Balancing Two-Player Stochastic Games with Soft Q-Learning
This work addresses the need for adjustable AI difficulty in video games, offering a method to avoid overly adversarial agents, though it is incremental as it builds on existing soft Q-learning techniques.
The paper tackles the problem of creating more engaging and tunable AI opponents in video games by generalizing soft Q-learning to stochastic games, enabling a continuous spectrum of gaming behaviors beyond optimal performance. The result includes theoretical proof of unique game values and empirical demonstration of reliable game balancing with high-dimensional representations.
Within the context of video games the notion of perfectly rational agents can be undesirable as it leads to uninteresting situations, where humans face tough adversarial decision makers. Current frameworks for stochastic games and reinforcement learning prohibit tuneable strategies as they seek optimal performance. In this paper, we enable such tuneable behaviour by generalising soft Q-learning to stochastic games, where more than one agent interact strategically. We contribute both theoretically and empirically. On the theory side, we show that games with soft Q-learning exhibit a unique value and generalise team games and zero-sum games far beyond these two extremes to cover a continuous spectrum of gaming behaviour. Experimentally, we show how tuning agents' constraints affect performance and demonstrate, through a neural network architecture, how to reliably balance games with high-dimensional representations.