Quantum mechanical approaches are attracted in computational sciences. The quantum annealing is a novel technique for optimization of various disordered problems. It utilizes the dynamical motion of quantum state driven by handling quantum fluctuations. As an important direction, we focus on realistic numerical methods to carry out the quantum annealing in classical computers. The mean field quantum annealing, which we discuss, is a rough but non-trivial method. It is applicable to large problem sizes and yields an answer fast. However it is not clear how the mean field method is valid. In our study we investigate the validity of this method for elementary models and clarify the property of this method in comparison with other known optimization techniques. In my presentation, I will report our results of numerical calculation and discuss the limitation of the mean field quantum annealing. I will also present an improvement of the mean field approximation. |
|
|