Skip to the main content

Preliminary communication

The shortest path algorithm performance comparison in graph and relational database on a transportation network

Mario Miler ; Faculty of geodesy, University of Zagreb
Damir Medak ; Faculty of geodesy, University of Zagreb
Dražen Odobašić ; Faculty of geodesy, University of Zagreb


Full text: english PDF 752 Kb

page 75-82

downloads: 3.867

cite


Abstract

In the field of geoinformation and transportation science, the shortest path is calculated on graph data mostly found in road and transportation networks. This data is often stored in various database systems. Many applications dealing with transportation network require calculation of the shortest path. The objective of this research is to compare the performance of Dijkstra shortest path calculation in PostgreSQL (with pgRouting) and Neo4j graph database for the purpose of determining if there is any difference regarding the speed of the calculation. Benchmarking was done on commodity hardware using OpenStreetMap road network. The first assumption is that Neo4j graph database would be well suited for the shortest path calculation on transportation networks but this does not come without some cost. Memory proved to be an issue in Neo4j setup when dealing with larger transportation networks.

Keywords

pgRouting; OpenStreetMap; Dijkstra; benchmark; Neo4j; PostgreSQL

Hrčak ID:

124144

URI

https://hrcak.srce.hr/124144

Publication date:

28.2.2014.

Visits: 5.678 *