PhiBE-Q-Learning: Bridging Off-Policy Reinforcement Learning and Continuous-Time Control

Abstract

In this paper, we develop an off-policy method for continuous-time reinforcement learning (CTRL), where the system dynamics are governed by an unknown stochastic differential equation (SDE) and only discrete-time trajectory data are available. A central challenge is that the classical discrete-time state--action value function loses its action-dependent signal in the continuous-time limit (Baird, 1994; Jia and Zhou, 2023; Tallec et al., 2019). On the other hand, continuous-time control provides local notions such as the instantaneous advantage function $q(s,a)$, but these typically rely on state value function $V(s)$. To obtain a formulation with a single unknown function, we introduce a new definition of the state--action value function $Q(s,a)$ in CTRL and derive its governing equation. Building on the PhiBE approximation (Zhu, 2024; Zhu et al., 2025), we propose iterative algorithms to approximate the optimal $Q$-function in both model-based and model-free settings using only discrete-time off-policy data. Under linear function approximation, we establish error bounds and explicit convergence rates for the algorithm in the popoluation level. The finite-data algorithm is evaluated numerically.