1SHSEQR(1)             LAPACK driver routine (version 3.2)            SHSEQR(1)
2
3
4

NAME

6       SHSEQR  - SHSEQR compute the eigenvalues of a Hessenberg matrix H  and,
7       optionally, the matrices T and Z from the Schur decomposition  H = Z  T
8       Z**T,  where  T  is an upper quasi-triangular matrix (the  Schur form),
9       and Z is the orthogonal matrix of Schur vectors
10

SYNOPSIS

12       SUBROUTINE SHSEQR( JOB, COMPZ, N, ILO, IHI, H, LDH,  WR,  WI,  Z,  LDZ,
13                          WORK, LWORK, INFO )
14
15           INTEGER        IHI, ILO, INFO, LDH, LDZ, LWORK, N
16
17           CHARACTER      COMPZ, JOB
18
19           REAL           H(  LDH, * ), WI( * ), WORK( * ), WR( * ), Z( LDZ, *
20                          )
21

PURPOSE

23          SHSEQR computes the eigenvalues of a Hessenberg matrix H
24          and, optionally, the matrices T and Z from the Schur decomposition
25          H = Z T Z**T, where T is an upper quasi-triangular matrix (the
26          Schur form), and Z is the orthogonal matrix of Schur vectors.
27          Optionally Z may be postmultiplied into an input orthogonal
28          matrix Q so that this routine can give the Schur factorization
29          of a matrix A which has been reduced to the Hessenberg form H
30          by the orthogonal matrix Q:  A = Q*H*Q**T = (QZ)*T*(QZ)**T.
31

ARGUMENTS

33       JOB   (input) CHARACTER*1
34             = 'E':  compute eigenvalues only;
35             = 'S':  compute eigenvalues and the Schur form T.  COMPZ  (input)
36             CHARACTER*1
37             = 'N':  no Schur vectors are computed;
38             =  'I':   Z is initialized to the unit matrix and the matrix Z of
39             Schur vectors of H is returned; = 'V':  Z must contain an orthog‐
40             onal matrix Q on entry, and the product Q*Z is returned.
41
42       N     (input) INTEGER
43             The order of the matrix H.  N .GE. 0.
44
45       ILO   (input) INTEGER
46             IHI    (input) INTEGER It is assumed that H is already upper tri‐
47             angular in rows and columns 1:ILO-1 and IHI+1:N. ILO and IHI  are
48             normally  set  by  a  previous call to SGEBAL, and then passed to
49             SGEHRD when the matrix output by SGEBAL is reduced to  Hessenberg
50             form.  Otherwise  ILO  and  IHI  should be set to 1 and N respec‐
51             tively.  If N.GT.0, then 1.LE.ILO.LE.IHI.LE.N.  If N  =  0,  then
52             ILO = 1 and IHI = 0.
53
54       H     (input/output) REAL array, dimension (LDH,N)
55             On  entry,  the  upper Hessenberg matrix H.  On exit, if INFO = 0
56             and JOB = 'S', then H contains the upper quasi-triangular  matrix
57             T  from the Schur decomposition (the Schur form); 2-by-2 diagonal
58             blocks (corresponding to complex conjugate pairs of  eigenvalues)
59             are  returned  in  standard  form,  with  H(i,i) = H(i+1,i+1) and
60             H(i+1,i)*H(i,i+1).LT.0. If INFO = 0 and JOB = 'E',  the  contents
61             of  H  are  unspecified  on  exit.   (The  output value of H when
62             INFO.GT.0 is given under the description of INFO below.)   Unlike
63             earlier versions of SHSEQR, this subroutine may explicitly H(i,j)
64             = 0 for i.GT.j and j = 1, 2, ... ILO-1 or j = IHI+1,  IHI+2,  ...
65             N.
66
67       LDH   (input) INTEGER
68             The leading dimension of the array H. LDH .GE. max(1,N).
69
70       WR    (output) REAL array, dimension (N)
71             WI     (output)  REAL array, dimension (N) The real and imaginary
72             parts, respectively, of the computed eigenvalues. If  two  eigen‐
73             values  are computed as a complex conjugate pair, they are stored
74             in consecutive elements of WR and WI, say the i-th  and  (i+1)th,
75             with WI(i) .GT. 0 and WI(i+1) .LT. 0. If JOB = 'S', the eigenval‐
76             ues are stored in the same order as on the diagonal of the  Schur
77             form returned in H, with WR(i) = H(i,i) and, if H(i:i+1,i:i+1) is
78             a 2-by-2 diagonal block,  WI(i)  =  sqrt(-H(i+1,i)*H(i,i+1))  and
79             WI(i+1) = -WI(i).
80
81       Z     (input/output) REAL array, dimension (LDZ,N)
82             If  COMPZ = 'N', Z is not referenced.  If COMPZ = 'I', on entry Z
83             need not be set and on exit, if INFO = 0, Z contains the orthogo‐
84             nal matrix Z of the Schur vectors of H.  If COMPZ = 'V', on entry
85             Z must contain an N-by-N matrix Q, which is assumed to  be  equal
86             to  the  unit matrix except for the submatrix Z(ILO:IHI,ILO:IHI).
87             On exit, if INFO = 0, Z contains Q*Z.  Normally Q is the orthogo‐
88             nal  matrix  generated  by  SORGHR after the call to SGEHRD which
89             formed the Hessenberg matrix H.  (The  output  value  of  Z  when
90             INFO.GT.0 is given under the description of INFO below.)
91
92       LDZ   (input) INTEGER
93             The  leading dimension of the array Z.  if COMPZ = 'I' or COMPZ =
94             'V', then LDZ.GE.MAX(1,N).  Otherwize, LDZ.GE.1.
95
96       WORK  (workspace/output) REAL array, dimension (LWORK)
97             On exit, if INFO = 0, WORK(1) returns an estimate of the  optimal
98             value  for  LWORK.   LWORK  (input)  INTEGER The dimension of the
99             array WORK.  LWORK .GE. max(1,N) is sufficient and delivers  very
100             good  and sometimes optimal performance.  However, LWORK as large
101             as 11*N may be required for  optimal  performance.   A  workspace
102             query is recommended to determine the optimal workspace size.  If
103             LWORK = -1, then SHSEQR does a workspace query.   In  this  case,
104             SHSEQR  checks  the  input  parameters  and estimates the optimal
105             workspace size for the given values of N, ILO and IHI.  The esti‐
106             mate  is  returned in WORK(1).  No error message related to LWORK
107             is issued by XERBLA.  Neither H nor Z are accessed.
108
109       INFO  (output) INTEGER
110             =  0:  successful exit
111             value
112             the eigenvalues.  Elements 1:ilo-1 and i+1:n of WR and WI contain
113             those  eigenvalues which have been successfully computed.  (Fail‐
114             ures are rare.)  If INFO .GT. 0 and JOB = 'E', then on exit,  the
115             remaining  unconverged  eigenvalues  are the eigen- values of the
116             upper Hessenberg matrix rows and columns ILO through INFO of  the
117             final,  output  value of H.  If INFO .GT. 0 and JOB   = 'S', then
118             on exit
119
120       (*)  (initial value of H)*U  = U*(final value of H)
121            where U is an orthogonal matrix.  The final value of  H  is  upper
122            Hessenberg and quasi-triangular in rows and columns INFO+1 through
123            IHI.  If INFO .GT. 0 and COMPZ = 'V', then on exit (final value of
124            Z)   =  (initial value of Z)*U where U is the orthogonal matrix in
125            (*) (regard- less of the value of JOB.)  If INFO .GT. 0 and  COMPZ
126            = 'I', then on exit (final value of Z)  = U where U is the orthog‐
127            onal matrix in (*) (regard- less of the value of  JOB.)   If  INFO
128            .GT. 0 and COMPZ = 'N', then Z is not accessed.
129
130
131
132 LAPACK driver routine (version 3.N2o)vember 2008                       SHSEQR(1)
Impressum