改进灰色马尔科夫模型及其在航空备件需求预测中的应用  

Improved Gray Markov Model for Demand Forecasting of Aviation Spare Parts

作　　者：廖乃智 涂继亮[1] 刘辉 邓泽平 赖国超 LIAO Naizhi;TU Jiiang;LIU Hui;DENG Zeping;LAI Guochao(College of Information Engineering,Nanchang Hanghong University,Nanchang 330063,China)

机构地区：[1]南昌航空大学信息工程学院,南昌330063

出　　处：《火力与指挥控制》2024年第12期55-61,共7页Fire Control & Command Control

基　　金：国家自然科学基金(72261027);南昌航空大学校级研究生创新专项基金资助项目(YC2023-026)。

摘　　要：针对传统灰色马尔科夫模型预测精度较低,没有量化区分各状态关联关系强弱的不足,提出一种改进的灰色马尔科夫航空备件预测模型。使用组合插值法对传统灰色模型背景值的求解进行改进,并且利用新陈代谢方法不断更新历史数据,使得模型的平均相对误差由原来的3.2%下降至2.17%,模型后验差比C为0.004,小误差概率ρ大于0.8,为良好等级,有较高的预测精度。再利用加权马尔科夫模型对新陈代谢模型的预测值进行修正,经修正后的2015年需求量预测相对误差从-4.04%下降到-2.82%。In response to the low prediction accuracy of the traditional Gray Markovmodel and its inability to quantify the strength of state relationships,an improved Gray Markov forecast-ing model for aviation spare parts is proposed.The model improves the solution of the traditional gray model's background value by usinga combination interpolation method.Additionally,the model continuously updates historical data through a metabolic method,reducing the averagerela-tive error from 3.2%to 2.17%.The model's posterior error coefficient(C)is 0.004.with a small error probability(p)greater than 0.8,indicating a good level with highprediction accuracy.Finally,the weighted Markov model is used to correct the forecasted values from the metabolic model.After correction,the relative error ofthe demand forecast for 2015 decreased from -4.04% to-2.82%.

关 键 词：航空备件 背景值改进 新陈代谢 灰色预测 加权马尔科夫 

分 类 号：V246[航空宇航科学与技术—飞行器设计]