Dresden 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

QI: Fachverband Quanteninformation

QI 13: Quantum Control

QI 13.1: Talk

Wednesday, March 11, 2026, 15:00–15:15, BEY/0245

A Statistical-Physics Approach to Quantum Control Landscape Exploration — •Malte Krug and Jürgen Stockburger — Institute for Complex Quantum Systems, Ulm, Germany

Conventional approaches to quantum optimal control rely on local optimizers that return a single high-fidelity pulse but offer little insight into the global structure of the cost landscape. A new method is presented, motivated by a statistical-physics inspired approach to stochastic control theory [1], that maps the quantum optimal control problem to an exploration of a high-dimensional landscape using ideas from protein-folding methods [2]. Instead of a single pulse, the method generates a distribution of control trajectories, represented by a Markov chain whose stationary distribution reflects the dominant regions of the cost landscape. This ensemble captures globally competitive pulses, reveals the diversity of near-optimal solutions, and allows for characterization of the local landscape through soft and stiff directions around optimal pulses. Moreover, the sampled trajectories provide high-quality initial guesses for conventional optimization methods, strongly biased toward the most promising regions of the landscape.
[1] Kappen, H. J.,Phys. Rev. Lett. 95, 200201 (2005).
[2] Trebst, S. & Troyer, M., in: Computer Simulations in Condensed Matter Systems, eds. Ferrario, M., Ciccotti, G. & Binder, K., Springer (2006).

Keywords: Quantum Control; Control Landscape; Markov Chain; Statistical Physics