基于改进云遗传算法的多跳自组网故障诊断

Fault Diagnosis of Multiple Hops Self-organizing Networks Based on Support Vector Machine Optimized Cloud Genetic Algorithm

通过收集多跳自组织网络下无线传感器故障历史数据,然后利用支持向量机对故障类型与特征之间关系进行建模,支持向量机参数通过遗传算法优化,同时利用云计算对遗传算法性能进行改善,防止神经网络训练时出现收敛速度慢和陷入局部极小等缺陷。仿真实验表明,相对于其它多跳自组织网络下无线传感器故障诊断模型,CGA-SVM提高了多跳自组织网络下无线传感器故障诊断正确率,能够满足多跳自组织网络下无线传感器故障诊断的要求。

The historical data multiple hops self-organizing networks wireless sensor fault are collected, and then the model is built by support vector machine to describe the relation among fault types and features of faults which the parametersare optimized by genetic algorithm, at the same time cloud model is used to optimize the performance of improved geneticalgorithm, lastly, the performance of the model is test. The results show that CGA-SVM improves the accuracy of fault diagnosis compared with other aviation engine fault diagnosis models, and it can meet the requirements of aero engine fault diagnosis.