A renormalization approach to quantum thermal annealing

The details of an efficient global optimization approach, quantum thermal a nnealing with renormalization (QTAR) (Y. H. Lee and B. J. Berne, J. Phys. C hem. A, in press (2000)) are presented in this paper. This method is based on the application of the Migdal-Kadanoff method for decimating Trotter tim e slices in the staging and primitive algorithms for sampling path integral s using Monte Carlo methods. In a nutshell, one starts in a strong quantum regime where the number of Trotter beads representing each quantum particle and the Value of Planck's constant are large, thereby allowing for efficie nt tunneling through the barriers of a rough energy landscape typical in th e folding of proteins, and anneals the system methodically to the classical limit where the values of the aforementioned quantities are 1 and 0, respe ctively. Global optimization of the system is achieved through the iterativ e use of such quantum-to-classical annealing cycles. The QTAR algorithm app lied to a highly frustrated BLN model protein with 46 residues more efficie ntly locates the global energy minimum than established methods like simula ted annealing.