Forecast nervousness in supply chains: Should we use consumer demand or retailer orders?

Li, Qinyun ORCID: https://orcid.org/0000-0001-5978-1903 and Disney, Stephen ORCID: https://orcid.org/0000-0003-2505-9271 2018. Forecast nervousness in supply chains: Should we use consumer demand or retailer orders? Presented at: OR60 Annual Conference of the Society of Operational Research, Lancaster, UK, 11-13 September 2018.

Full text not available from this repository.

Abstract

Retailers observe and satisfy consumer demand. They replenish their in-store inventory by placing orders onto manufacturers. They usually place firm orders to the manufacturer in the morning to be dispatched later that day. They also provide future predicted daily orders over the coming weeks; these are future order forecasts. We have recently delivered a measure of the accuracy of these order forecast streams that we call nervousness (Li and Disney, 2017). The nervousness measure places a weight on the variances of the n-period forecast errors to combine them into a single metric. The geometrically decreasing weight assigns more importance to short-term forecast errors (presumably they are more costly) than forecast errors in the distant future (likely they are easier to cope with). The natural question to ask is should the manufacturer use the retailer’s future order forecasts based on the end consumer demand? Or should the manufacturer ignore the retailer’s future order forecasts and create his forecast stream based on the retailer’s actual order history? We answer by considering a two-echelon order-up-to policy based supply chain. Both the retailer and the manufacturer use exponential smoothing to forecast. We investigate the case of general lead-times with a z-transform analysis. For i.i.d. demand we find exact closed-form expressions for the nervousness experienced by the manufacturer. We also obtain closed-form order and inventory variances at both echelons, both with and without information sharing. We also investigate how the nature of demand pattern influences these characteristics via a frequency response analysis.

Item Type:	Conference or Workshop Item - unpublished
Date Type:	Completion
Status:	Unpublished
Schools:	Schools > Business (Including Economics)
Related URLs:	Organisation
Last Modified:	25 Aug 2026 14:25
URI:	https://orca.cardiff.ac.uk/id/eprint/116382

Actions (repository staff only)

Edit Item