摘要
针对传统备件消耗预测模型在组合赋权方面存在的缺陷,提出了一种基于改进Theil不等系数的诱导有序加权调和平均(induced ordered weighted harmonic averaging,IOWHA)算子和马尔科夫链(Markov chain,MC)的导弹备件消耗预测模型。首先运用最小二乘法、数据变换技术和加权理论对单项预测模型进行改进;然后建立基于Theil不等系数的IOWHA算子组合预测模型;提出利用MC定性推导出组合模型中各单项预测模型在待预测时点上的预测精度状态,进而应用遗传算法(genetic algorithm,GA)求解得到待预测时点上的组合模型权系数。实例结果表明,所提出的组合预测模型大大降低了预测误差。
To overcome the disadvantage of the traditional combination consumptive forecast model for spare parts in combination weighting, a combination consumptive forecast model of missile spare parts based on improved Theil coefficient induced ordered weighted harmonic averaging （IOWHA） operator and Markov chain （MC） is proposed. Firstly, the single item forecast method is improved with the least square model, date transformation technology and weighing theory. The combination forecast model based on Theil coefficient and IOWHA operator is established. By MC, the forecast accuracy condition of each single item forecast model at the forecast time point can be qualitatively surmised, thus its weighted coefficient at the forecast time point could be obtained by genetic algorithm （GA）. Experimental results show that the proposed combination forecast model significantly reduces forecast errors.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2013年第8期1681-1686,共6页
Systems Engineering and Electronics
基金
军队科研项目资助课题
关键词
组合预测
Theil不等系数
诱导有序加权调和平均算子
加权最小二乘支持向量机
遗传算法
马尔科夫链
备件
combination forecast
Theil coefficient operator
weighted least squares support vector machine （MC）
spare parts induced ordered weighted harmonic averaging （IOWHA） （WLS-SVM）
genetic algorithm （GA）
Markov chain
