TY - JOUR
T1 - Experimental quantum control landscapes
T2 - Inherent monotonicity and artificial structure
AU - Roslund, Jonathan
AU - Rabitz, Herschel
PY - 2009/8/6
Y1 - 2009/8/6
N2 - Unconstrained searches over quantum control landscapes are theoretically predicted to generally exhibit trap-free monotonic behavior. This paper makes an explicit experimental demonstration of this intrinsic monotonicity for two controlled quantum systems: frequency unfiltered and filtered second-harmonic generation (SHG). For unfiltered SHG, the landscape is randomly sampled and interpolation of the data is found to be devoid of landscape traps up to the level of data noise. In the case of narrow-band-filtered SHG, trajectories are taken on the landscape to reveal a lack of traps. Although the filtered SHG landscape is trap free, it exhibits a rich local structure. A perturbation analysis around the top of these landscapes provides a basis to understand their topology. Despite the inherent trap-free nature of the landscapes, practical constraints placed on the controls can lead to the appearance of artificial structure arising from the resultant forced sampling of the landscape. This circumstance and the likely lack of knowledge about the detailed local landscape structure in most quantum control applications suggests that the a priori identification of globally successful (un)constrained curvilinear control variables may be a challenging task.
AB - Unconstrained searches over quantum control landscapes are theoretically predicted to generally exhibit trap-free monotonic behavior. This paper makes an explicit experimental demonstration of this intrinsic monotonicity for two controlled quantum systems: frequency unfiltered and filtered second-harmonic generation (SHG). For unfiltered SHG, the landscape is randomly sampled and interpolation of the data is found to be devoid of landscape traps up to the level of data noise. In the case of narrow-band-filtered SHG, trajectories are taken on the landscape to reveal a lack of traps. Although the filtered SHG landscape is trap free, it exhibits a rich local structure. A perturbation analysis around the top of these landscapes provides a basis to understand their topology. Despite the inherent trap-free nature of the landscapes, practical constraints placed on the controls can lead to the appearance of artificial structure arising from the resultant forced sampling of the landscape. This circumstance and the likely lack of knowledge about the detailed local landscape structure in most quantum control applications suggests that the a priori identification of globally successful (un)constrained curvilinear control variables may be a challenging task.
UR - http://www.scopus.com/inward/record.url?scp=68549133378&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=68549133378&partnerID=8YFLogxK
U2 - 10.1103/PhysRevA.80.013408
DO - 10.1103/PhysRevA.80.013408
M3 - Article
AN - SCOPUS:68549133378
SN - 1050-2947
VL - 80
JO - Physical Review A - Atomic, Molecular, and Optical Physics
JF - Physical Review A - Atomic, Molecular, and Optical Physics
IS - 1
M1 - 013408
ER -