An Improved Frequency Weighted Balanced Reduction Technique

In this paper,we present an improved frequency weighted balanced reduction technique.The method is based on an accuracy-enhancing idea.The reduced order models are guaranteed to be stable in case of double-sided weighting.Meanwhile,the approximation error is relatively smaller than other well-known methods.Simple,elegant and easily calculated a priori error bounds are also derived.Two numerical examples and comparison with other well-known techniques show the effectiveness of the proposed technique.