experimentargo.py

from protargo import *
import os
import numpy as np
import itertools

if __name__ == "__main__":
    # The number of agents
    AGENTS_NUM_MIN = 2
    AGENTS_NUM_MAX = 100
    AGENTS_NUM_STEP = 5
    # The number of arguments attacking the 
    # issue directly
    ROOT_BRANCH_MIN = 2
    ROOT_BRANCH_MAX = 10
    ROOT_BRANCH_STEP = 2
    # The maximum number of arguements 
    # per branch
    MAX_ARGUMENTS_PER_BRANCH_MIN = 2
    MAX_ARGUMENTS_PER_BRANCH_MAX = 1000
    MAX_ARGUMENTS_PER_BRANCH_STEP = 10
    # The random seed that was used 
    RANDOM_SEED_MIN = 1
    RANDOM_SEED_MAX = 1000000
    RANDOM_SEED_STEP = 10000
    # The maximum number of arguments to be 
    # played at once
    MAX_ARGUMENTS_AT_ONCE_MIN = 1
    MAX_ARGUMENTS_AT_ONCE_MAX = 10
    MAX_ARGUMENTS_AT_ONCE_STEP = 1
    experiment_counter = 0

    # Génération des combinaisons de paramètres
    rand_seeds = np.arange(RANDOM_SEED_MIN, RANDOM_SEED_MAX, RANDOM_SEED_STEP)
    max_arguments_at_once = np.arange(MAX_ARGUMENTS_AT_ONCE_MIN, MAX_ARGUMENTS_AT_ONCE_MAX, MAX_ARGUMENTS_AT_ONCE_STEP)
    agent_nums = np.arange(AGENTS_NUM_MAX, AGENTS_NUM_MIN, -AGENTS_NUM_STEP)
    root_branchs = np.arange(ROOT_BRANCH_MAX, ROOT_BRANCH_MIN, -ROOT_BRANCH_STEP)
    max_arguments_per_branchs = np.arange(MAX_ARGUMENTS_PER_BRANCH_MIN, MAX_ARGUMENTS_PER_BRANCH_MAX, MAX_ARGUMENTS_PER_BRANCH_STEP)

    print(rand_seeds,max_arguments_at_once,agent_nums)
    # générer les permutations
    permutations = itertools.product(rand_seeds, max_arguments_at_once, agent_nums, root_branchs, max_arguments_per_branchs)

    # exécuter la boucle sur les permutations
    for perm in permutations:
        rand_seed, max_arguments_at_once, agent_num, root_branch, max_arguments_per_branch = perm
        if root_branch * max_arguments_per_branch < rand_seed: 
            continue
        if root_branch * max_arguments_per_branch < max_arguments_at_once: 
            continue
        print(f"Experiment {experiment_counter}")
        command = f"Command: python3 main.py --agents {agent_num} --root-branch {root_branch} --max-arguments-per-branch {max_arguments_per_branch} --rand-seed {rand_seed} --max-arguments-at-once {max_arguments_at_once} --nodebug --batch-mode"
        print(command)
        os.system(command)
        experiment_counter+=1