Dynamic coarse graining in complex system simulation

The high computational burden in complex system simulation, particularly for a polymer system, prohibits long term simulation that provides information to predict the system attributes. We propose a coarse graining simulation method, stemming from our earlier work on state reduction through a modified local feature analysis (LFA), so that, based on short term system dynamics, one can automatically identify a low number of "seeds" based on correlations in the dynamic motion of all states. The trajectories of the "seeds" are then extrapolated. A simple matrix transformation is proposed to calculate trajectories of the whole system from the extrapolated "seed" trajectories. As the recovered system dynamics are derived from the low dimensional seed trajectories, we call it coarse grained dynamics. Simulation is carried out to illustrate the application of the developed algorithm to the PVC polymer dynamics.