Balanced truncation (linear)

powerful technique to reduce the state-space dimension of a dynamical system with linear input equation

computational task

MaRDI QIDQ6684551FDOCommunityMathModDB

Description

In the case of a linear control system, a useful property of balanced truncation is that it admits easy control of the approximation error when truncating states.

Contained entities

control system matrix B (reduced)	$\tilde{B}\equiv T^{-1}BT$	output
	$B$ represents control system matrix B
	$T$ represents MOR transformation matrix
	$\tilde{B}$ represents control system matrix B (reduced)
Lyapunov equation observability	$A^{\ast }{W}_o+W_{o}A+C^{\ast }C=0$
	$A$ represents control system matrix A
	$C$ represents control system matrix C
	$W_o$ represents observability Gramian
control system output linear	$y(t)=Cx(t)$
	$C$ represents control system matrix C
	$t$ represents time
	$x$ represents control system state
	$y$ represents control system output
initial condition for control state	$x(t=0)=x_0$	initial condition
	$t$ represents time
	$x$ represents control system state
	$x_0$ represents initial value of control system
control system output linear (reduced)	$y(t)=\tilde{C}\tilde{x}(t)$
	$\tilde{C}$ represents control system matrix C (reduced)
	$\tilde{x}$ represents control system state (reduced)
	$t$ represents time
	$y$ represents control system output
balancing transformation	$T^{-1}{W}_c{\left(T^{-1}\right)}^{\ast }=T^{\ast }{W}_{o}T=\left(\begin{array}{ccc}{\sigma }_1& & 0\\ & \ddots & \\ 0& & {\sigma }_n\end{array}\right)=\Sigma$
	$T$ represents MOR transformation matrix
	$W_c$ represents controllability Gramian
	$W_o$ represents observability Gramian
	$\sigma$ represents Hankel singular value
control system input linear (reduced)	$\dot{\tilde{x}}(t)=\tilde{A}\tilde{x}(t)+\tilde{B}u(t)$
	$\tilde{A}$ represents control system matrix A (reduced)
	$\tilde{B}$ represents control system matrix B (reduced)
	$\tilde{x}$ represents control system state (reduced)
	$t$ represents time
	$u$ represents control system input
Lyapunov equation controllability	$A{W}_c+W_c{A}^{\ast }+B{B}^{\ast }=0$
	$A$ represents control system matrix A
	$B$ represents control system matrix B
	$W_c$ represents controllability Gramian
control system matrix B		input
control system matrix C		input
control system matrix A (reduced)	$\tilde{A}\equiv T^{-1}AT$	output
	$A$ represents control system matrix A
	$T$ represents MOR transformation matrix
	$\tilde{A}$ represents control system matrix A (reduced)
initial control state (reduced)	$\tilde{x}(t=0)\equiv {\tilde{x}}_0$	input
	$\tilde{x}$ represents control system state (reduced)
	$t$ represents time
	${\tilde{x}}_0$ represents initial control state (reduced)
control system matrix C (reduced)	$\tilde{C}\equiv CT$	output
	$C$ represents control system matrix C
	$T$ represents MOR transformation matrix
	$\tilde{C}$ represents control system matrix C (reduced)
control system matrix A		input
MOR transformation matrix		output
control system input linear	$\dot{x}(t)=Ax(t)+Bu(t)$
	$A$ represents control system matrix A
	$B$ represents control system matrix B
	$t$ represents time
	$u$ represents control system input
	$x$ represents control system state

Further links

Item	Property
balanced truncation	specialized by
balanced truncation (bi-linear)	specialized by
linear time-invariant system	used by
model order reduction	specialized by

---

---

Edit balanced truncation (linear) item

This page was built for task: balanced truncation (linear)