스탠포드 CS221: 인공지능 - 게임 이론과 탐색 알고리즘 기초

def V_eval(game, policies, state):
    if game.is_end(state):
        return game.utility(state)
    player = game.player(state)
    policy = policies[player]
    value = 0
    for action, prob in policy(state).items():
        next_state = game.successors(state)[action]
        value += prob * V_eval(game, policies, next_state)
    return value

def V_minimax(game, state):
    if game.is_end(state):
        return game.utility(state), None
    player = game.player(state)
    successors = game.successors(state)
    values = [(V_minimax(game, next_state)[0], action) for action, next_state in successors.items()]
    if player == "agent":
        return max(values)
    else:
        return min(values)