Policy Optimization with Differentiable MPC: Convergence Analysis under Uncertainty
This addresses the challenge of model accuracy dependence in model predictive control for control applications, representing an incremental improvement.
The paper tackles the problem of model-based policy optimization for control applications by combining gradient-based policy optimization with recursive system identification, demonstrating convergence to optimal controller designs in several control examples.
Model-based policy optimization is a well-established framework for designing reliable and high-performance controllers across a wide range of control applications. Recently, this approach has been extended to model predictive control policies, where explicit dynamical models are embedded within the control law. However, the performance of the resulting controllers, and the convergence of the associated optimization algorithms, critically depends on the accuracy of the models. In this paper, we demonstrate that combining gradient-based policy optimization with recursive system identification ensures convergence to an optimal controller design and showcase our finding in several control examples.