Optimum Order Estimation of Reduced Macromodels Based on a Geometric Approach for Projection-Based MOR Methods

Estimation of the optimal order of reduced models in existing macromodeling techniques is a challenging task and is often based on heuristics. In this paper, a new algorithm is described for estimating the minimum acceptable order for reduced models of linear systems to ensure accurate as well as efficient transient behavior. The precise determination of the optimum order for a reduced system is based on evaluation of the number of false nearest neighbors.