SALT. A unified framework for all shortest-path query variants on road networks

TL;DR

SALT is a comprehensive framework that efficiently handles various shortest-path and k-nearest neighbor queries on road networks, suitable for dynamic environments with frequent traffic updates.

Contribution

It introduces a unified framework capable of answering multiple shortest-path variants and k-nearest neighbor queries with fast query times and short preprocessing, adaptable to dynamic road networks.

Findings

Achieves excellent query performance

Features very short preprocessing times

Suitable for dynamic road networks with frequent updates

Abstract

Although recent scientific output focuses on multiple shortest-path problem definitions for road networks, none of the existing solutions does efficiently answer all different types of SP queries. This work proposes SALT, a novel framework that not only efficiently answers SP related queries but also k-nearest neighbor queries not handled by previous approaches. Our solution offers all the benefits needed for practical use-cases, including excellent query performance and very short preprocessing times, thus making it also a viable option for dynamic road networks, i.e., edge weights changing frequently due to traffic updates. The proposed SALT framework is a deployable software solution capturing a range of network-related query problems under one "algorithmic hood".