Deconstructing Actor-Critic: A Large-scale Empirical Study of Design Components for Practitioners
arXiv:2607.13274
The paper empirically evaluates how design choices in actor‑critic reinforcement learning algorithms affect performance and robustness on a real‑world water‑treatment control task, analyzing over 33,000 experiments.
Abstract
Reinforcement learning is increasingly being considered for controlling real-world systems, from fusion plasma and autonomous vehicles to drug discovery and drinking water treatment, where reliability is essential and tuning budgets are limited. Actor-critic algorithms share a set of design decisions, such as how the policy is updated, how it represents the distribution over actions, how its gradient is estimated, and how often it is updated relative to the value estimator. Using a control task derived from a real water treatment plant, we analyze over 33,000 experiments to determine how these components affect variability across runs and sensitivity to hyperparameters. Common defaults, such as Gaussian action distributions with pathwise gradient estimators, are among the least reliable configurations, whereas bounded distributions with adaptive update schedules remain robust across a wide range of settings. These findings offer empirical guidance to practitioners across scientific and engineering domains for understanding and making component-level decisions when adapting actor-critic methods to new real-world control settings.
38 pages, 22 figures