1DLAED4(1)                LAPACK routine (version 3.1)                DLAED4(1)
2
3
4

NAME

6       DLAED4  -  compute  the I-th updated eigenvalue of a symmetric rank-one
7       modification to a diagonal matrix whose elements are given in the array
8       d, and that   D(i) < D(j) for i < j  and that RHO > 0
9

SYNOPSIS

11       SUBROUTINE DLAED4( N, I, D, Z, DELTA, RHO, DLAM, INFO )
12
13           INTEGER        I, INFO, N
14
15           DOUBLE         PRECISION DLAM, RHO
16
17           DOUBLE         PRECISION D( * ), DELTA( * ), Z( * )
18

PURPOSE

20       This  subroutine  computes  the  I-th updated eigenvalue of a symmetric
21       rank-one modification to a diagonal matrix whose elements are given  in
22       the  array  d,  and  that no loss in generality.  The rank-one modified
23       system is thus
24
25                  diag( D )  +  RHO *  Z * Z_transpose.
26
27       where we assume the Euclidean norm of Z is 1.
28
29       The method consists of approximating the rational functions in the sec‐
30       ular equation by simpler interpolating rational functions.
31
32

ARGUMENTS

34       N      (input) INTEGER
35              The length of all arrays.
36
37       I      (input) INTEGER
38              The index of the eigenvalue to be computed.  1 <= I <= N.
39
40       D      (input) DOUBLE PRECISION array, dimension (N)
41              The original eigenvalues.  It is assumed that they are in order,
42              D(I) < D(J)  for I < J.
43
44       Z      (input) DOUBLE PRECISION array, dimension (N)
45              The components of the updating vector.
46
47       DELTA  (output) DOUBLE PRECISION array, dimension (N)
48              If N .GT. 2, DELTA contains (D(j) - lambda_I) in its  j-th  com‐
49              ponent.   If  N = 1, then DELTA(1) = 1. If N = 2, see DLAED5 for
50              detail. The vector DELTA contains the information  necessary  to
51              construct the eigenvectors by DLAED3 and DLAED9.
52
53       RHO    (input) DOUBLE PRECISION
54              The scalar in the symmetric updating formula.
55
56       DLAM   (output) DOUBLE PRECISION
57              The computed lambda_I, the I-th updated eigenvalue.
58
59       INFO   (output) INTEGER
60              = 0:  successful exit
61              > 0:  if INFO = 1, the updating process failed.
62

PARAMETERS

64       Logical  variable  ORGATI  (origin-at-i?)  is  used  for distinguishing
65       whether D(i) or D(i+1) is treated as the origin.
66
67       ORGATI = .true.    origin at i ORGATI = .false.   origin at i+1
68
69       Logical variable SWTCH3 (switch-for-3-poles?) is for noting if  we  are
70       working with THREE poles!
71
72       MAXIT is the maximum number of iterations allowed for each eigenvalue.
73
74       Further Details ===============
75
76       Based  on contributions by Ren-Cang Li, Computer Science Division, Uni‐
77       versity of California at Berkeley, USA
78
79
80
81 LAPACK routine (version 3.1)    November 2006                       DLAED4(1)