Promotion Optimization for Multiple Items in Supermarkets
Name
SSRN-id3061451.pdf
Description
Submitted version
Size
611.83 KB
Format
Adobe PDF
Checksum (MD5)
073b2c2ff70b5c724c60e3ef3efe319d
Author(s)
Kalas, Jeremy J.
Perakis, Georgia
Date Issued
June 2020
Journal
Management Science
Publisher
Institute for Operations Research and the Management Sciences (INFORMS)
Citation
Duguay, Raphael et al. “Promotion Optimization for Multiple Items in Supermarkets.” Management Science (June 2020) © 2020 The Author(s)
Version
Original manuscript
Abstract
Promotions are a critical decision for supermarket managers, who must decide the price promotions for a large number of items. Retailers often use promotions to boost the sales of the different items by leveraging the cross-item effects. We formulate the promotion optimization problem for multiple items as a nonlinear integer program. Our formulation includes several business rules as constraints. Our demand models can be estimated from data and capture the postpromotion dip effect and cross-item effects (substitution and complementarity). Because demand functions are typically nonlinear, the exact formulation is intractable. To address this issue, we propose a general class of integer programming approximations. For demand models with additive cross-item effects, we prove that it is sufficient to account for unilateral and pairwise contributions and derive parametric bounds on the performance of the approximation. We also show that the unconstrained problem can be solved efficiently via a linear program when items are substitutable and the price set has two values. For more general cases, we develop efficient rounding schemes to obtain an integer solution. We conclude by testing our method on realistic instances and convey the potential practical impact for retailers.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Sloan School of Management
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1287/MNSC.2020.3641
