Zur NavigationLogin



Accelerated-MD & MC

Mittwoch, 20. September 2017 10:00 Uhr;

JGU Mainz, Physics, Lorentz-Raum

Sprecher: Giovanni Cicotti; University of Rome, La Sapienza

The talk addresses a case of the challenges associated, for MD/MC practitioners, with rare events, events whose statistical properties cannot be computed by brute force simulations. In particular we present a technique, useful to compute the free energy landscape associated with activated processes, and illustrate it in a simple but instructive case. The technique derives from a procedure, introduced by E. Vanden Eijnden & L. Maragliano (Chem. Phys. Lett. 426, 168 (2006); Single-sweep methods for free energy calculations, J. Chem. Phys., 2008, 128, 184110), called TAMD, which gets the free energy landscape, in few collective variables (CV’s) describing the process of interest, by extending the phase space of the system with the addition of the collective variables taken as new, slow, degrees of freedom, coupled to the original system by a suitable biasing potential. The extended system can be let to evolve for the fast, physical, variables at the physical temperature while the slow variables are thermostatted at a much higher temperature. The result is that the system can be shown to evolve sampling the full free energy landscape in spite of the presence of rare events. In the case, frequent in applications, when the CV’s cannot be expresses in a analytical form, the procedure can be substituted by a temperature accelerated Monte Carlo (TAMC) since the MC “evolution” needs only the energies and no more the forces associated to the sampling. The new techniques will be demonstrated and, then, illustrated with the characterization of a simple nucleation process.


iCal-Datei herunterladen


Diese Seite verwendet Cookies. Mit der Nutzung dieser Seite erklären Sie sich damit einverstanden, dass wir Cookies setzen.

Weitere Informationen