issues
search
smarianimore
/
RL-swarms
Python ports of popular NetLogo models for MARL experiments
1
stars
8
forks
source link
algorithms and tools implemented for the exam project for Daniel Rossi's Distribute Artificial Intelligence
#9
Closed
ProjectoOfficial
closed
1 year ago
ProjectoOfficial
commented
1 year ago
added DQN
Centralized: one network which predicts the current action for each agent
Decentralized: each agent has its own network used to predict actions
added SARSA
added new environment state: pheromone quantity within the neighborhood (np.ndarray([float, float, ...])
reorganized agents directory
added utils