摘要
为降低航空公司维修成本,增强送修等级决策科学性,保障飞行安全,提出基于PSO-SVM的民航发动机送修等级决策算法。首先利用改进的粒子群优化(Particle Swarm Optimiza-tion,PSO)算法对支持向量机(Support Vector Machine,SVM)参数进行寻优,并提出将交叉验证(Cross Validation,CV)的平均分类精度作为PSO的适应度值。对某型发动机送修等级的真实数据进行了决策对比研究,研究数据表明:与传统的Grid和GA算法相比,PSO的参数寻优效果要更优;在小样本分类时,PSO-SVM的分类精度要远高于常用的神经网络分类模型径向基函数(RadialBasis Function,RBF)模型和学习向量量化(Learning Vector Quantization)模型。
In order to reduce airline repairment cost,enhance the scientific nature of maintenance level decision and ensure flight safety,the aeroengine maintenance level decision algorithm based on the PSO-SVM was developped.The improved particle swarm optimization(PSO) was used to optimize parameters of support vector machine(SVM)and the average classified precision based on cross validation(CV) was used as PSO fitness value.The decision comparison study on the real data of engine maintenance level was carried out.The research data shows that the PSO parameter optimization is superior to the traditional Grid and GA optimization algorithm.In small sample classification,the PSO-SVM classified precision is better than that of neural network model RBF and LVQ.
出处
《推进技术》
EI
CAS
CSCD
北大核心
2013年第5期687-692,共6页
Journal of Propulsion Technology
基金
中国民航飞行学院青年基金项目(Q2010-67)
关键词
粒子群优化算法
支持向量机
交叉验证
送修等级决策
Particle swarm optimization
Support vector machine
Cross validation
Maintenance level decision