Commit Graph
30 Commits
Author SHA1 Message Date
Julian Schönberger 7706d1f704 Store code submission last state for paper 2024-10-23 11:00:14 +02:00
Julian Schönberger c56bb3f188 Reword TSP runner 2024-08-09 16:36:39 +02:00
Julian Schönberger 0a8eaf8827 Increased maximum number of nodes for TSP agent to deal with all potential dirt placements in dirt_quadrant 2024-08-09 16:36:05 +02:00
Julian Schönberger 973f3e9fc8 Added various plotting methods + Fixed action maps plotting 2024-08-09 16:35:08 +02:00
Julian Schönberger fcd2eaf773 Added option for random dirt positioning 2024-08-09 16:33:43 +02:00
Julian Schönberger 50ef0c94e9 Updated RL_runner 2024-08-09 16:33:06 +02:00
Julian Schönberger 4c81e4b865 Reworked configurations 2024-08-09 16:31:08 +02:00
Julian Schönberger 8e8e925278 Reworked differentiation between train and eval execution + Renamed cfgs + Added algorithm seeding + Included early stopping functionality + Added weights&biases logging 2024-08-09 16:30:04 +02:00
Julian Schönberger 81b12612ed Updated required python version 2024-05-27 18:45:28 +02:00
Julian Schönberger a0852e805a Updated pomdp_r comment + Added some additional comments + Restructured experiment calling + Added Readme and requirements.txt 2024-05-27 18:23:11 +02:00
Julian Schönberger 41a1ec0a5b Update utility plotting + Added alternative methods for TSP emergence prevention 2024-05-27 13:25:52 +02:00
Julian Schönberger a78e0dd8a3 Added commentary to configs 2024-05-25 18:31:18 +02:00
Julian Schönberger dca930636e Add non-generic utility plots 2024-05-25 17:36:49 +02:00
Julian Schönberger ac35e46310 Refactored a2c_dirt file 2024-05-25 01:45:09 +02:00
Julian Schönberger 81f0f6e209 Code cleaning part 2 2024-05-24 23:56:00 +02:00
Julian Schönberger 6e6ce9dc5d Equalize rendering for TSP and RL agents 2024-05-24 18:54:38 +02:00
Julian Schönberger c8336e8f78 Reset tsp route caching + renamed and moved configs + removed unnecessary files 2024-05-24 16:12:05 +02:00
Julian Schönberger 4571dc1cd1 Added shared piles option to dirt_quadrant eval + Changed dirt_quadrant layout and adapted configs 2024-05-17 17:52:52 +02:00
Julian Schönberger 5e9e59c843 Added code for tsp_runs + Updated eval configs so that every episode only takes a maximum number of steps 2024-05-10 17:38:34 +02:00
Julian Schönberger d0d31b964c Added 'shared' dirt piles option for eval + Fixed usage of renderer + Added recorder option 2024-05-10 14:54:00 +02:00
Julian Schönberger a25b04e092 Included method to tackle emergence in two_rooms_one_door_modified + Better access of different settings in marl_adapted + Added and modified a lot of config files 2024-05-10 11:57:26 +02:00
Julian Schönberger 89ce723690 Fixed gitignore ignoring study_out base folder 2024-05-06 12:48:07 +02:00
Julian Schönberger 3c54d04f9f All relevant functional code for A2C Dirt Quadrant setting with small changes to the environment + Different configs for single agent and multiagent settings 2024-05-06 12:33:37 +02:00
Julian Schönberger 55026eda12 Add Independent A2C implementation 2024-05-02 11:10:23 +02:00
Julian Schönberger 175a02f997 Add various RL adapted configs part 2 2024-05-02 11:01:20 +02:00
Julian Schönberger c7c2c4e5a3 Add various RL adapted configs 2024-05-02 11:00:35 +02:00
Julian Schönberger 48d708bbcd environment code changes for RL settings 2024-05-02 10:54:46 +02:00
Julian Schönberger 50bcf5d995 Merge branch 'unit_testing' into marl_refactor 2024-04-19 09:46:46 +02:00
Julian Schönberger 63bed3f4ba Include modified version of two_rooms_one_door_config + small changes on two other configs and a rl study file 2024-03-27 17:06:23 +01:00
Julian Schönberger 086a921929 Adapt base_ac.py and utils.py to be compatible with refactored environment 2024-03-27 17:04:14 +01:00