1ZPBEQU(1)                LAPACK routine (version 3.1)                ZPBEQU(1)
2
3
4

NAME

6       ZPBEQU  -  row  and column scalings intended to equilibrate a Hermitian
7       positive definite band matrix A and reduce its condition  number  (with
8       respect to the two-norm)
9

SYNOPSIS

11       SUBROUTINE ZPBEQU( UPLO, N, KD, AB, LDAB, S, SCOND, AMAX, INFO )
12
13           CHARACTER      UPLO
14
15           INTEGER        INFO, KD, LDAB, N
16
17           DOUBLE         PRECISION AMAX, SCOND
18
19           DOUBLE         PRECISION S( * )
20
21           COMPLEX*16     AB( LDAB, * )
22

PURPOSE

24       ZPBEQU  computes row and column scalings intended to equilibrate a Her‐
25       mitian positive definite band matrix A and reduce its condition  number
26       (with  respect  to the two-norm).  S contains the scale factors, S(i) =
27       1/sqrt(A(i,i)), chosen so that the scaled matrix B with elements B(i,j)
28       = S(i)*A(i,j)*S(j) has ones on the diagonal.  This choice of S puts the
29       condition number of B within a factor N of the smallest possible condi‐
30       tion number over all possible diagonal scalings.
31
32

ARGUMENTS

34       UPLO    (input) CHARACTER*1
35               = 'U':  Upper triangular of A is stored;
36               = 'L':  Lower triangular of A is stored.
37
38       N       (input) INTEGER
39               The order of the matrix A.  N >= 0.
40
41       KD      (input) INTEGER
42               The  number of superdiagonals of the matrix A if UPLO = 'U', or
43               the number of subdiagonals if UPLO = 'L'.  KD >= 0.
44
45       AB      (input) COMPLEX*16 array, dimension (LDAB,N)
46               The upper or lower triangle of the  Hermitian  band  matrix  A,
47               stored in the first KD+1 rows of the array.  The j-th column of
48               A is stored in the j-th column of the array AB as  follows:  if
49               UPLO  =  'U', AB(kd+1+i-j,j) = A(i,j) for max(1,j-kd)<=i<=j; if
50               UPLO = 'L', AB(1+i-j,j)    = A(i,j) for j<=i<=min(n,j+kd).
51
52       LDAB     (input) INTEGER
53                The leading dimension of the array A.  LDAB >= KD+1.
54
55       S       (output) DOUBLE PRECISION array, dimension (N)
56               If INFO = 0, S contains the scale factors for A.
57
58       SCOND   (output) DOUBLE PRECISION
59               If INFO = 0, S contains the ratio of the smallest S(i)  to  the
60               largest  S(i).   If  SCOND >= 0.1 and AMAX is neither too large
61               nor too small, it is not worth scaling by S.
62
63       AMAX    (output) DOUBLE PRECISION
64               Absolute value of largest matrix  element.   If  AMAX  is  very
65               close to overflow or very close to underflow, the matrix should
66               be scaled.
67
68       INFO    (output) INTEGER
69               = 0:  successful exit
70               < 0:  if INFO = -i, the i-th argument had an illegal value.
71               > 0:  if INFO = i, the i-th diagonal element is nonpositive.
72
73
74
75 LAPACK routine (version 3.1)    November 2006                       ZPBEQU(1)
Impressum