Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Guokai, Gao, Pin, Jasin, Stefanus, Wang, Zizhuo
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2507.10834
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918202737426432
author Li, Guokai
Gao, Pin
Jasin, Stefanus
Wang, Zizhuo
author_facet Li, Guokai
Gao, Pin
Jasin, Stefanus
Wang, Zizhuo
contents Assortment optimization seeks to select a subset of substitutable products, subject to constraints, to maximize expected revenue. The problem is NP-hard due to its combinatorial and nonlinear nature and arises frequently in industries such as e-commerce, where platforms must solve thousands of such problems each minute. We propose a graph convolutional network (GCN) framework to efficiently solve constrained assortment optimization problems. Our approach constructs a graph representation of the problem, trains a GCN to learn the mapping from problem parameters to optimal assortments, and develops three inference policies based on the GCN's output. Owing to the GCN's ability to generalize across instance sizes, patterns learned from small-scale samples can be transferred to large-scale problems. Numerical experiments show that a GCN trained on instances with 20 products achieves over 85% of the optimal revenue on problems with up to 2,000 products within seconds, outperforming existing heuristics in both accuracy and efficiency. We further extend the framework to settings with an unknown choice model using transaction data and demonstrate similar performance and scalability.
format Preprint
id arxiv_https___arxiv_org_abs_2507_10834
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle From Small to Large: A Graph Convolutional Network Approach for Solving Assortment Optimization Problems
Li, Guokai
Gao, Pin
Jasin, Stefanus
Wang, Zizhuo
Machine Learning
Assortment optimization seeks to select a subset of substitutable products, subject to constraints, to maximize expected revenue. The problem is NP-hard due to its combinatorial and nonlinear nature and arises frequently in industries such as e-commerce, where platforms must solve thousands of such problems each minute. We propose a graph convolutional network (GCN) framework to efficiently solve constrained assortment optimization problems. Our approach constructs a graph representation of the problem, trains a GCN to learn the mapping from problem parameters to optimal assortments, and develops three inference policies based on the GCN's output. Owing to the GCN's ability to generalize across instance sizes, patterns learned from small-scale samples can be transferred to large-scale problems. Numerical experiments show that a GCN trained on instances with 20 products achieves over 85% of the optimal revenue on problems with up to 2,000 products within seconds, outperforming existing heuristics in both accuracy and efficiency. We further extend the framework to settings with an unknown choice model using transaction data and demonstrate similar performance and scalability.
title From Small to Large: A Graph Convolutional Network Approach for Solving Assortment Optimization Problems
topic Machine Learning
url https://arxiv.org/abs/2507.10834