Prescriptive Analytics for Flexible Capacity Management

46 Pages Posted: 3 Jun 2019

See all articles by Pascal M. Notz

Pascal M. Notz

University of Würzburg - Business Administration & Economics

Richard Pibernik

University of Würzburg - Business Administration & Economics

Date Written: April 2019

Abstract

Motivated by the real-world problem of a logistics company, this paper proposes and studies two novel data-driven, distribution-free prescriptive analytics approaches to solve a complex two-stage capacity planning problem with multivariate demand and vector-valued capacity decisions. Our approaches use integrated machine learning algorithms to prescribe capacities directly from historical demand and numerous features (co-variates) without having to make assumptions about the underlying multivariate demand distribution. We provide extensive analytical insights for both approaches and derive out-of-sample performance guarantees and proofs of consistency and convergence for our approach that relies on the wellestablished machine learning principle of empirical risk minimization (ERM). We demonstrate the applicability of both approaches to a real-world planning problem and evaluate their performance relative to traditional approaches (using estimated multivariate demand distributions) and a conventional data-driven approach (sample average approximation). Our results suggest that using our prescriptive analytics approaches can result in substantial performance improvements. Moreover, the results of additional numerical analyses demonstrate that, in our specific case, the prescriptive approaches are much more robust to variations of exogenous cost parameters than traditional approaches are. Based on a discussion of our numerical results, we offer detailed explanations for this attractive property of our novel
approaches.

Keywords: Prescriptive Analytics, Machine Learning, Data-driven OM, Capacity Management

Suggested Citation

Notz, Pascal M. and Pibernik, Richard, Prescriptive Analytics for Flexible Capacity Management (April 2019). Available at SSRN: https://ssrn.com/abstract=3387866 or http://dx.doi.org/10.2139/ssrn.3387866

Pascal M. Notz (Contact Author)

University of Würzburg - Business Administration & Economics ( email )

Sanderring 2
Wuerzburg, D-97070
Germany

Richard Pibernik

University of Würzburg - Business Administration & Economics ( email )

Sanderring 2
Wuerzburg, D-97070
Germany

Register to save articles to
your library

Register

Paper statistics

Downloads
43
Abstract Views
210
PlumX Metrics