Model order reduction for delay systems by iterative interpolation

Abstract

Adaptive algorithms for computing the reduced-order model of time-delay systems (TDSs) are proposed in this work. The algorithms are based on interpolating the transfer function at multiple expansion points and greedy iterations for selecting the expansion points. The (Formula presented.) -error of the reduced transfer function is used as the criterion for choosing the next new expansion point. One heuristic greedy algorithm and one algorithm based on the error system and adaptive sub-interval selection are developed. Results on four TDSs with tens of delays from electromagnetic applications are presented and show the efficiency of the proposed algorithms.