A systematic assessment framework is presented that compares four prominent DRL controllers with a classical control baseline across a diverse set of applied control problems, including non-minimum phase dynamics, flexible mechanical systems, nonlinear marine control, and aerial robotics, and clarifies the trade-offs between learning-based and conventional control.
A generalized framework that combines the adaptive, high-performance nature of deep reinforcement learning (DRL) with the formal safety guarantees of model predictive control (MPC) is proposed, demonstrating successful exploration and stable policy convergence on physical hardware.
George Schafer, Jakob Rehrl, Stefan Huber et al.· 0 citations
Advanced control of heating and cooling systems can substantially reduce energy costs and pollution. However, real-world adoption of popular algorithms among researchers, such as model predictive control (MPC) and reinforcement learning (RL), remains limited due in part to their high deployment and commissioning costs....
W. G. Dierking, Arash Khabbazi, Levi D. Reyes Premer et al.· arXiv.org· 0 citations
Improvements of 13% under normal conditions and 17% in the presence of system noise achieved by the proposed approach, compared to integer-order difference (IOD)-RL are demonstrated.
Mohsen Shiri, Hadi Delavari· International Journal of Dyn...· 1 citation
A deep reinforcement learning (DRL)-based PID online gain scheduling control framework for precise tracking of a DC motor achieves higher tracking accuracy and effectively mitigates control signal oscillations, thereby enhancing overall control robustness and practical feasibility.
This work begins with dynamic modeling using the Euler-Lagrange formulation and demonstrates that the proposed hybrid controller outperforms traditional methods in terms of tracking accuracy, settling time, and disturbance rejection.
Anita Verma· International Journal of Int...· 0 citations
PEARL employs an actor-adjoint algorithm that leverages automatic differentiation to compute policy gradients over short horizons and adjoint-based sensitivities of future returns approximated via neural networks, significantly reducing the number of environment interactions, while mitigating long-term gradient instabi...
