Incorporating Rivalry in Reinforcement Learning for a Competitive Game