Neural TSP Solver with Progressive Distillation
Dongxiang Zhang, Ziyang Xiao, Yuan Wang, Mingli Song, Gang Chen
Abstract
Travelling salesman problem (TSP) is NP-Hard with exponential search space. Recently, the adoption of encoder-decoder models as neural TSP solvers has emerged as an attractive topic because they can instantly obtain near-optimal results for small-scale instances. Nevertheless, their training efficiency and solution quality degrade dramatically when dealing with large-scale problems. To address the issue, we propose a novel progressive distillation framework, by adopting curriculum learning to train TSP samples in increasing order of their problem size and progressively distilling high-level knowledge from small models to large models via a distillation loss. In other words, the trained small models are used as the teacher network to guide action selection when training large models. To accelerate training speed, we also propose a Delaunary-graph based action mask and a new attention-based decoder to reduce decoding cost. Experimental results show that our approach establishes clear advantages over existing encoder-decoder models in terms of training effectiveness and solution quality. In addition, we validate its usefulness as an initial solution generator for the state-of-the-art TSP solvers, whose probability of obtaining the optimal solution can be further improved in such a hybrid manner.
BibTeX
@article{Zhang_Xiao_Wang_Song_Chen_2023, title={Neural TSP Solver with Progressive Distillation}, volume={37}, url={https://ojs.aaai.org/index.php/AAAI/article/view/26432}, DOI={10.1609/aaai.v37i10.26432}, abstractNote={Travelling salesman problem (TSP) is NP-Hard with exponential search space. Recently, the adoption of encoder-decoder models as neural TSP solvers has emerged as an attractive topic because they can instantly obtain near-optimal results for small-scale instances. Nevertheless, their training efficiency and solution quality degrade dramatically when dealing with large-scale problems. To address the issue, we propose a novel progressive distillation framework, by adopting curriculum learning to train TSP samples in increasing order of their problem size and progressively distilling high-level knowledge from small models to large models via a distillation loss. In other words, the trained small models are used as the teacher network to guide action selection when training large models. To accelerate training speed, we also propose a Delaunary-graph based action mask and a new attention-based decoder to reduce decoding cost. Experimental results show that our approach establishes clear advantages over existing encoder-decoder models in terms of training effectiveness and solution quality. In addition, we validate its usefulness as an initial solution generator for the state-of-the-art TSP solvers, whose probability of obtaining the optimal solution can be further improved in such a hybrid manner.}, number={10}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Zhang, Dongxiang and Xiao, Ziyang and Wang, Yuan and Song, Mingli and Chen, Gang}, year={2023}, month={Jun.}, pages={12147-12154} }