Learn to Compare Nodes in Branch and Bound

Explore the intricacies of node comparison in Branch and Bound algorithms during this 24-minute DS4DM Coffee Talk presented by Abdel Ghani Labassi from Johns Hopkins University. Delve into the importance of efficient node selection strategies for optimizing NP-hard problem solutions. Discover various approaches to node selection, with a focus on data-based methods. Examine a novel contribution utilizing graph neural networks to compare nodes represented as bipartite graphs. Learn how this model, capable of handling variable dimension data, demonstrates faster resolution and reduced branching trees across three NP-hard benchmarks while offering natural generalization to larger instances.