graph-convnet-tsp
Code for the paper 'An Efficient Graph Convolutional Network Technique for the Travelling Salesman Problem' (arXiv Pre-print)
view repo
This paper introduces a new learning-based approach for approximately solving the Travelling Salesman Problem on 2D Euclidean graphs. We use deep Graph Convolutional Networks to build efficient TSP graph representations and output tours in a non-autoregressive manner via highly parallelized beam search. Our approach outperforms all recently proposed autoregressive deep learning techniques in terms of solution quality, inference speed and sample efficiency for problem instances of fixed graph sizes. In particular, we reduce the average optimality gap from 0.52 1.39 approaches for TSP, our approach falls short of standard Operations Research solvers.
READ FULL TEXT
We present a learning-based approach to computing solutions for certain
...
10/25/2018 ∙ by Zhuwen Li, et al. ∙
0
∙
share
read it
Graph convolutional networks(GCNs) have become the most popular approach...
07/21/2019 ∙ by Yi Ma, et al. ∙
0
∙
share
read it
This paper explores the recently proposed Graph Convolutional Network
ar...
03/12/2018 ∙ by Sean Billings, et al. ∙
0
∙
share
read it
In recent years, there has been a surge of interest in developing deep
l...
06/03/2018 ∙ by Federico Monti, et al. ∙
0
∙
share
read it
Graph convolutional networks (GCNs) for learning graph representation of...
11/24/2018 ∙ by Hyeoncheol Cho, et al. ∙
0
∙
share
read it
Graph Convolutional Networks (GCNs) and their variants have experienced
...
02/19/2019 ∙ by Felix Wu, et al. ∙
0
∙
share
read it
Geometric deep learning provides a principled and versatile manner for t...
03/11/2019 ∙ by Anees Kazi, et al. ∙
0
∙
share
read it
Code for the paper 'An Efficient Graph Convolutional Network Technique for the Travelling Salesman Problem' (arXiv Pre-print)