zbMATH — the first resource for mathematics

A relative error bound for balanced stochastic truncation. (English) Zbl 0659.93061
The stochastic model reduction technique known as balanced stochastic truncation is shown to satisfy an infinity norm bound on the relative error between the frequency responses of the original and reduced order models. The error bound is in terms of the canonical correlation coefficients associated with the stochastic realization.
The balanced stochastic truncation algorithm can therefore be used to provide a computationally simple method, involving the solution of an algebraic Riccati equation, for relative error model reduction.
Reviewer: M.Green

93E03 Stochastic systems in control theory (general)
93B15 Realizations from input-output data
93E25 Computational methods in stochastic control (MSC2010)
15A24 Matrix equations and identities
Full Text: DOI