From Small to Large: A Graph Convolutional Network Approach for Solving Assortment Optimization Problems
Researchers have developed a novel Graph Convolutional Network (GCN) framework to tackle the complex and computationally intensive problem of assortment optimization. This method represents assortment problems as graphs, enabling the GCN to learn optimal product selections. The framework demonstrates remarkable scalability, with models trained on small datasets effectively generalizing to much larger problems, achieving over 85% of optimal revenue on instances with up to 2,000 products within seconds. AI
IMPACT This research offers a scalable and efficient AI-driven solution for a common e-commerce problem, potentially improving revenue for online platforms.