Solving large-scale instances of the Traveling Salesman Problem (TSP) exactly is computationally expensive. Researchers often employ graph sparsification methods to improve computational efficiency. Traditional sparsification methods typically rely on fixed heuristics and fail to fully exploit instance-specific structural information. This paper proposes a learning-based sparsification approach called Graph Edge Sparsification (GES) for Euclidean TSP.
By incorporating geometric structural information and combinatorial optimization technology, our method adaptively generates a sparsification graph for different instances, significantly reducing the graph size and accelerating the solving process. Experimental results demonstrate that our sparsification method can prune up to 95% of edges on the MATILDA dataset while keeping the solution gap within 1% of the optimal value. Moreover, our approach exhibits strong generalization capability on the TSPLIB benchmark. In some large-scale instances, the pruning rate exceeds 99%, while the optimality gap remains below 1%.
Blogger's Review: The GES method innovatively leverages a learning mechanism for efficient sparsification tailored to specific instances, enhancing TSP solving efficiency. This showcases how to integrate geometric and combinatorial optimization in practical applications, holding significant theoretical value and broad application prospects.