A reliability-based fuzzified model for automotive spare part forecasting

Due to the importance of spare parts forecasting for inventory management and its significant economic impacts, several studies were done to achieve an accurate demand estimation of spare parts. Most of these methods depend on historical data which usually are not available or experts’ judgments that produce biased and inconsistent results. In this study, a mathematical model of automotive spare part forecasting was proposed based on product reliability and the part failure rate. Fuzzy numbers were used to avoid the uncertainty of the reliability and failure rate parameters. Both triangular and trapezoidal fuzzy numbers were used. Four different defuzzification methods were used namely, weighted average, centroid, graded integration representation, and signed distance methods. Sensitivity analysis of both reliability and failure rate parameters was done to study their effects on the demand deviation. The fuzzifications of both part failure rate means and the product reliability mean have the largest effects on the demand deviation with demand deviation ۷.۵% and ۲۰%, respectively. The maximum demand deviation reaches around ۲۲.۵% by applying the centroid method at the latest quarter of lifetime under fuzzifications of all reliability and failure rate parameters. Accurate results are obtained by using trapezoidal fuzzy numbers compared with Triangular Fuzzy Number (TFNs)..

کلیدواژه ها:

Spare parts ، Forecasting Methods ، Reliability ، failure rate ، Fuzzy number