Tree-Deletion Pruning in Label-Correcting Algorithms for the Multiobjective Shortest Path Problem

TL;DR

This paper re-evaluates label-correcting algorithms for the multiobjective shortest path problem, showing that node-selection strategies outperform label-selection when carefully implemented, and introduces an effective pruning method tested on real-world networks.

Contribution

The paper demonstrates the superiority of node-selection over label-selection strategies and introduces a new pruning method for multiobjective shortest path algorithms.

Findings

Node-selection strategies outperform label-selection with careful implementation.

The new pruning method performs well on real-world road networks.

Algorithms tested on instances with up to 15 objectives and 160,000 nodes.

Abstract

In this paper, we re-evaluate the basic strategies for label correcting algorithms for the multiobjective shortest path (MOSP) problem, i.e., node and label selection. In contrast to common believe, we show that---when carefully implemented---the node-selection strategy usually beats the label-selection strategy. Moreover, we present a new pruning method which is easy to implement and performs very well on real-world road networks. In this study, we test our hypotheses on artificial MOSP instances from the literature with up to 15 objectives and real-world road networks with up to almost 160,000 nodes.