 
 
 
 
 
 
 
 
 
 
Although the KCF looks quite complicated in the general case, most
matrix pencils have a more simple Kronecker structure.
If  is
 is  , where
, where  , then for almost all
, then for almost all 
 and
 and  it will have the same KCF, depending only on
 it will have the same KCF, depending only on  and
 and  .
This corresponds to the generic case when
.
This corresponds to the generic case when  has full rank
for any complex (or real) value of
 has full rank
for any complex (or real) value of  . 
Accordingly, generic rectangular pencils have no regular part.
The generic Kronecker structure for
. 
Accordingly, generic rectangular pencils have no regular part.
The generic Kronecker structure for  with
 with  
 is
  
diag(L_, ..., L_, L_+ 1, 
..., L_+ 1), 
where
 is
  
diag(L_, ..., L_, L_+ 1, 
..., L_+ 1), 
where 
 , the total number of blocks is
, the total number of blocks is  , and
the number of
, and
the number of  blocks is
 blocks is  (which is 0 when
 (which is 0 when  divides
divides  ) [446,116].
The same statement holds for
) [446,116].
The same statement holds for 
 if we replace
 if we replace 
 in (8.31) by
 in (8.31) by 
 .
For example, a generic pencil of size
.
For example, a generic pencil of size  has
an
 has
an  block as its KCF.
 block as its KCF.
Square pencils are generically regular; i.e., 
 if and only if
 
if and only if  is an eigenvalue. 
Moreover, the most generic regular pencil is diagonalizable
with distinct finite eigenvalues.
The generic singular pencils of size
 is an eigenvalue. 
Moreover, the most generic regular pencil is diagonalizable
with distinct finite eigenvalues.
The generic singular pencils of size  have the
Kronecker structures [456]:
  
diag(L_j, L_n-j-1^T),     j = 0, ..., n-1. 
Only if a singular
 have the
Kronecker structures [456]:
  
diag(L_j, L_n-j-1^T),     j = 0, ..., n-1. 
Only if a singular  is rank deficient (for some
 is rank deficient (for some  ),  
the associated KCF may be more complicated and possibly include a regular part, 
as well as right and left singular blocks. 
This situation corresponds to the nongeneric (or degenerate) 
case, which is the real challenge from a computational point of view.
Degenerate rectangular pencils have several applications in
control theory, for example, to compute the 
controllable subspace and uncontrollable modes
of a linear descriptor system [447,120].
),  
the associated KCF may be more complicated and possibly include a regular part, 
as well as right and left singular blocks. 
This situation corresponds to the nongeneric (or degenerate) 
case, which is the real challenge from a computational point of view.
Degenerate rectangular pencils have several applications in
control theory, for example, to compute the 
controllable subspace and uncontrollable modes
of a linear descriptor system [447,120].
 
 
 
 
 
 
 
 
