Heidelberg 2015 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

Q: Fachverband Quantenoptik und Photonik

Q 10: Quantum Information: Concepts and Methods II

Q 10.2: Vortrag

Montag, 23. März 2015, 15:00–15:15, K/HS1

Landscape Engineering: Removing local traps for state to state transfer — •Niklas Rach, Matthias Müller, Tommaso Calarco, and Simone Montangero — Institut für komplexe Quantensysteme, Universität Ulm, D-89069 Ulm, Germany

In Quantum Optimal Control, the success of optimization algorithms are drastically determined by the landscape of the cost functional F. It is known, that the landscape of an unconstrained state to state transfer problem in a controllable system contains only global maxima, however, the landscape of constrained optimizations are characterized by the presence of traps [1,2]. Considering an Ising chain with broken symmetry we study the influence of these traps in the Chopped Random Bases algorithm (CRAB), as the expansion of the control into a truncated basis introduces a constraint on the control. We show that with increasing number of basis functions the success probability converges to one. In addition, we introduce an iterative version of CRAB which allows to engineer the landscape in a way that removes these false traps and converges always to a global maximum regardless of the number of basis functions involved.

[1] Herschel A. Rabitz et al. Science 303 , 1998 (2004).

[2] K.W. Moore, H. Rabitz, J. Chem. Phys. 137, 134113 (2012).