play_tournament.py

from agents.BNN_PIMC_Player import BNNPIMCPlayer
from agents.PIMC_Player import PIMCPlayer
from agents.RandomPlayer import RandomPlayer

from game_environment import SchafkopfEnv

def main():
    pimc_player = PIMCPlayer(10, 40, RandomPlayer())
    bnn_pimc_player = BNNPIMCPlayer(10, 40, RandomPlayer())

    participants = [pimc_player, pimc_player, bnn_pimc_player, bnn_pimc_player]

    number_of_games = 100

    for i in range(len(participants)):
        for j in range(i + 1, len(participants)):
            p1 = participants[0]
            p2 = participants[1]
            p3 = participants[2]
            p4 = participants[3]

            cummulative_reward = [0, 0, 0, 0]
            players = [p1, p2, p3, p4]
            for k in range(2):  # run the same tournament twice with different positions of players
                print(' ')
                schafkopf_env = SchafkopfEnv(seed=1)
                        #    if k == 0:
                        #        players = [p1, p1, p2, p2]
                        #    else:
                        #        players = [p2, p2, p1, p1]
                        #        cummulative_reward.reverse()

                            # tournament loop
                for game_nr in range(1, number_of_games + 1):
                    state, reward, terminal = schafkopf_env.reset()
                    while not terminal:
                        action, prob = players[state["game_state"].current_player].act(state)
                        state, reward, terminal = schafkopf_env.step(action, prob)

                    cummulative_reward = [cummulative_reward[m] + reward[m] for m in range(4)]

                    if game_nr % 100 == 0:
                        print('.', end='')
                            # schafkopf_env.print_game()

                print(cummulative_reward)
                print("player " + str(i) + " vs. player " + str(j) + " = " + str(
                    (cummulative_reward[2] + cummulative_reward[3]) / (2 * 2 * number_of_games)) + " to " + str(
                    (cummulative_reward[0] + cummulative_reward[1]) / (2 * 2 * number_of_games)))
                        # print("--------Episode: " + str(i_episode) + " game simulation (s) = " + str(t1 - t0))
                        # print("--------Cummulative reward: " + str(cummulative_reward))
                        # print("--------per game reward: " + str([i /i_episode for i in cummulative_reward] ))
                        # print("--------monte_carlo_tree_search rewards: " + str(((cummulative_reward[1] + cummulative_reward[3]) / i_episode)/2))
                        # print("--------monte_carlo_tree_search rewards: " + str(((cummulative_reward[1] + cummulative_reward[3]) / i_episode)/2))


if __name__ == '__main__':
    main()