This repository contains the code and instructions for reproducing the experiments presented in:

"Successful Misunderstandings: Learning to Coordinate Without Being Understood", 
Nikolaos Kondylidis, Anil Yaman, Frank van Harmelen, Erman Acar, Annette ten Teije, 
Presented at the 22nd European Conference on Multi-Agent Systems (EUMAS 2025), Bucharest

Experiment Reproducibility Instructions:

1 Download or clone this git

2 Create python environment and install required libraries

python3 -m venv suc_mis_venv
source suc_mis_venv/bin/activate
pip install -r requirements.txt 

(change "experiment.num_workers=10" to set number of threads to work in parallel)

Experiment 1:

python main.py experiment.init_population=2 experiment.final_population=3 experiment.reward_table=random_simple experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5 experiment.num_workers=10

Experiment 2:

python main.py experiment.init_population=3 experiment.final_population=4 experiment.reward_table=random_simple experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5 experiment.num_workers=10

Experiment 3:

python main.py experiment.population_size=3 experiment.reward_table=random_simple experiment.experiment_setting=agents_initially_grouped_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5 experiment.num_workers=10

Experiments for Validating with more complicated Reward functions:

Non-symmeic reward function:

% Experiment 1
python main.py experiment.init_population=2 experiment.final_population=3 experiment.reward_table=random_simple_non_symmetric experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5
% Experiment 2
python main.py experiment.init_population=3 experiment.final_population=4 experiment.reward_table=random_simple_non_symmetric experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5
% Experiment 3
python main.py experiment.population_size=3 experiment.reward_table=random_simple_non_symmetric experiment.experiment_setting=agents_initially_grouped_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5

3x3 Reward table, 3 States & 3 Actions:

% Experiment 1
python main.py experiment.init_population=2 experiment.final_population=3 experiment.reward_table=random_3x3 experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5
% Experiment 2
python main.py experiment.init_population=3 experiment.final_population=4 experiment.reward_table=random_3x3 experiment.experiment_setting=population_increase_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5
% Experiment 3
python main.py experiment.population_size=3 experiment.reward_table=random_3x3 experiment.experiment_setting=agents_initially_grouped_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5

Experiment 3 with 4 agents:

python main.py experiment.population_size=4 experiment.reward_table=random_simple experiment.experiment_setting=agents_initially_grouped_experiment experiment.episodes=10000 experiment.repetitions=1000 agent.apply_epsilon_in_episode_ratio=0.5