Bidirectional-Reachable Hierarchical Reinforcement Learning with Mutually Responsive Policies
This addresses a bottleneck in HRL for complex long-horizon tasks, though it is incremental as it builds on existing HRL methods.
The paper tackles the problem of subgoal reachability in hierarchical reinforcement learning by introducing a mutual response mechanism for real-time bilateral information sharing and error correction, resulting in BrHPO outperforming state-of-the-art baselines with higher exploration efficiency and robustness.
Hierarchical reinforcement learning (HRL) addresses complex long-horizon tasks by skillfully decomposing them into subgoals. Therefore, the effectiveness of HRL is greatly influenced by subgoal reachability. Typical HRL methods only consider subgoal reachability from the unilateral level, where a dominant level enforces compliance to the subordinate level. However, we observe that when the dominant level becomes trapped in local exploration or generates unattainable subgoals, the subordinate level is negatively affected and cannot follow the dominant level's actions. This can potentially make both levels stuck in local optima, ultimately hindering subsequent subgoal reachability. Allowing real-time bilateral information sharing and error correction would be a natural cure for this issue, which motivates us to propose a mutual response mechanism. Based on this, we propose the Bidirectional-reachable Hierarchical Policy Optimization (BrHPO)--a simple yet effective algorithm that also enjoys computation efficiency. Experiment results on a variety of long-horizon tasks showcase that BrHPO outperforms other state-of-the-art HRL baselines, coupled with a significantly higher exploration efficiency and robustness.