Optimal control of molecular states in a learning loop with a parameterization in frequency and time domain

Learning loops, consisting of femtosecond pulse shaper, feedback signal and evolutionary algorithm, can be used to optimize coherent control processes in complex molecular systems. Using different parameterizations for the al gorithm-shaper-interface not only in the frequency but also in the time dom ain, where the genes of the algorithm directly represent the applied electr ic field, the optimization process can be dramatically improved and gives d irect insight about the physical mechanism responsible for the achieved con trol. With the prototype system K-2 combined with fs-DFNM we demonstrate co ntrol over the vibrational dynamics through parameterizations of linear chi rp, impulsive-stimulated-Raman-scattering, and phase-locked double pulses. (C) 2000 Elsevier Science B.V. All rights reserved.