The model matching problem corresponds to the
special case of the standard problem of
control
where the plant G has no feedback term
(that is, D=0) and when K lies in 2.
The standard problem of
control
when
2 was shown in Lemmas3.2
and
3.3
to reduce to a model matching problem.
Here we shall always take
.
The traditional
control model matching problem leads to T =
A + BQC. Thus the interpolating condition becomes
where
.
This typically is a one
complex dimensional curve in
or a
finite set of such curves in
.
Thus we do not have a finite set of
interpolating constraints.
Agler's Theorem says that for a solution to the
control problem to exist Agler's Inequality applied to every
finite set of points satisfying this interpolation constraint
must have a solution. That is, Agler's Inequality is a
necessary condition for
control. How close this is to sufficient is not known.
However, it is quite likely that as one takes larger and
larger finite sets F of interpolating conditions, then
the performance
bounds
obtained from Agler's inequality
approach being sufficient .
Several cases of model matching which lead to a finite number of interpolation constraints are when there are two controls and one measurement, and when there are two controls and two measurements (in general position).