def reward_function(self, agent: Agent):
max_score = 0
for player in self.players:
score = self.scores[player.index]
if score > max_score:
max_score = score
if self.terminal_test():
if self.scores[agent.index] == max_score:
reward = 100
else:
reward = -150
else:
multiplier = agent.get_multiplier_last_action()
player_score_intial = agent.last_score
current_score = self.scores[agent.index]
player_won_turn = current_score > player_score_intial
if player_won_turn:
reward = 5 * multiplier
else:
reward = -10 * multiplier
return reward
