An isolated from the environment n-level controlled quantum system is described by the Schrödinger equation for unitary evolution operator U t :Here H 0 and V are n × n Hermitian matrices and f (t) ∈ L 1 ([0, T ]; R) is the control. We assume that [H 0 , V ] = 0. The goal is to find a control f (t) which maximizes the objectivewhere ρ 0 is the initial density matrix of the system, A is a Hermitian matrix and T > 0 is the final time. The functional J A describes the average value of an observable A at time T . An important problem in quantum control is the analysis of local maxima and minima (called traps) of the objective functional [1, 2, 3] because traps, if they exist, can be obstacles to find globally optimal solutions. The absence of traps in typical control problems for quantum systems was suggested in [1,2]. The absence of traps for two-level quantum systems for sufficiently large T was proved in [4,5]. The influence of constraints in the controls on the appearance of traps for two-level Landau-Zener system was investigated in [6]. For constrained control of this system with sufficiently large T only a special kind of traps may exist. These traps can be eliminated by a modification of gradient search algorithm so that in this sense the quantum control landscape practically appears as trap-free [6]. In [5] the following statement is proved.