Inventory survival analysis for semiconductor memory manufacturing

Abstract

The high variety of and intermittent demand for semiconductor memory products frequently limits the use of forecast error normalization in estimating inventory. Inventory turnover is a practical performance indicator that is used to calculate the number of days for which a company retains inventory before selling a product. Although previous studies on inventory level settings have primarily applied information regarding demand variability and forecast error, few studies have investigated the inventory turnover for inventory decisions. Inventory turnover data are time scaled, suited for a small sample, and right censored to fit the input of survival analysis. In this study, a model in which inventory turnover and survival analysis were integrated was developed to estimate the production inventory survival function used to determine inventory level. Data analysis results based on real settings indicated the viability of using inventory survival analysis to determine semiconductor memory inventory level settings.