基于径向基神经网络的舰船备件需求预测  

Ship's Spare Parts Forecast Based on RBF Nerve Network

作　　者：唐成[1] 盖强[2] 田峰[3] 刘勇[2] 黄俊添 

机构地区：[1]中国人民解放军海军大连舰艇学院训练舰支队,辽宁大连116018 [2]中国人民解放军海军大连舰艇学院舰炮系,辽宁大连116018 [3]中国人民解放军海军大连舰艇学院舰船指挥系,辽宁大连116018

出　　处：《科技视界》2013年第36期152-153,共2页Science & Technology Vision

摘　　要：针对舰船装备保障过程中备件数量的确定方法缺乏科学性和实际操作繁琐的现状,分析了影响舰船备件数量主要因素,研究了径向基神经网络的工作原理,提出基于径向基函数神经网络的舰船备件需求预测方法。最后给出了预测实例,并与BP神经网络预测结果进行对比。结果表明,径向基神经网络预测方法操作简单,预测结果符合实际情况,拟合效果优于BP。Due to the fact that in ship maintenance process, the method of determining the uumber of spare parts is not scientific and the actual operation is complicated, this paper analyzes the 4 major factors affecting the number of ship＇s spare parts （number of planned main operations, total times of disassembling in maintenance, accumulated working time, mean time between failures）, it also establishes the spare parts demand forecast model based on the affecting factors and RBF nerve network. Finally, the paper provides forecast examples and makes a comparison between the examples with the BP nerve network forecast results. It shows that the operation of radial basis function nerve network forecast is simple, the forecast result fits the actual situation and the fitting effect is better than BP.

关 键 词：备件需求预测 神经网络 精确保障 

分 类 号：TQ051.21[化学工程]