This is an elective course that covers the basic concepts and methods of multi-agent pathfinding, as well as its practical applications. The course assumes basic knowledge of logic and search algorithms at the undergraduate level. The course aims to give students an overview of fundamental methods and concepts of multi-agent pathfinding, familiarize them with their practical usage, and present the current open problems in the research community (which are perfect topics for bachelor’s/master’s theses).
Organization
The dates and times of the lecture will be determined based on the agreement with the students. The lecture will most likely be held in the corridor on the second floor of the Malostranské náměstí building (rooms starting with S).
To pass the course, the student must pass an exam. The exam is oral, with time for written preparation. The requirements correspond to the syllabus to the extent presented during the lectures.
Useful links and literature
- Webs
- Papers
- Multi-Agent Pathfinding: Definitions, Variants, and Benchmarks
- A Comprehensive Review on Leveraging Machine Learning for Multi-Agent Path Finding
- 42+ Flavors of Multi-Agent Path Finding
- Repositories and implementation
Lectures
| 26.3. | slides | Introduction, overview of topics to be discussed during the class |
| 5.3. | slides | Definitions, single agent pathfinding (A*, SIPP), CBS, Reduction to SAT, BCP |
| 12.3. | slides | Sub-optimal solvers |
| 19.3. | video | PIBT, LaCAM and current research by Keisuke Okumura |
| 26.3. | slides | Decentralized solvers and ML in MAPF |
| 2.4. | slides | MAPF variants |
| 9.4. | slides | MAPF on robots and safe execution policies |
| 16.4. | slides | Intersection management |
| 23.4. | slides | Games |
| 30.4. | slides | Routing problem |
| 7.5. | Cancelled | |
| 14.5. | slides | Demos, videos, and open problems |
| 21.5. | Exam |