Complexity reduction of a thin film deposition model using a trajectory based nonlinear model reduction technique

The paper proposes the use of a complexity reduction procedure based on nonlinear system identification techniques on a kinetic Monte Carlo simulation model of thin film deposition. This procedure combines frequency domain identification with nonlinear structure selection techniques and generalised basis function parameterization of a Volterra structure. It is used successfully to model the roughness response to time-varying temperature inputs during the deposition of germanium on a cubic Ge(001) lattice, reducing the original model by three orders of magnitude.