A stochastic quasi Newton method for molecular simulations

Abstract

In this thesis the Langevin equation with a space-dependent alternative mobility matrix has been considered. Simulations of a complex molecular system with many different length and time scales based on the fundamental equations of motion take a very long simulation time before capturing the functional and relevant motions. This problem is called critical slowing down. To avoid this problem multi-scale simulation methods are applied, which permits the use of different size time steps and thus enables acceleration of the relevant (slow) movements. The aim of this thesis is to develop a stochastic quasi Newton method, such that by incorporating multi-scaling the relevant motions are effectively taken with larger time step. Due to the integration of the slow motions with a larger time step, critical slowing down can be avoided. The proposed stochastic quasi Newton method enables automatic multi-scaling in the Langevin dynamics and contributes to efficient calculation of the noise term. The construction of the proposed method also enables the construction of a limited memory version for the mobility. This results in a method where less storage is needed. Together with the reduction of the computation time and the multi-scaling property, a powerful method for molecular simulations has been provided.