You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

claesy.f 6.3 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218
  1. *> \brief \b CLAESY computes the eigenvalues and eigenvectors of a 2-by-2 complex symmetric matrix.
  2. *
  3. * =========== DOCUMENTATION ===========
  4. *
  5. * Online html documentation available at
  6. * http://www.netlib.org/lapack/explore-html/
  7. *
  8. *> \htmlonly
  9. *> Download CLAESY + dependencies
  10. *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/claesy.f">
  11. *> [TGZ]</a>
  12. *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/claesy.f">
  13. *> [ZIP]</a>
  14. *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/claesy.f">
  15. *> [TXT]</a>
  16. *> \endhtmlonly
  17. *
  18. * Definition:
  19. * ===========
  20. *
  21. * SUBROUTINE CLAESY( A, B, C, RT1, RT2, EVSCAL, CS1, SN1 )
  22. *
  23. * .. Scalar Arguments ..
  24. * COMPLEX A, B, C, CS1, EVSCAL, RT1, RT2, SN1
  25. * ..
  26. *
  27. *
  28. *> \par Purpose:
  29. * =============
  30. *>
  31. *> \verbatim
  32. *>
  33. *> CLAESY computes the eigendecomposition of a 2-by-2 symmetric matrix
  34. *> ( ( A, B );( B, C ) )
  35. *> provided the norm of the matrix of eigenvectors is larger than
  36. *> some threshold value.
  37. *>
  38. *> RT1 is the eigenvalue of larger absolute value, and RT2 of
  39. *> smaller absolute value. If the eigenvectors are computed, then
  40. *> on return ( CS1, SN1 ) is the unit eigenvector for RT1, hence
  41. *>
  42. *> [ CS1 SN1 ] . [ A B ] . [ CS1 -SN1 ] = [ RT1 0 ]
  43. *> [ -SN1 CS1 ] [ B C ] [ SN1 CS1 ] [ 0 RT2 ]
  44. *> \endverbatim
  45. *
  46. * Arguments:
  47. * ==========
  48. *
  49. *> \param[in] A
  50. *> \verbatim
  51. *> A is COMPLEX
  52. *> The ( 1, 1 ) element of input matrix.
  53. *> \endverbatim
  54. *>
  55. *> \param[in] B
  56. *> \verbatim
  57. *> B is COMPLEX
  58. *> The ( 1, 2 ) element of input matrix. The ( 2, 1 ) element
  59. *> is also given by B, since the 2-by-2 matrix is symmetric.
  60. *> \endverbatim
  61. *>
  62. *> \param[in] C
  63. *> \verbatim
  64. *> C is COMPLEX
  65. *> The ( 2, 2 ) element of input matrix.
  66. *> \endverbatim
  67. *>
  68. *> \param[out] RT1
  69. *> \verbatim
  70. *> RT1 is COMPLEX
  71. *> The eigenvalue of larger modulus.
  72. *> \endverbatim
  73. *>
  74. *> \param[out] RT2
  75. *> \verbatim
  76. *> RT2 is COMPLEX
  77. *> The eigenvalue of smaller modulus.
  78. *> \endverbatim
  79. *>
  80. *> \param[out] EVSCAL
  81. *> \verbatim
  82. *> EVSCAL is COMPLEX
  83. *> The complex value by which the eigenvector matrix was scaled
  84. *> to make it orthonormal. If EVSCAL is zero, the eigenvectors
  85. *> were not computed. This means one of two things: the 2-by-2
  86. *> matrix could not be diagonalized, or the norm of the matrix
  87. *> of eigenvectors before scaling was larger than the threshold
  88. *> value THRESH (set below).
  89. *> \endverbatim
  90. *>
  91. *> \param[out] CS1
  92. *> \verbatim
  93. *> CS1 is COMPLEX
  94. *> \endverbatim
  95. *>
  96. *> \param[out] SN1
  97. *> \verbatim
  98. *> SN1 is COMPLEX
  99. *> If EVSCAL .NE. 0, ( CS1, SN1 ) is the unit right eigenvector
  100. *> for RT1.
  101. *> \endverbatim
  102. *
  103. * Authors:
  104. * ========
  105. *
  106. *> \author Univ. of Tennessee
  107. *> \author Univ. of California Berkeley
  108. *> \author Univ. of Colorado Denver
  109. *> \author NAG Ltd.
  110. *
  111. *> \ingroup complexSYauxiliary
  112. *
  113. * =====================================================================
  114. SUBROUTINE CLAESY( A, B, C, RT1, RT2, EVSCAL, CS1, SN1 )
  115. *
  116. * -- LAPACK auxiliary routine --
  117. * -- LAPACK is a software package provided by Univ. of Tennessee, --
  118. * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  119. *
  120. * .. Scalar Arguments ..
  121. COMPLEX A, B, C, CS1, EVSCAL, RT1, RT2, SN1
  122. * ..
  123. *
  124. * =====================================================================
  125. *
  126. * .. Parameters ..
  127. REAL ZERO
  128. PARAMETER ( ZERO = 0.0E0 )
  129. REAL ONE
  130. PARAMETER ( ONE = 1.0E0 )
  131. COMPLEX CONE
  132. PARAMETER ( CONE = ( 1.0E0, 0.0E0 ) )
  133. REAL HALF
  134. PARAMETER ( HALF = 0.5E0 )
  135. REAL THRESH
  136. PARAMETER ( THRESH = 0.1E0 )
  137. * ..
  138. * .. Local Scalars ..
  139. REAL BABS, EVNORM, TABS, Z
  140. COMPLEX S, T, TMP
  141. * ..
  142. * .. Intrinsic Functions ..
  143. INTRINSIC ABS, MAX, SQRT
  144. * ..
  145. * .. Executable Statements ..
  146. *
  147. *
  148. * Special case: The matrix is actually diagonal.
  149. * To avoid divide by zero later, we treat this case separately.
  150. *
  151. IF( ABS( B ).EQ.ZERO ) THEN
  152. RT1 = A
  153. RT2 = C
  154. IF( ABS( RT1 ).LT.ABS( RT2 ) ) THEN
  155. TMP = RT1
  156. RT1 = RT2
  157. RT2 = TMP
  158. CS1 = ZERO
  159. SN1 = ONE
  160. ELSE
  161. CS1 = ONE
  162. SN1 = ZERO
  163. END IF
  164. ELSE
  165. *
  166. * Compute the eigenvalues and eigenvectors.
  167. * The characteristic equation is
  168. * lambda **2 - (A+C) lambda + (A*C - B*B)
  169. * and we solve it using the quadratic formula.
  170. *
  171. S = ( A+C )*HALF
  172. T = ( A-C )*HALF
  173. *
  174. * Take the square root carefully to avoid over/under flow.
  175. *
  176. BABS = ABS( B )
  177. TABS = ABS( T )
  178. Z = MAX( BABS, TABS )
  179. IF( Z.GT.ZERO )
  180. $ T = Z*SQRT( ( T / Z )**2+( B / Z )**2 )
  181. *
  182. * Compute the two eigenvalues. RT1 and RT2 are exchanged
  183. * if necessary so that RT1 will have the greater magnitude.
  184. *
  185. RT1 = S + T
  186. RT2 = S - T
  187. IF( ABS( RT1 ).LT.ABS( RT2 ) ) THEN
  188. TMP = RT1
  189. RT1 = RT2
  190. RT2 = TMP
  191. END IF
  192. *
  193. * Choose CS1 = 1 and SN1 to satisfy the first equation, then
  194. * scale the components of this eigenvector so that the matrix
  195. * of eigenvectors X satisfies X * X**T = I . (No scaling is
  196. * done if the norm of the eigenvalue matrix is less than THRESH.)
  197. *
  198. SN1 = ( RT1-A ) / B
  199. TABS = ABS( SN1 )
  200. IF( TABS.GT.ONE ) THEN
  201. T = TABS*SQRT( ( ONE / TABS )**2+( SN1 / TABS )**2 )
  202. ELSE
  203. T = SQRT( CONE+SN1*SN1 )
  204. END IF
  205. EVNORM = ABS( T )
  206. IF( EVNORM.GE.THRESH ) THEN
  207. EVSCAL = CONE / T
  208. CS1 = EVSCAL
  209. SN1 = SN1*EVSCAL
  210. ELSE
  211. EVSCAL = ZERO
  212. END IF
  213. END IF
  214. RETURN
  215. *
  216. * End of CLAESY
  217. *
  218. END