Spatial AI

Resource Search in Spatial Environments

Resources in a spatial environments represent objects being located at a certain position which might appear and disappear over time. Examples are parking spots, charging stations or even customers looking for a roaming cab. Due to novel sensor techniques historical and real time information about the availability of resources gets more and more available. Our goal in this project is to develop strategies for minimizing the search time for an available resource in a spatial environment.

Contributors:

  • Gregor Jossé
  • Klaus Schmid
  • Sebastian Schmoll
  • Matthias Schubert

publications:

  • Sebastian Schmoll, Matthias Schubert (2018). Dynamic Resource Routing using Real-Time Dynamic Programming. In IJCAI'18, Stockholm, Sweden (to appear)
  • Sebastian Schmoll, Matthias Schubert (2018). Dynamic Resource Routing using Real-Time Information. In EDBT'18, Vienna, Austria, pp.4
  • Gregor Jossé, Klaus Arthur Schmid, Matthias Schubert (2015). Probabilistic Resource Route Queries with Reappearance. In EDBT'15, pp. 445–456.
  • Gregor Jossé, Matthias Schubert, Hans-Peter Kriegel (2013). Probabilistic Parking Queries using Aging Functions. In ACMGIS'13, pp. 4

Multicriteria Shortest Path Search

Finding cost optimal paths in graphs strongly depends on the employed cost function. However, in many applications like routing multiple types of cost should be considered. Thus, optimal solutions depend on the given preference function combining multiple cost criteria. In this project, we focus on computing the set of all potential optimal solution considering a monotonous combination function. We develop fast A* based search algorithm and the linear skyline to directly compute a very important subset of all pareto optimal paths.

Contributors:

  • Gregor Jossé
  • Michael Shekelyan
  • Matthias Schubert
  • Matthias Renz

publications:

  • Michael Shekelyan, Gregor Jossé, Matthias Schubert (2015). Linear Path Skyline Computation in Multicriteria Networks. In ICDE'15, pp. 459-470.
  • Michael Shekelyan, Gregor Jossé, Matthias Schubert, Hans-Peter Kriegel (2014). Linear Path Skyline Computation in Bicriteria Networks. In DASFAA'14, pp. 173-187.
  • Adrian Balteanu, Gregor Jossé, Matthias Schubert (2013). Mining Driving Preferences in Multi-cost Networks. In SSTD'13, pp. 74-91
  • Matthias Schubert, Hans-Peter Kriegel (2011). LOCAR: Local Compression of Alternative Routes. In ACMGIS'11, pp. 441–444
  • Matthias Schubert, Matthias Renz, Hans-Peter Kriegel (2010). Route skyline queries: a multi-preference path planning approach. In ICDE'10, pp. 261–272.

Spatial Environments and Routing Frameworks

Though spatial environments and routing frameworks are not directly a AI projects, they serve as virtual environments for testing and simulating spatial AI solutions. We developed multiple frameworks such as PAROS, MARIO and Easy-EV. Our latest Open Streetmap-based mobility planing system is Navigazelle. Navigazelle allows multi-model routing, time dependent travel costs and is based on a web interface for easy interaction.

Contributors:

  • Gregor Jossé
  • Klaus Schmid
  • Franz Graf
  • Robert Greil
  • Sebastian Schmoll
  • Ludwig Zellner
  • Matthias Schubert

publications:

  • Gregor Jossé, Matthias Schubert, Ludwig Zellner (2015). EasyEV: Monitoring and Querying System for Electric Vehicle Fleets Using Smart Car Data. In SSTD'15, pp. 497-502.
  • Franz Graf, Hans-Peter Kriegel, Matthias Renz, Matthias Schubert (2011). MARiO: Multi Attribute Routing in Open Street Map. In SSTD'11, pp. 486–490.
  • Franz Graf, Hans-Peter Kriegel, Matthias Renz, Matthias Schubert (2010). PAROS: Pareto Optimal Route Selection. In SIGMOD'10, pp. 1199–1202.