Reinforcement Learning for Mobility

In this project, we investigate various tasks where a single or multiple mobile agents move in dynamically changing, constrained environments. Dynamic change might reflect varying travel conditions or the availability of local resources. Applications include emergency management systems, parking management, and logistics services.

Research Focus

Controlling movement of autonomous agents in constrained, dynamic environments is a inherent skill that has to be learned in various tasks such as autonomous driving, robotics or computer games. Dynamic environments yield inherent challanges as the most attractive location to reach for the agents might change as well as the cost-optimal path to reach these goals.

Publications

  • Niklas Strauß and Matthias Schubert: Spatial-Aware Deep Reinforcement Learning for the Traveling Officer Problem, to appear at the SIAM Int. Conference on Data Mining (SDM'24), 18.04-20-04.20 in Houston, Texas, U.S.
  • Lukas Rottkamp, Niklas Strauß, Matthias Schubert: DEAR: Dynamic Electric Ambulance Redeployment, in Proceedings of the 18th International Symposium on Spatial and Temporal Data (SSTD 23),23-25 August 23, Calgary, AB, Canada
  • Lukas Rottkamp, Matthias Schubert, Niklas Strauß: Efficient on-street parking sensor placement. IWCTS@SIGSPATIAL 2022: 13:1-13:8
  • Niklas Strauß, Max Berrendorf, Tom Haider, Matthias Schubert: A Comparison of Ambulance Redeployment Systems on Real-World Data. ICDM (Workshops) 2022: 1-8
  • Niklas Strauß, David Winkel, Max Berrendorf, Matthias Schubert: Reinforcement Learning for Multi-Agent Stochastic Resource Collection. ECML/PKDD (4) 2022: 200-215
  • Niklas Strauß, Lukas Rottkamp, Sebastian Schmoll, Matthias Schubert: Efficient Parking Search using Shared Fleet Data. MDM 2021: 115-120
  • Lukas Rottkamp, Matthias Schubert: Quantifying the potential of data-driven mobility support systems. IWCTS@SIGSPATIAL 2020: 2:1-2:10
  • Sebastian Schmoll, Matthias Schubert: Semi-Markov Reinforcement Learning for Stochastic Resource Collection. IJCAI 2020: 3349-3355
  • Sabrina Friedl, Sebastian Schmoll, Felix Borutta, Matthias Schubert: SMART-Env. MDM 2020: 234-235
    Evgeniy Faerman, Felix Borutta, Julian Busch, Matthias Schubert: Ada-LLD: Adaptive Node Similarity
  • Sebastian Schmoll, Sabrina Friedl, Matthias Schubert: Scaling the Dynamic Resource Routing Problem. SSTD 2019: 80-89

What are you looking for?

More information about Google data transfer in LMU's Privacy Policy.