Discretizing Continuous Action Space for On-Policy Optimization

Abstract

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.

Accepted at AAAI Conference on Artificial Intelligence (2020) in New York, NY, USA. An open source implementation can be found at https://github.com/robintyh1/onpolicybaselines

References in corpus (8)

• Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
• Addressing Function Approximation Error in Actor-Critic Methods
• Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation
• Learning Dexterous In-Hand Manipulation
• Simple random search provides a competitive approach to reinforcement learning
• Fast Approximate Natural Gradient Descent in a Kronecker-factored Eigenbasis
• Latent Space Policies for Hierarchical Reinforcement Learning
• Implicit Policy for Reinforcement Learning

Cited by in corpus (5)

• Discrete and Continuous Action Representation for Practical RL in Video Games
• Consensus Multi-Agent Reinforcement Learning for Volt-VAR Control in Power Distribution Networks
• Discrete Action On-Policy Learning with Action-Value Critic
• PFPN: Continuous Control of Physically Simulated Characters using Particle Filtering Policy Network
• RLCache: Automated Cache Management Using Reinforcement Learning