artificial im- mune algorithm is applied to choose the parameters of support vector machine. Artificial immune algorithm has strong global search ability ,which can choose the parameters of support vector machine classifier effectively. Normal SVM are applied to com- pare with the proposed method. The experimental results demonstrate that the proposed method has higher life prediction accuracy for mechanical products than normal SVM.

关 键 词：机械产品 免疫算法 支持向量机 寿命 预测 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]