Abstract:Model predictive control (MPC) solves a receding-horizon optimization problem in real-time, which can be computationally demanding when there are thousands of constraints. To accelerate online computation of MPC, we utilize data to adaptively remove the constraints while maintaining the MPC policy unchanged. Specifically, we design the removal rule based on the Lipschitz continuity of the MPC policy. This removal rule can use the information of historical data according to the Lipschitz constant and the distance between the current state and historical states. In particular, we provide the explicit expression for calculating the Lipschitz constant by the model parameters. Finally, simulations are performed to validate the effectiveness of the proposed method.
What problem does this paper attempt to address?
The paper attempts to address the issue of excessive computational burden in real-time optimization calculations in Model Predictive Control (MPC) when there are a large number of constraints. Specifically, the paper proposes a method to adaptively remove constraints using historical data to accelerate online MPC computations while maintaining the MPC strategy unchanged.
### Background and Motivation
- **Challenges of MPC**: MPC calculates control inputs by solving an optimization problem over a finite time horizon at each time step to ensure the system meets the constraints. However, when there are a large number of constraints, solving this optimization problem becomes computationally intensive, especially in linear dynamic systems.
- **Limitations of existing methods**: Existing methods to accelerate MPC computations include model reduction, explicit MPC, custom solvers, and constraint removal. Among these, constraint removal methods further reduce computation time by utilizing online data, but existing methods mainly rely on current data and fail to fully utilize historical data.
### Main Contributions of the Paper
- **New constraint removal method**: The paper proposes a new constraint removal method that uses historical state-input data to further accelerate online computations.
- **Utilization of Lipschitz continuity**: By analyzing the Lipschitz continuity of the MPC strategy in the state space, the paper designs a constraint removal rule. The Lipschitz constant can be computed offline and used to evaluate and remove redundant constraints based on historical data.
- **Theoretical guarantee**: The paper proves that the proposed constraint removal method does not alter the closed-loop behavior of the system, meaning the optimal solution of MPC remains unchanged after removing constraints.
### Method Overview
1. **Calculation of Lipschitz constant**: The paper provides an explicit expression for the Lipschitz constant, which can be computed offline and depends only on the model parameters.
2. **Design of outer approximation set**: Using Lipschitz continuity, an outer approximation set \( M(x) \) is designed to evaluate and remove redundant constraints.
3. **Algorithm implementation**: The paper proposes a constraint removal algorithm based on historical data, which measures the current state at each time step, finds the historical state closest to the current state, and removes redundant constraints based on historical data.
### Experimental Validation
- **Simulation results**: The paper validates the effectiveness of the proposed method through simulation experiments on a double integrator system. The results show that the proposed adaptive constraint MPC method can significantly reduce the number of constraints and greatly improve computation speed while maintaining the same trajectory as the original MPC.
### Conclusion
The paper proposes an MPC method that adaptively removes constraints using historical data. By analyzing Lipschitz continuity and designing an outer approximation set, the method significantly reduces computational burden and improves computation speed without altering the closed-loop behavior of the system. Experimental results validate the effectiveness of the proposed method.