Optimization-Free Fast Optimal Control: Bang-Ride Property, Monotonicity, and Applications to Fast Battery Charging

Shengling Shi, Jacob Sass, Jiaen Wu, Minsu Kim, Yingjie Ma, Sungho Shin, Richard D. BraatzarXiv 2025 · DOI · arXiv

Introduction

Solving optimal control problems (OCPs) in real-time is a significant computational hurdle, particularly for applications like fast battery charging. High-fidelity electrochemical models involve complex nonlinear constraints that typically require iterative optimization. This paper introduces a theoretical foundation for "bang-ride" control, demonstrating that for a specific class of systems, these complex problems can be replaced by simple, high-speed simulations that are orders of magnitude faster than traditional methods while remaining mathematically optimal.

Theoretical Framework: Monotonicity and Bang-Ride

The core of the paper's contribution lies in identifying the conditions under which an optimal control strategy follows a "bang-ride" structure. In this context, "bang" refers to operating at the maximum or minimum input limit, while "ride" refers to operating exactly on a state or path constraint boundary.

The authors leverage Pontryagin’s Maximum Principle (PMP) to analyze systems with scalar inputs. They identify two critical properties:

  1. Monotonicity: In many physical systems, a larger input consistently leads to a larger state value. If the objective is to reach a target state as quickly as possible, the intuition is to apply the maximum feasible input at all times.
  2. External Positivity: For linear systems, the authors show that if the costate dynamics (representing the sensitivity of the cost to state changes) remain positive, the bang-ride strategy is guaranteed to be optimal.

The Hybrid Simulation Algorithm

Based on these theoretical insights, the authors propose the Hybrid Simulation algorithm. Instead of using a gradient-based optimizer, the system is simulated forward in time. The algorithm applies the maximum input until a state constraint is reached, at which point it switches to the specific input required to "ride" along that constraint boundary.

The hybrid simulation perfectly matches the optimal input profile when monotonicity and positivity conditions are met.

This approach transforms a complex optimization problem into a straightforward integration task, which can be executed on low-power embedded hardware in milliseconds. This is a massive improvement over traditional gradient descent methods, which can take years of CPU time for detailed battery models.

Applications to Fast Battery Charging

The paper applies this theory to the Single-Particle Model (SPM), a standard representation of lithium-ion battery dynamics. A major result is the formal mathematical proof that the widely used Constant-Current Constant-Voltage (CC-CV) charging profile is indeed the optimal solution for minimizing charging time under voltage constraints in the SPM.

The hybrid simulation maintains the system on the constraint boundary, matching the optimal trajectory.

By using the hybrid simulation approach, the authors achieved computational speedups of several orders of magnitude compared to conventional solvers. This allows for high-fidelity, optimal charging strategies to be implemented in real-time Battery Management Systems (BMS), providing a mathematical "certificate" for the optimality of simple charging rules.

Conclusion and Future Work

While the bang-ride property provides a powerful shortcut for many control problems, the authors note that it is not universal. The optimality of the hybrid simulation depends on the positivity of the costate variables. Future research will focus on extending these results to systems with multiple inputs and more complex "mixed" monotonicity, where different constraints may respond differently to input changes. This work paves the way for a new generation of "optimization-free" controllers that combine mathematical rigor with extreme computational efficiency.

Reference: Shi, S., Sass, J., Wu, J., et al. (2025). Bang-Ride Optimal Control: Monotonicity, External Positivity, and Fast Battery Charging. arXiv:2508.09010v3.

Summary updated

BibTeX
@misc{shi2025optimization,
  author = {Shengling Shi and Jacob Sass and Jiaen Wu and Minsu Kim and Yingjie Ma and Sungho Shin and Richard D. Braatz},
  title = {{Optimization-Free Fast Optimal Control: Bang-Ride Property, Monotonicity, and Applications to Fast Battery Charging}},
  year = {2025},
  eprint = {2508.09010},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2508.09010},
}
Figure