Action Gaps and Advantages in Continuous-Time Distributional Reinforcement Learning
This addresses performance issues in high-frequency decision-making for RL applications like algorithmic trading, though it appears incremental as it extends existing distributional RL concepts.
The paper tackles the problem of reinforcement learning agents performing poorly at high decision frequencies by showing that distributional RL agents are also sensitive to decision frequency, with return distributions collapsing as frequency increases. It introduces a superiority distribution as a probabilistic generalization of advantage and demonstrates through simulations in option trading that modeling this distribution improves controller performance at high frequencies.
When decisions are made at high frequency, traditional reinforcement learning (RL) methods struggle to accurately estimate action values. In turn, their performance is inconsistent and often poor. Whether the performance of distributional RL (DRL) agents suffers similarly, however, is unknown. In this work, we establish that DRL agents are sensitive to the decision frequency. We prove that action-conditioned return distributions collapse to their underlying policy's return distribution as the decision frequency increases. We quantify the rate of collapse of these return distributions and exhibit that their statistics collapse at different rates. Moreover, we define distributional perspectives on action gaps and advantages. In particular, we introduce the superiority as a probabilistic generalization of the advantage -- the core object of approaches to mitigating performance issues in high-frequency value-based RL. In addition, we build a superiority-based DRL algorithm. Through simulations in an option-trading domain, we validate that proper modeling of the superiority distribution produces improved controllers at high decision frequencies.