MDP State object

mdp-ttt-state.py

class State:
    def __init__(self, player: Player, state_desc: State_Desc):
        # print(state_desc)
        assert (len(state_desc) == 9)
        self.possible_actions: List = [action[1].name for action in enumerate(Action) if state_desc[action[0]] == '-']
        self.p: Dict = {action: Qvalue_Frequency((RewardType(0), 0)) for action in self.possible_actions}
        self.v_star = -np.inf
        self.state_desc: State_Desc = state_desc
        self.policy = None  # Represents best action to be taken
        self.player: Player = player