| Summary: | 碩士 === 國防管理學院 === 資源管理研究所 === 94 === Weapon equipment is an important condition of national defence security and combat effectiveness. Therefore, weapon system must maintain well arranged situation frequently. The spare parts is an important turning factor to maintain the appropriate condition of weapon. However, how many quantity to store can fit both economic benefits and combat readiness demanding simultaneously is already becoming an very important issue of military.
The number of the spare parts in weapon equipment are countless, and most of them belong to precision instruments. The troops usually use the historical parameters of demanding time and quantity as the factor to evaluate the demanding quantity on the spare parts. However, the demanding quantity of the spare parts often possess characteristics of indefinite. They involve several affective factors. Therefore, the actual and predictive value always different if the key factor does not bring into consider. This research takes air defence missile system as the target will thinks over some integrate logistic informations as the key factor, such as, the Mean Time between Failure (MTBF), Mean Time between Maintance Action, (MTBMA), Mean Repair Time (MRT), Inherent Availability (IA), and Operational Availability (OA), etc. This research will apply the ability of Artificial Neural Network (ANN) to establish a model which can forecast the demanding of the spare parts. By the way, the result of the model will compare with some forecast methods that the troops often used to apply, are Multiple Regression analysis, Moving Average method, and Exponential Smooth method.
The result of this research shows that the forecast of spare parts is a nonlinear problem, and the effect of the Back Propagation Network (BPN) Model is better than the other kinds of forecasting methods.
|
|---|
