Flow-based Policy With Distributional Reinforcement Learning in Trajectory Optimization
This work addresses the problem of capturing multimodal solutions in reinforcement learning for control tasks, offering an incremental improvement over existing methods.
The paper tackles the limitations of traditional RL policies, which use diagonal Gaussian distributions and mean returns, by proposing FP-DRL, a method combining flow-based policies with distributional RL to handle multimodal distributions and improve performance. Experiments on MuJoCo benchmarks show that FP-DRL achieves state-of-the-art results in most tasks.
Reinforcement Learning (RL) has proven highly effective in addressing complex control and decision-making tasks. However, in most traditional RL algorithms, the policy is typically parameterized as a diagonal Gaussian distribution, which constrains the policy from capturing multimodal distributions, making it difficult to cover the full range of optimal solutions in multi-solution problems, and the return is reduced to a mean value, losing its multimodal nature and thus providing insufficient guidance for policy updates. In response to these problems, we propose a RL algorithm termed flow-based policy with distributional RL (FP-DRL). This algorithm models the policy using flow matching, which offers both computational efficiency and the capacity to fit complex distributions. Additionally, it employs a distributional RL approach to model and optimize the entire return distribution, thereby more effectively guiding multimodal policy updates and improving agent performance. Experimental trails on MuJoCo benchmarks demonstrate that the FP-DRL algorithm achieves state-of-the-art (SOTA) performance in most MuJoCo control tasks while exhibiting superior representation capability of the flow policy.