Test-Time Augmentation for Traveling Salesperson Problem

TL;DR

This paper introduces Test-Time Augmentation (TTA) for the Traveling Salesperson Problem, improving solution quality by leveraging permutation-based augmentations during inference.

Contribution

It presents a novel TTA scheme based on node permutation for combinatorial optimization, enhancing deep learning model performance on TSP.

Findings

TTA yields shorter TSP solutions than recent models.

Solution quality improves with larger augmentation sizes.

TTA increases the probability of near-optimal solutions.

Abstract

We propose Test-Time Augmentation (TTA) as an effective technique for addressing combinatorial optimization problems, including the Traveling Salesperson Problem. In general, deep learning models possessing the property of invariance, where the output is uniquely determined regardless of the node indices, have been proposed to learn graph structures efficiently. In contrast, we interpret the permutation of node indices, which exchanges the elements of the distance matrix, as a TTA scheme. The results demonstrate that our method is capable of obtaining shorter solutions than the latest models. Furthermore, we show that the probability of finding a solution closer to an exact solution increases depending on the augmentation size.