Estimating relaxation times from a single trajectory

Takuma Akimoto,Eiji Yamamoto,Takashi Uneyama
DOI: https://doi.org/10.7566/JPSJ.92.074005
2023-05-19
Abstract:Complex systems such as protein conformational fluctuations and supercooled liquids exhibit a long relaxation time and are considered to posses multiple relaxation times. We analytically obtain the exact correlation function for stochastic processes with multiple relaxation times. We show that the time-averaged correlation function calculated by a trajectory whose length is shorter than the longest relaxation time exhibits an apparent aging behavior. We propose a method to extract relaxation times from a single trajectory. This method successfully extracts relaxation times of a stochastic process with multiple states when a state can be characterized by the values of the trajectory. As an application of this method, we estimate several relaxation times smaller than the longest relaxation time in conformational fluctuations of a small protein.
Soft Condensed Matter,Data Analysis, Statistics and Probability
What problem does this paper attempt to address?