J. Phys. B: At. Mol. Opt. Phys. 44, 154012 (2011)
We introduce a new approach to assess the error of control problems we aim to optimize. The method offers a strategy to define new control pulses that are not necessarily optimal but still able to yield an error not larger than some fixed a priori threshold, and therefore provide control pulses that might be more amenable for an experimental implementation. The formalism is applied to an exactly solvable model and to the Landau–Zener model, whose optimal control problem is solvable only numerically. The presented method is of importance for applications where a high degree of controllability of the dynamics of quantum systems is required.