Emergency action termination for immediate reaction in hierarchical reinforcement learning
This addresses a gap in hierarchical RL for large dynamical systems, offering an incremental improvement by combining hierarchical and flat RL advantages.
The paper tackles the problem of hierarchical reinforcement learning (RL) where higher-level goals may become obsolete due to environmental randomness, leading to inefficiencies. They propose a method that constantly verifies and replaces inadequate goals, achieving fast training and immediate reactivity, as demonstrated experimentally on seven benchmark environments.
Hierarchical decomposition of control is unavoidable in large dynamical systems. In reinforcement learning (RL), it is usually solved with subgoals defined at higher policy levels and achieved at lower policy levels. Reaching these goals can take a substantial amount of time, during which it is not verified whether they are still worth pursuing. However, due to the randomness of the environment, these goals may become obsolete. In this paper, we address this gap in the state-of-the-art approaches and propose a method in which the validity of higher-level actions (thus lower-level goals) is constantly verified at the higher level. If the actions, i.e. lower level goals, become inadequate, they are replaced by more appropriate ones. This way we combine the advantages of hierarchical RL, which is fast training, and flat RL, which is immediate reactivity. We study our approach experimentally on seven benchmark environments.