Home

DEVELOPMENT OF A REINFORCEMENT LEARNING MODULE BASED ON Q-LEARNING OF A MULTI-AGENT SYSTEM FOR EMERGENCY EVACUATION SIMULATION (both dynamic and static fire approaches):

Paper

Used frameworks (ABSS):

1.- MESA framework:

https://github.com/projectmesa/mesa/wiki

2.- SOBA framework:

http://soba.readthedocs.io/en/latest/sobaoverview.html

3.- SEBA framework:

http://soba.readthedocs.io/en/latest/projects/seba/index.html

Used Q-Learning implementation:

https://github.com/GGP00/Q-Learning-maze-construction-E-greedy-approach-

Sneak peek of the algorithm convergence process: