Bridging from micro to macro by means of mesoscopic simulations in physics and biology

Politecnico di Torino, Torino, Italy

An impressive range of results has been obtained in the field of Molecular Dynamics in the last decade for systems up to the size of about 100 nm. Due to the enormous potential for technological applications at the macroscopic level it is necessary to go well beyond such a limit. However, the computational costs of increasing the size of the system of even one or two orders of magnitude are absolutely prohibitive. A suitable alternative may be to split up the problem according to the different length scales involved, i.e. Multiscale Modelling (MM). Several approaches have been proposed for MM: in general they start from atomistic simulations, replacing, however, the interatomic interactions with e.g. interactions among dislocations or, for larger systems, among dislocation densities or grains. However, since our major concern lies in the macroscopic domain (e.g. NDE or materials characterization), we propose to reverse this "bottom-up" strategy. The approach we propose starts with a very fine discretization of the macroscopic specimen under study, which brings us to the mesoscopic level. For this bridging from a microscopic to a macroscopic scale by means of mesoscopic simulations, we plan to use the local Interaction Simulation Approach (LISA), whose "local" nature makes it ideally suitable for the task in hand, particularly in the case of strong local heterogeneities. Due to the complexity and vastness of scope of our goal, rather than searching for a general recipe, we propose to investigate several "case studies", which should demonstrate the feasibility of the approach and its validity. A similar approach can be used in biology and medicine to bridge between cell dynamics and macroscopic problems, such as the simulation of growth and therapy of tumors.