QuantFPFlow: Quantum Amplitude Estimation for Fokker--Planck Policy Optimisation in Continuous Reinforcement Learning

In three linesQuantFPFlow integrates quantum amplitude estimation (Grover) into stochastic policy optimization via Fokker-Planck formulation. Provable quadratic speedup O(1/ε) vs O(1/ε²) classical. On continuous multimodal task, outperforms SAC (1295.7 vs 1284.0 reward) and finds global optimum 10.4% more frequently (33.9% vs 30.7%).

Summary generated by Claude — human-verified