The role of forecasting parameters in reducing bullwhip effect

Abstract

This paper investigates how selecting appropriate forecasting parameters could be useful in reducing ordering variances (i.e. bullwhip effect) moving up stream a four-level supply chain. We examined 40 different scenarios for each echelon of supply chain in order to find governing rules in determination the best forecasting parameters. Relying on extensive spreadsheet computation we found some interesting results which need more investigations to be accepted as a general rule for all forecasting techniques. The results shown in tables and charts demonstrate that increasing number of periods used for initial average calculating in smoothing formula (T) decreases ordering variances. Moreover, increasing weighting factor () in exponential smoothing formula results in increasing variances of orders; that is, it results in intensified bullwhip effect. Regarding these two findings, we will show that simultaneous increase of forecasting parameters (T,) will alleviate ordering variances or bullwhip effect. The most important limitation of the research is that there are some informal local producers whose levels of production could not be exactly determined. We had to simply add their production variances to the demand variance of each echelon. However, designing computerized systems for archiving data during the time will be useful for finding the variances pattern. This investigation was carried out on the real supply chain of a dairy company (Mahghol Co.) and the results made good progress in improving supply chain management and inventory policy.

Bullwhip Effect in a Supply Chain with Trend Demand: An Agent-based Simulation View

“…This section describes and builds a four-echelon supply chain, which is a common multi-echelon model [13]. The supply chain contains six roles, i.e., an end customer, a retailer, a wholesaler, a distributor, a manufacturer and an out supplier.…”

“…It aims at warming up the simulation model quickly. (13) Then order information is passed to the upstream. This order quantity becomes the demand of an adjacent upstream player.…”