Mining Optimal Policies: A Pattern Recognition Approach to Model Analysis

44 Pages Posted: 27 Nov 2017 Last revised: 22 Aug 2018

See all articles by Fernanda Bravo

Fernanda Bravo

University of California, Los Angeles (UCLA) - Anderson School of Management

Yaron Shaposhnik

University of Rochester - Simon Business School

Date Written: August 19, 2018

Abstract

This paper studies the application of Machine Learning (ML) for discovering structural properties of optimal policies in numerically obtained solutions to optimization problems. We propose a framework based on ML for conducting model analysis in a systematic way, which complements theoretical and numerical methods. As a proof of concept, we apply the framework to core operations problems, such as inventory management, queuing admission control, multi-armed bandit (MAB), and revenue management problems. We demonstrate how this approach can be used to identify optimal threshold-based policies (inventory management and admission control) and index policies (MAB), as well as for developing new heuristics for revenue management problems. For the MAB problem, our approach leads to a new efficient algorithm for computing optimal index policies. The main contribution of this work is methodological, in proposing and demonstrating the potential of using ML algorithms to analyzing optimization problems and devising interpretable policies.

Keywords: Mathematical modeling, Machine learning, Dynamic programming, Interpretability

JEL Classification: C44

Suggested Citation

Bravo, Fernanda and Shaposhnik, Yaron, Mining Optimal Policies: A Pattern Recognition Approach to Model Analysis (August 19, 2018). Available at SSRN: https://ssrn.com/abstract=3069690 or http://dx.doi.org/10.2139/ssrn.3069690

Fernanda Bravo

University of California, Los Angeles (UCLA) - Anderson School of Management ( email )

110 Westwood Plaza
Los Angeles, CA 90095-1481
United States

Yaron Shaposhnik (Contact Author)

University of Rochester - Simon Business School ( email )

Rochester, NY 14627
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
341
Abstract Views
1,661
Rank
180,546
PlumX Metrics