Purpose
To obtain the state-space model (A,B,C,D) for the cascaded inter-connection of two systems, each given in state-space form.Specification
SUBROUTINE AB05MD( UPLO, OVER, N1, M1, P1, N2, P2, A1, LDA1, B1,
$ LDB1, C1, LDC1, D1, LDD1, A2, LDA2, B2, LDB2,
$ C2, LDC2, D2, LDD2, N, A, LDA, B, LDB, C, LDC,
$ D, LDD, DWORK, LDWORK, INFO )
C .. Scalar Arguments ..
CHARACTER OVER, UPLO
INTEGER INFO, LDA, LDA1, LDA2, LDB, LDB1, LDB2, LDC,
$ LDC1, LDC2, LDD, LDD1, LDD2, LDWORK, M1, N, N1,
$ N2, P1, P2
C .. Array Arguments ..
DOUBLE PRECISION A(LDA,*), A1(LDA1,*), A2(LDA2,*), B(LDB,*),
$ B1(LDB1,*), B2(LDB2,*), C(LDC,*), C1(LDC1,*),
$ C2(LDC2,*), D(LDD,*), D1(LDD1,*), D2(LDD2,*),
$ DWORK(*)
Arguments
Mode Parameters
UPLO CHARACTER*1
Indicates whether the user wishes to obtain the matrix A
in the upper or lower block diagonal form, as follows:
= 'U': Obtain A in the upper block diagonal form;
= 'L': Obtain A in the lower block diagonal form.
OVER CHARACTER*1
Indicates whether the user wishes to overlap pairs of
arrays, as follows:
= 'N': Do not overlap;
= 'O': Overlap pairs of arrays: A1 and A, B1 and B,
C1 and C, and D1 and D (for UPLO = 'L'), or A2
and A, B2 and B, C2 and C, and D2 and D (for
UPLO = 'U'), i.e. the same name is effectively
used for each pair (for all pairs) in the routine
call. In this case, setting LDA1 = LDA,
LDB1 = LDB, LDC1 = LDC, and LDD1 = LDD, or
LDA2 = LDA, LDB2 = LDB, LDC2 = LDC, and LDD2 = LDD
will give maximum efficiency.
Input/Output Parameters
N1 (input) INTEGER
The number of state variables in the first system, i.e.
the order of the matrix A1. N1 >= 0.
M1 (input) INTEGER
The number of input variables for the first system.
M1 >= 0.
P1 (input) INTEGER
The number of output variables from the first system and
the number of input variables for the second system.
P1 >= 0.
N2 (input) INTEGER
The number of state variables in the second system, i.e.
the order of the matrix A2. N2 >= 0.
P2 (input) INTEGER
The number of output variables from the second system.
P2 >= 0.
A1 (input) DOUBLE PRECISION array, dimension (LDA1,N1)
The leading N1-by-N1 part of this array must contain the
state transition matrix A1 for the first system.
LDA1 INTEGER
The leading dimension of array A1. LDA1 >= MAX(1,N1).
B1 (input) DOUBLE PRECISION array, dimension (LDB1,M1)
The leading N1-by-M1 part of this array must contain the
input/state matrix B1 for the first system.
LDB1 INTEGER
The leading dimension of array B1. LDB1 >= MAX(1,N1).
C1 (input) DOUBLE PRECISION array, dimension (LDC1,N1)
The leading P1-by-N1 part of this array must contain the
state/output matrix C1 for the first system.
LDC1 INTEGER
The leading dimension of array C1.
LDC1 >= MAX(1,P1) if N1 > 0.
LDC1 >= 1 if N1 = 0.
D1 (input) DOUBLE PRECISION array, dimension (LDD1,M1)
The leading P1-by-M1 part of this array must contain the
input/output matrix D1 for the first system.
LDD1 INTEGER
The leading dimension of array D1. LDD1 >= MAX(1,P1).
A2 (input) DOUBLE PRECISION array, dimension (LDA2,N2)
The leading N2-by-N2 part of this array must contain the
state transition matrix A2 for the second system.
LDA2 INTEGER
The leading dimension of array A2. LDA2 >= MAX(1,N2).
B2 (input) DOUBLE PRECISION array, dimension (LDB2,P1)
The leading N2-by-P1 part of this array must contain the
input/state matrix B2 for the second system.
LDB2 INTEGER
The leading dimension of array B2. LDB2 >= MAX(1,N2).
C2 (input) DOUBLE PRECISION array, dimension (LDC2,N2)
The leading P2-by-N2 part of this array must contain the
state/output matrix C2 for the second system.
LDC2 INTEGER
The leading dimension of array C2.
LDC2 >= MAX(1,P2) if N2 > 0.
LDC2 >= 1 if N2 = 0.
D2 (input) DOUBLE PRECISION array, dimension (LDD2,P1)
The leading P2-by-P1 part of this array must contain the
input/output matrix D2 for the second system.
LDD2 INTEGER
The leading dimension of array D2. LDD2 >= MAX(1,P2).
N (output) INTEGER
The number of state variables (N1 + N2) in the resulting
system, i.e. the order of the matrix A, the number of rows
of B and the number of columns of C.
A (output) DOUBLE PRECISION array, dimension (LDA,N1+N2)
The leading N-by-N part of this array contains the state
transition matrix A for the cascaded system.
If OVER = 'O', the array A can overlap A1, if UPLO = 'L',
or A2, if UPLO = 'U'.
LDA INTEGER
The leading dimension of array A. LDA >= MAX(1,N1+N2).
B (output) DOUBLE PRECISION array, dimension (LDB,M1)
The leading N-by-M1 part of this array contains the
input/state matrix B for the cascaded system.
If OVER = 'O', the array B can overlap B1, if UPLO = 'L',
or B2, if UPLO = 'U'.
LDB INTEGER
The leading dimension of array B. LDB >= MAX(1,N1+N2).
C (output) DOUBLE PRECISION array, dimension (LDC,N1+N2)
The leading P2-by-N part of this array contains the
state/output matrix C for the cascaded system.
If OVER = 'O', the array C can overlap C1, if UPLO = 'L',
or C2, if UPLO = 'U'.
LDC INTEGER
The leading dimension of array C.
LDC >= MAX(1,P2) if N1+N2 > 0.
LDC >= 1 if N1+N2 = 0.
D (output) DOUBLE PRECISION array, dimension (LDD,M1)
The leading P2-by-M1 part of this array contains the
input/output matrix D for the cascaded system.
If OVER = 'O', the array D can overlap D1, if UPLO = 'L',
or D2, if UPLO = 'U'.
LDD INTEGER
The leading dimension of array D. LDD >= MAX(1,P2).
Workspace
DWORK DOUBLE PRECISION array, dimension (LDWORK)
The array DWORK is not referenced if OVER = 'N'.
LDWORK INTEGER
The length of the array DWORK.
LDWORK >= MAX( 1, P1*MAX(N1, M1, N2, P2) ) if OVER = 'O'.
LDWORK >= 1 if OVER = 'N'.
Error Indicator
INFO INTEGER
= 0: successful exit;
< 0: if INFO = -i, the i-th argument had an illegal
value.
Method
After cascaded inter-connection of the two systems
X1' = A1*X1 + B1*U
V = C1*X1 + D1*U
X2' = A2*X2 + B2*V
Y = C2*X2 + D2*V
(where ' denotes differentiation with respect to time)
the following state-space model will be obtained:
X' = A*X + B*U
Y = C*X + D*U
where matrix A has the form ( A1 0 ),
( B2*C1 A2)
matrix B has the form ( B1 ),
( B2*D1 )
matrix C has the form ( D2*C1 C2 ) and
matrix D has the form ( D2*D1 ).
This form is returned by the routine when UPLO = 'L'. Note that
when A1 and A2 are block lower triangular, the resulting state
matrix is also block lower triangular.
By applying a similarity transformation to the system above,
using the matrix ( 0 I ), where I is the identity matrix of
( J 0 )
order N2, and J is the identity matrix of order N1, the
system matrices become
A = ( A2 B2*C1 ),
( 0 A1 )
B = ( B2*D1 ),
( B1 )
C = ( C2 D2*C1 ) and
D = ( D2*D1 ).
This form is returned by the routine when UPLO = 'U'. Note that
when A1 and A2 are block upper triangular (for instance, in the
real Schur form), the resulting state matrix is also block upper
triangular.
References
NoneNumerical Aspects
The algorithm requires P1*(N1+M1)*(N2+P2) operations.Further Comments
NoneExample
Program Text
* AB05MD EXAMPLE PROGRAM TEXT
*
* .. Parameters ..
INTEGER NIN, NOUT
PARAMETER ( NIN = 5, NOUT = 6 )
INTEGER N1MAX, N2MAX, NMAX, M1MAX, P1MAX, P2MAX
PARAMETER ( N1MAX = 20, N2MAX = 20, NMAX = N1MAX+N2MAX,
$ M1MAX = 20, P1MAX = 20, P2MAX = 20 )
INTEGER LDA, LDA1, LDA2, LDB, LDB1, LDB2, LDC, LDC1,
$ LDC2, LDD, LDD1, LDD2, LDWORK
PARAMETER ( LDA = NMAX, LDA1 = N1MAX, LDA2 = N2MAX,
$ LDB = NMAX,LDB1 = N1MAX, LDB2 = N2MAX,
$ LDC = P2MAX, LDC1 = P1MAX, LDC2 = P2MAX,
$ LDD = P2MAX, LDD1 = P1MAX, LDD2 = P2MAX,
$ LDWORK = P1MAX*N1MAX )
* .. Local Scalars ..
CHARACTER*1 OVER, UPLO
INTEGER I, INFO, J, M1, N, N1, N2, P1, P2
* .. Local Arrays ..
DOUBLE PRECISION A(LDA,NMAX), A1(LDA1,N1MAX), A2(LDA2,N2MAX),
$ B(LDB,M1MAX), B1(LDB1,M1MAX), B2(LDB2,P1MAX),
$ C(LDC,NMAX), C1(LDC1,N1MAX), C2(LDC2,N2MAX),
$ D(LDD,M1MAX), D1(LDD1,M1MAX), D2(LDD2,P1MAX),
$ DWORK(LDWORK)
* .. External Subroutines ..
EXTERNAL AB05MD
* .. Executable Statements ..
*
UPLO = 'Lower'
OVER = 'N'
WRITE ( NOUT, FMT = 99999 )
* Skip the heading in the data file and read the data.
READ ( NIN, FMT = '()' )
READ ( NIN, FMT = * ) N1, M1, P1, N2, P2
IF ( N1.LE.0 .OR. N1.GT.N1MAX ) THEN
WRITE ( NOUT, FMT = 99992 ) N1
ELSE
READ ( NIN, FMT = * ) ( ( A1(I,J), J = 1,N1 ), I = 1,N1 )
IF ( M1.LE.0 .OR. M1.GT.M1MAX ) THEN
WRITE ( NOUT, FMT = 99991 ) M1
ELSE
READ ( NIN, FMT = * ) ( ( B1(I,J), I = 1,N1 ), J = 1,M1 )
IF ( P1.LE.0 .OR. P1.GT.P1MAX ) THEN
WRITE ( NOUT, FMT = 99990 ) P1
ELSE
READ ( NIN, FMT = * ) ( ( C1(I,J), J = 1,N1 ), I = 1,P1 )
READ ( NIN, FMT = * ) ( ( D1(I,J), J = 1,M1 ), I = 1,P1 )
IF ( N2.LE.0 .OR. N2.GT.N2MAX ) THEN
WRITE ( NOUT, FMT = 99989 ) N2
ELSE
READ ( NIN, FMT = * )
$ ( ( A2(I,J), J = 1,N2 ), I = 1,N2 )
READ ( NIN, FMT = * )
$ ( ( B2(I,J), I = 1,N2 ), J = 1,P1 )
IF ( P2.LE.0 .OR. P2.GT.P2MAX ) THEN
WRITE ( NOUT, FMT = 99988 ) P2
ELSE
READ ( NIN, FMT = * )
$ ( ( C2(I,J), J = 1,N2 ), I = 1,P2 )
READ ( NIN, FMT = * )
$ ( ( D2(I,J), J = 1,P1 ), I = 1,P2 )
* Find the state-space model (A,B,C,D).
CALL AB05MD( UPLO, OVER, N1, M1, P1, N2, P2, A1,
$ LDA1, B1, LDB1, C1, LDC1, D1, LDD1,
$ A2, LDA2, B2, LDB2, C2, LDC2, D2,
$ LDD2, N, A, LDA, B, LDB, C, LDC, D,
$ LDD, DWORK, LDWORK, INFO )
*
IF ( INFO.NE.0 ) THEN
WRITE ( NOUT, FMT = 99998 ) INFO
ELSE
WRITE ( NOUT, FMT = 99997 )
DO 20 I = 1, N
WRITE ( NOUT, FMT = 99996 )
$ ( A(I,J), J = 1,N )
20 CONTINUE
WRITE ( NOUT, FMT = 99995 )
DO 40 I = 1, N
WRITE ( NOUT, FMT = 99996 )
$ ( B(I,J), J = 1,M1 )
40 CONTINUE
WRITE ( NOUT, FMT = 99994 )
DO 60 I = 1, P2
WRITE ( NOUT, FMT = 99996 )
$ ( C(I,J), J = 1,N )
60 CONTINUE
WRITE ( NOUT, FMT = 99993 )
DO 80 I = 1, P2
WRITE ( NOUT, FMT = 99996 )
$ ( D(I,J), J = 1,M1 )
80 CONTINUE
END IF
END IF
END IF
END IF
END IF
END IF
STOP
*
99999 FORMAT (' AB05MD EXAMPLE PROGRAM RESULTS',/1X)
99998 FORMAT (' INFO on exit from AB05MD = ',I2)
99997 FORMAT (' The state transition matrix of the cascaded system is ')
99996 FORMAT (20(1X,F8.4))
99995 FORMAT (/' The input/state matrix of the cascaded system is ')
99994 FORMAT (/' The state/output matrix of the cascaded system is ')
99993 FORMAT (/' The input/output matrix of the cascaded system is ')
99992 FORMAT (/' N1 is out of range.',/' N1 = ',I5)
99991 FORMAT (/' M1 is out of range.',/' M1 = ',I5)
99990 FORMAT (/' P1 is out of range.',/' P1 = ',I5)
99989 FORMAT (/' N2 is out of range.',/' N2 = ',I5)
99988 FORMAT (/' P2 is out of range.',/' P2 = ',I5)
END
Program Data
AB05MD EXAMPLE PROGRAM DATA 3 2 2 3 2 1.0 0.0 -1.0 0.0 -1.0 1.0 1.0 1.0 2.0 1.0 1.0 0.0 2.0 0.0 1.0 3.0 -2.0 1.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0 -3.0 0.0 0.0 1.0 0.0 1.0 0.0 -1.0 2.0 0.0 -1.0 0.0 1.0 0.0 2.0 1.0 1.0 0.0 1.0 1.0 -1.0 1.0 1.0 0.0 1.0Program Results
AB05MD EXAMPLE PROGRAM RESULTS The state transition matrix of the cascaded system is 1.0000 0.0000 -1.0000 0.0000 0.0000 0.0000 0.0000 -1.0000 1.0000 0.0000 0.0000 0.0000 1.0000 1.0000 2.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 -3.0000 0.0000 0.0000 -3.0000 2.0000 -1.0000 1.0000 0.0000 1.0000 0.0000 2.0000 0.0000 0.0000 -1.0000 2.0000 The input/state matrix of the cascaded system is 1.0000 2.0000 1.0000 0.0000 0.0000 1.0000 0.0000 1.0000 -1.0000 0.0000 0.0000 2.0000 The state/output matrix of the cascaded system is 3.0000 -1.0000 1.0000 1.0000 1.0000 0.0000 0.0000 1.0000 0.0000 1.0000 1.0000 -1.0000 The input/output matrix of the cascaded system is 1.0000 1.0000 0.0000 1.0000