hide
Free keywords:
statistical sampling; efficient simulations; Langevin equation;
Markov chain Monte Carlo; deep learning
Abstract:
This thesis deals with the development and formalization of algorithms designed for an efficient simulation of biological systems. This work is separated into two different parts, and in each part a different algorithm is investigated. In the first part of the thesis, an algorithm that is used to simulate biological systems at the mesoscopic scale is outlined. The aforementioned algorithm is studied in detail, and several improvements, theoretical, algorithmic and technical, are presented. In the second part of the thesis, a novel sampling method is outlined, which uses deep-learning to accelerate the computation of equilibrium properties of systems defined with atomistic detail. The two parts lead to applications at different scales, and, in the future, methods and concepts developed in this thesis can be useful for the investigation of biological processes defined with mesoscopic or microscopic detail.
