A Data-Driven Newsvendor Problem: From Data to Decision

27 Pages Posted: 26 Dec 2017

See all articles by Jakob Huber

Jakob Huber

University of Mannheim - Data and Web Science Group

Sebastian Müller

University of Mannheim, Business School

Moritz Fleischmann

University of Mannheim, Business School

Heiner Stuckenschmidt

University of Mannheim - Data and Web Science Group

Date Written: December 8, 2017

Abstract

Retailers that order perishable items are required to make ordering decisions for hundreds of products on a daily basis. This task is non-trivial because the risk of ordering too much or too little is associated with overstocking costs and unsatisfied customers. Traditionally, this problem is solved in a two-step procedure. First, the parameters of a given demand distribution are estimated, and second, an optimization problem based on this distribution is solved to obtain the order quantity. However, in reality, the true demand distribution is almost never known to the decision maker. Therefore, we present a novel solution method based on Artificial Neural Networks and Quantile Regression that does not require the assumption of a specifc demand distribution. We provide an empirical evaluation of our method with point-of-sales data for a large German bakery chain. We find that our method outperforms well-established standard approaches in most cases.

Keywords: inventory, newsvendor, artifcial neural networks, quantile regression

Suggested Citation

Huber, Jakob and Müller, Sebastian and Fleischmann, Moritz and Stuckenschmidt, Heiner, A Data-Driven Newsvendor Problem: From Data to Decision (December 8, 2017). Available at SSRN: https://ssrn.com/abstract=3090901 or http://dx.doi.org/10.2139/ssrn.3090901

Jakob Huber (Contact Author)

University of Mannheim - Data and Web Science Group ( email )

L 5, 2 - 2. OG
68161 Mannheim
Germany

Sebastian Müller

University of Mannheim, Business School ( email )

University of Mannheim, Business School
P.O. Box 10 34 62
Mannheim, 68131
Germany

Moritz Fleischmann

University of Mannheim, Business School ( email )

University of Mannheim
P.O. Box 10 34 62
Mannheim, 68131
Germany

Heiner Stuckenschmidt

University of Mannheim - Data and Web Science Group ( email )

L 5, 2 - 2. OG
68161 Mannheim
Germany

Register to save articles to
your library

Register

Paper statistics

Downloads
105
rank
245,850
Abstract Views
374
PlumX Metrics