Reinforced Model Predictive Control via Trust-Region Quasi-Newton Policy Optimization

TL;DR

This paper introduces a trust-region Quasi-Newton method for optimizing model predictive control policies, achieving faster convergence and better data efficiency compared to first-order reinforcement learning methods.

Contribution

It develops a novel second-order policy optimization algorithm for model predictive control using Quasi-Newton updates with trust-region constraints.

Findings

Outperforms first-order RL algorithms in data efficiency

Achieves superlinear convergence rate

Effective second-order derivative computation via linear system solutions

Abstract

Model predictive control can optimally deal with nonlinear systems under consideration of constraints. The control performance depends on the model accuracy and the prediction horizon. Recent advances propose to use reinforcement learning applied to a parameterized model predictive controller to recover the optimal control performance even if an imperfect model or short prediction horizons are used. However, common reinforcement learning algorithms rely on first order updates, which only have a linear convergence rate and hence need an excessive amount of dynamic data. Higher order updates are typically intractable if the policy is approximated with neural networks due to the large number of parameters. In this work, we use a parameterized model predictive controller as policy, and leverage the small amount of necessary parameters to propose a trust-region constrained Quasi-Newton training algorithm for policy optimization with a superlinear convergence rate. We show that the required second order derivative information can be calculated by the solution of a linear system of equations. A simulation study illustrates that the proposed training algorithm outperforms other algorithms in terms of data efficiency and accuracy.