Learning-based control methods typically assume stationary system dynamics, an assumption often violated in real-world systems due to drift, wear, or changing operating conditions. We study reinforcement learning for control under time-varying dynamics. We consider a continual model-based reinforcement learning setting in which an agent repeatedly learns and controls a dynamical system whose transition dynamics evolve across episodes. We analyze the problem using Gaussian process dynamics models under frequentist variation-budget assumptions. Our analysis shows that persistent non-stationarity requires explicitly limiting the influence of outdated data to maintain calibrated uncertainty and meaningful dynamic regret guarantees. Motivated by these insights, we propose a practical optimistic model-based reinforcement learning algorithm with adaptive data buffer mechanisms and demonstrate improved performance on continuous control benchmarks with non-stationary dynamics.

@article{iten2026model,
	author = {Klemens Iten and Bruce Lee and Chenhao Li and Lenart Treven and Andreas Krause and Bhavya Sukhija},
	journal = {arXiv preprint arXiv:2604.02260},
	month = {April},
	day = {2},
	pdf = {https://arxiv.org/pdf/2604.02260},
	title = {Model-Based Reinforcement Learning for Control under Time-Varying Dynamics},
	year = {2026}}