TY - JOUR
T1 - Efficient learning of decision-making models
T2 - A penalty block coordinate descent algorithm for data-driven inverse optimization
AU - Gupta, Rishabh
AU - Zhang, Qi
N1 - Publisher Copyright:
© 2022 Elsevier Ltd
PY - 2023/2
Y1 - 2023/2
N2 - Decision-making problems are commonly formulated as optimization problems, which are then solved to make optimal decisions. In this work, we consider the inverse problem where we use prior decision data to uncover the underlying decision-making process in the form of a mathematical optimization model. This statistical learning problem is referred to as data-driven inverse optimization. We focus on problems where the underlying decision-making process is modeled as a convex optimization problem whose parameters are unknown. We formulate the inverse optimization problem as a bilevel program and propose an efficient block coordinate descent-based algorithm to solve large problem instances. Numerical experiments on synthetic datasets demonstrate the computational advantage of our method compared to standard commercial solvers. Moreover, the real-world utility of the proposed approach is highlighted through two realistic case studies in which we consider estimating risk preferences and learning local constraint parameters of agents in a multiplayer Nash bargaining game.
AB - Decision-making problems are commonly formulated as optimization problems, which are then solved to make optimal decisions. In this work, we consider the inverse problem where we use prior decision data to uncover the underlying decision-making process in the form of a mathematical optimization model. This statistical learning problem is referred to as data-driven inverse optimization. We focus on problems where the underlying decision-making process is modeled as a convex optimization problem whose parameters are unknown. We formulate the inverse optimization problem as a bilevel program and propose an efficient block coordinate descent-based algorithm to solve large problem instances. Numerical experiments on synthetic datasets demonstrate the computational advantage of our method compared to standard commercial solvers. Moreover, the real-world utility of the proposed approach is highlighted through two realistic case studies in which we consider estimating risk preferences and learning local constraint parameters of agents in a multiplayer Nash bargaining game.
KW - Bilevel optimization
KW - Block coordinate descent
KW - Data-driven inverse optimization
KW - Statistical learning
UR - http://www.scopus.com/inward/record.url?scp=85144904595&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85144904595&partnerID=8YFLogxK
U2 - 10.1016/j.compchemeng.2022.108123
DO - 10.1016/j.compchemeng.2022.108123
M3 - Article
AN - SCOPUS:85144904595
SN - 0098-1354
VL - 170
JO - Computers and Chemical Engineering
JF - Computers and Chemical Engineering
M1 - 108123
ER -