This paper proposes a novel H-infinity model reduction framework for linear non-Markovian quantum systems, addressing the computational challenges posed by high dimensionality in augmented models. The approach formulates the model reduction as an optimization problem with physical realizability constraints, providing guaranteed worst-case error bounds essential for robust control applications.
Key findings
Proposes an H-infinity model reduction framework for non-Markovian quantum systems.
Derives conditions for physical realizability of reduced-order models.
Develops convex relaxation techniques for non-convex constraints.
Presents an iterative algorithm based on semidefinite programming.
Validates the method through numerical experiments on prototypical non-Markovian quantum systems.
Limitations & open questions
Further research needed for broader application and scalability.