I am a lecturer and researcher of computer science at the College of Engineering & Computer Science (CECS) at the ANU. Before joining the ANU in October 2018 I completed my PhD in computer science at the University of Freiburg in Germany, supervised by Prof. Dr. Bernhard Nebel. During my PhD I was part of the Karis Pro project, where our group developed the localisation, navigation and path planning for a fleet of robots. The project was a collaboration between many industrial partners (with the Robert Bosch GmbH and quattro GmbH as end users), the Karlsruhe Institute of Technology (KIT) and the Foundations of AI group of the University of Freiburg. A second industrial collaboration project during my PhD was concerned with intralogistics and multi-agent path-finding. At ANU I am part of the DONUT project, in collaboration with the Airbus Central Research and Technology group. The project focuses on planning and scheduling under uncertainty for aircraft and satellite applications. Besides project work I also supervise students, organized workshops between different universities, and I am a program committee member of the AAAI Conference on Artificial Intelligence (AAAI), the International Conference on Automated Planning and Scheduling (ICAPS), and the International Joint Conference on Artificial Intelligence (IJCAI).
Note that the ß in Geißer is a german letter (not a b) and you can also write Geisser instead.
My research interests span the areas of deterministic and probabilistic planning and scheduling, as well as multi-agent path planning, which is both strongly connected to combinatorial and heuristic search. My PhD thesis was concerned with planning problems where the cost of an action is not constant, but depends on the state where the action is applied. While considering state-dependent action costs is rather unusual in the field of classical planning, it is common in probabilistic planning, for example when calculating rewards in Markov Decision Processes (MDPs). A key data structure for these type of problems are decision diagrams, which play a major role in many areas of AI, such as model checking and verification, but also more recently in Operations Research. Together with Benedict Wright I developed the decision diagram library LEMON-DD, which allows to construct decision diagrams over different types of carrier sets.
I am core developer of the probabilistic planning system Prost, and the algorithms we use in the planner are also applied on some of the industrial problems we face in our collaboration projects. A central theme of my work is a strong emphasis on a sound theoretical background, which serves as the foundation for scalable, feasible and maintainable systems.
Geißer, Poveda, Werndl-Trevizan, Bondouy, Teichteil-Königsbuch, and Thiebaux. Optimal and Heuristic Approaches for Constrained Flight Planning under Weather Uncertainty. ICAPS 2020, to appear.
Speck, Geißer, and Mattmüller. When Perfect is not Good Enough: On the Search Behaviour of Symbolic Heuristic Search. ICAPS 2020, to appear.
Corraya, Geißer, Speck, and Mattmüller. An Empirical Study of the Usefulness of State-Dependent Action Costs in Planning. In Proc. 42th KI, 2019.
- Speck, Geißer, Mattmüller, and Torralba. Symbolic Planning with Axioms. In Proc. 29th ICAPS, 2019.
- Geißer, Speck, and Keller. An Analysis of the Probabilistic Track of the IPC 2018. In Workshop on the International Planning Competition (WIPC), 2019.
- Geißer. On Planning with State-dependent Action Costs. PhD thesis, University of Freiburg, 2019.
- Speck, Geißer, and Mattmüller. Symbolic Planning with Edge-Valued Multi-Valued Decision Diagrams. In Proc. 28th ICAPS, 2018.
- Mattmüller, Geißer, Wright, and Nebel. On the Relationship Between State-Dependent Action Costs and Conditional Effects in Planning. In Proc. 32nd AAAI, 2018.
- Sun, Geißer, and Nebel. Towards Effective Localization in Dynamic Environments. In Proc. IEEE IROS Conference 2016, 2016.
- Keller, Pommerening, Seipp, Geißer, and Mattmüller. State-dependent Cost Partitionings for Cartesian Abstractions in Classical Planning. In Proc. 25th IJCAI, 2016.
- Geißer, Keller, and Mattmüller. Abstractions for Planning with State-Dependent Action Costs. In Proc. 26th ICAPS, 2016.
- Keller and Geißer. Better Be Lucky Than Good: Exceeding Expectations in MDP Evaluation. In Proc. 24th AAAI, 2015.
- Geißer, Keller, and Mattmüller. Delete Relaxations for Planning with State-Dependent Action Costs. In Proc. 24th IJCAI, 2015.
- Geißer, Keller, and Mattmüller. Past, Present, and Future: An Optimal Online Algorithm for Single-Player GDL-II Games. In Proc. 21st ECAI, 2014.