Design and implementation of a Multi-Agent Pathfinding library in C++ for Automated logistic warehouses
Files
Mans_15291800_Mahy_39771800_2023.pdf
Open access - Adobe PDF
- 8.13 MB
Details
- Supervisors
- Faculty
- Degree label
- Abstract
- This thesis elaborates on an associated c++ library in which the most common optimal algorithms for solving the Multi-Agent PathFinding (MAPF) problem were implemented. It contains a summary of these methods, implementation details and numerical experiments to compare their performance and check the impact of the various enhancements on success rate and running time. The implemented algorithms and their enhancements include A* in space or space-time search, in single-agent and multi-agent search, with and without operator decomposition; Independence Detection (ID) with or without Conflict Avoidance Table (CAT), as well as its Enhanced version (EID); Conflict Based Search (CBS) with or without CAT, with basic splitting or disjoint splitting.