Alternative routing

Abstract

Shortest path computation is a fundamental problem in road networks with application in various domains in research and industry. However, returning only the shortest path is often not satisfying; users are also interested in alternative paths which might be longer but have other advantages, e.g., less frequent traffic congestion. In this paper, we formally introduce the k-Shortest Paths with Limited Overlap (k-SPwLO) problem seeking to recommend k alternative paths which are (a) as short as possible and (b) sufficiently dissimilar based on a user-controlled similarity threshold. We propose two algorithms that examine the paths from a source s to a target t in increasing order of their length and progressively construct the result set. The baseline algorithm BSL builds upon a standard algorithm for computing k-Shortest Paths, followed by a filter step. The OnePass algorithm considers the overlap constraint in each expansion step while traversing the network. We evaluate the performance of both algorithms on real road networks and show that OnePass always outperforms BSL.

Multi-formalism Models for Performance Engineering

“…The first one is the alternative route planning that consists of the identification of k alternative paths from source to destination. This step is referred to as the K-shortest paths with limited overlap [30] problem, since the alternative paths should overlap less than a given threshold. The shortest path in terms of distance or average traveling time might not be the optimal one, as several other parameters may have a significant influence on the computation of the best solution.…”