{"id":"arxiv/2609.38745","url":"https://arxiv.org/abs/2609.38745","tags":["math.OC"],"score":0,"title":"Model-free Reinforcement Learning for Continuous Time and State: A Stochastic Maximum Principle Approach","domain":"math.OC","source":"arxiv","authors":["Lijun Bo","Yijie Huang","Jingfei Wang"],"reviews":0,"subject":"math.OC","summary":"This paper develops a model-free reinforcement learning (RL) algorithm based on the stochastic maximum principle for continuous-time stochastic control problems with continuous state and action spaces. For a parameterized Markovian policy, we establish the existence of the decoupling field for the adjoint backward stochastic differential equation, which allows the Hamiltonian gradient to be represented as a deterministic function of time and state. We then learn this Hamiltonian gradient directly from data and incorporate it into a policy gradient scheme with inexact gradients. We establish…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-30","reactions":{}}