Combining Reinforcement Learning and Optimal Transport for the Traveling Salesman Problem

The traveling salesman problem is a fundamental combinatorial optimization problem with strong exact algorithms. However, as problems scale up, these exact algorithms fail to provide a solution in a reasonable time. To resolve this, current works look at utilizing deep learning to construct reasonab...

Full description

Saved in:
Bibliographic Details
Published inarXiv.org
Main Authors Yong Liang Goh, Wee Sun Lee, Bresson, Xavier, Thomas, Laurent, Lim, Nicholas
Format Paper
LanguageEnglish
Published Ithaca Cornell University Library, arXiv.org 02.03.2022
Subjects
Algorithms
Combinatorial analysis
Constraint modelling
Deep learning
Machine learning
Optimization
Traveling salesman problem
Online AccessGet full text

Cover

More Information
Summary:The traveling salesman problem is a fundamental combinatorial optimization problem with strong exact algorithms. However, as problems scale up, these exact algorithms fail to provide a solution in a reasonable time. To resolve this, current works look at utilizing deep learning to construct reasonable solutions. Such efforts have been very successful, but tend to be slow and compute intensive. This paper exemplifies the integration of entropic regularized optimal transport techniques as a layer in a deep reinforcement learning network. We show that we can construct a model capable of learning without supervision and inferences significantly faster than current autoregressive approaches. We also empirically evaluate the benefits of including optimal transport algorithms within deep learning models to enforce assignment constraints during end-to-end training.
ISSN:2331-8422