Date of Award
Master of Science
Industrial and Manufacturing Systems Engineering
The effective management of returned products is enhanced by reliable forecasts of returns. We explore a method of forecasting the returns in an electronic product remanufacturing environment under the assumptions that each item will be returned and the movement of each individual item can be tracked. We investigate the benefit of information from early returns on the predictability of future return times under a variety of conditions described by the variability in the sale time and the mean time to return. We also explore the effects of prior knowledge of the return time distribution on the precision of estimates. We estimate the parameters of the distribution of the time until items are returned using maximum likelihood estimation and determine approximate confidence intervals for the estimates. The approximate confidence intervals are also derived for both the mean time until items are returned and the proportion of items that will be returned before that specified product is obsolete. The method is illustrated under two scenarios depending on the prior knowledge about the return time distribution. The results show that prior knowledge about the return time distribution contributes greatly to the precision in estimating and that the widths of the confidence intervals for the estimates are also affected by the variability in sales time distribution. However, even without prior knowledge about return distribution characteristics, both the mean of the return time distribution and the proportion of items that will be returned before obsolescence can by estimated satisfactorily.
Chittamvanich, Suphalat, "Using data on early returns of electronic products to forecast future availability" (2003). Retrospective Theses and Dissertations. 20053.