Exploring the Space of Coarse-Grained Models

Using the exactly renormalizable Gaussian network model, we extend upon a previous study which explored the impact of resolution upon information and entropy in coarse-grained models. In this work, we exploit an intuitive decomposition of the coarse-grained Potential of Mean Force (PMF) under a given mapping into entropic and energetic terms. Focusing on the entropic term as a measure of information loss, we explore the space of all mappings using Monte Carlo simulations in order to characterize the structure and features of this space. Applying a statistical mechanical analysis to this system yields valuable insight into the "mapping problem" of coarse-grained modeling.