摘要
备件需求量的预测是备件配置的重要内容,针对当前装备备件需求非稳态的特点,提出一种基于变分模态分解的备件需求预测方法。运用变分模态分解将非稳态备件需求序列分解为若干模态分量,引入模糊熵的概念,将周期性、随机性和长期性特征明显的模态分量进行有效聚合,提高计算效率,进而运用预测效果较好的径向基神经网络预测法对聚合后的模态分量分别进行预测,将各分量预测结果进行整合形成最终的备件需求预测值。通过案例分析与实验对比,结果表明提出的方法能够有效挖掘非稳态备件需求序列的深层次信息,实现非稳态备件需求序列的较好拟合,并与其他非稳态时间序列预测方法对比具有较高的预测精度,为适应新时代实战实训背景下备件需求的特点提供了有效的方法支撑。
The prediction of spare parts demand is an important part of spare parts allocation.According to the non-steady characteristics of current spare parts demand,a prediction method of spare parts demand based on variational mode decomposition is proposed.Firstly,the demand sequence of non-steady state spare parts is decomposed into several modal components by variational modal decomposition,and then the concept of fuzzy entropy is introduced to effectively aggregate the modal components with obvious periodicity,randomness and long-term characteristics,in order to improve the efficiency of calculation,radial basis function neural network prediction method is used to predict the aggregated modal components respectively,and then the predicted results of each component are integrated to form the final demand predict value of spare parts.Through case analysis and experiment comparison,the results show that the method proposed in this paper can effectively mine the deep level information of the non-steady state spare parts demand sequences,and can realize the better fitting of the non-steady state spare parts demand sequences,compared with other non-steady state sequence prediction methods,it provides an effective method support for be adapted to the characteristics of spare parts demand in the new era.
作者
秦海峰
杨超
侯兴明
徐庆尧
侯翔
QIN Hai-feng;YANG Chao;HOU Xing-ming;XU Qing-yao;HOU Xiang(Space Engineering University,Beijing 101416,China;NCO Institute of Army Armored Forces Academy,Changchun 130117,China)
出处
《火力与指挥控制》
CSCD
北大核心
2021年第11期99-105,共7页
Fire Control & Command Control
基金
全军军事类研究生重点资助课题(JY2018B141)。
关键词
变分模态分解
非稳态
备件
需求
预测
variational mode decomposition
non-steady state
spare parts
demand
prediction
