Discretizing Continuous Action Space for On-Policy Optimization
This method addresses the problem of efficient on-policy optimization for high-dimensional continuous control tasks, offering a simple yet effective solution for researchers and practitioners in reinforcement learning.
The paper tackles the challenge of continuous control in reinforcement learning by discretizing the action space, showing that this approach leads to significant performance gains, such as improved results with PPO, TRPO, and ACKTR on high-dimensional tasks.
In this work, we show that discretizing action space for continuous control is a simple yet powerful technique for on-policy optimization. The explosion in the number of discrete actions can be efficiently addressed by a policy with factorized distribution across action dimensions. We show that the discrete policy achieves significant performance gains with state-of-the-art on-policy optimization algorithms (PPO, TRPO, ACKTR) especially on high-dimensional tasks with complex dynamics. Additionally, we show that an ordinal parameterization of the discrete distribution can introduce the inductive bias that encodes the natural ordering between discrete actions. This ordinal architecture further significantly improves the performance of PPO/TRPO.