Kron-based Model-order Reduction of Open Mass-action Kinetics Chemical Reaction Networks
Abstract
We propose a Kron-based model-order reduction method for mass-action kinetics chemical reaction networks (CRN) with constant inflow and proportional outflow. The reduced-order models preserve the CRN structure and we establish that the resulting reduced-order models have the same DC-gain or zero-moment as that of the full-order ones. Subsequently, we present the spectrum interlacing property of the Kron-reduced open CRN and propose the use of Gramians-based approach for single-species single-substrate chemical network to get the upper-bound of approximation error and to use it in determining a good set of nodes to be removed systematically. Finally, we evaluate the applicability and efficacy of our results in two well-known biochemical kinetic models: the activated sludge model (ASM) 1 and McKeithan's T-cell receptor model.
Keywords
Cite
@article{arxiv.2203.16135,
title = {Kron-based Model-order Reduction of Open Mass-action Kinetics Chemical Reaction Networks},
author = {Mohamad Agung Prawira Negara and Azka Muji Burohman and Bayu Jayawardhana},
journal= {arXiv preprint arXiv:2203.16135},
year = {2022}
}
Comments
Submitted to IEEE Transactions on Automatic Control