Chaos and momentum diffusion of the classical and quantum kicked rotor

The de Broglie-Bohm (BB) approach to quantum mechanics gives trajectories similar to classical trajectories except that they are also determined by a quantum potential. The quantum potential is a "fictitious potential" in the sense that it is part of the quantum kinetic energy. We use quantum trajectories to treat quantum chaos in a manner similar to classical chaos. For the kicked rotor, which is a bounded system, we use the Benettin, et al. method to calculate both classical and quantum Lyapunov exponents as a function of control parameter K and find chaos in both cases. Within the chaotic sea we find in both cases nonchaotic stability regions for K equal to multiples of pi. For even multiples of pi the stability regions are associated with classical accelerator mode islands and for odd multiples of pi they are associated with new oscillator modes. We examine the structure of these regions.; Momentum diffusion of the quantum kicked rotor is studied with both BB and standard quantum mechanics (SQM). A general analytical expression is given for the momentum diffusion at quantum resonance of both BB and SQM. We obtain agreement between the two approaches in numerical experiments. For the case of nonresonance the quantum potential is not zero and must be included as part of the quantum kinetic energy for agreement. The numerical data for momentum diffusion of classical kicked rotor is well fit by a power law DNb. in the number of kicks N. In the anomalous momentum diffusion regions due to accelerator modes the exponent beta(K) is slightly less than quadratic, except for a slight dip, in agreement with an upper bound (K2/2)N 2. The corresponding coefficient D(K) in these regions has three distinct sections, most likely due to accelerator modes with period greater than one.; We also show that the local Lyapunov exponent of the classical kicked rotor has a plateau for a duration that depends on the initial separation and then decreases asymptotically as O(t-1 Int), where t is the time. This behavior is consistent with an upper bound that is determined analytically.