Skip Navigation Links
Numerical Libraries
Linear Algebra
Differential Equations
Optimization
Samples
Skip Navigation Links
Linear Algebra
CSLapack
CSBlas
   1:  #region Translated by Jose Antonio De Santiago-Castillo.
   2:   
   3:  //Translated by Jose Antonio De Santiago-Castillo. 
   4:  //E-mail:JAntonioDeSantiago@gmail.com
   5:  //Web: www.DotNumerics.com
   6:  //
   7:  //Fortran to C# Translation.
   8:  //Translated by:
   9:  //F2CSharp Version 0.71 (November 10, 2009)
  10:  //Code Optimizations: None
  11:  //
  12:  #endregion
  13:   
  14:  using System;
  15:  using DotNumerics.FortranLibrary;
  16:   
  17:  namespace DotNumerics.CSLapack
  18:  {
  19:      /// <summary>
  20:      /// -- LAPACK driver routine (version 3.1) --
  21:      /// Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
  22:      /// November 2006
  23:      /// Purpose
  24:      /// =======
  25:      /// 
  26:      /// DGGLSE solves the linear equality-constrained least squares (LSE)
  27:      /// problem:
  28:      /// 
  29:      /// minimize || c - A*x ||_2   subject to   B*x = d
  30:      /// 
  31:      /// where A is an M-by-N matrix, B is a P-by-N matrix, c is a given
  32:      /// M-vector, and d is a given P-vector. It is assumed that
  33:      /// P .LE. N .LE. M+P, and
  34:      /// 
  35:      /// rank(B) = P and  rank( (A) ) = N.
  36:      /// ( (B) )
  37:      /// 
  38:      /// These conditions ensure that the LSE problem has a unique solution,
  39:      /// which is obtained using a generalized RQ factorization of the
  40:      /// matrices (B, A) given by
  41:      /// 
  42:      /// B = (0 R)*Q,   A = Z*T*Q.
  43:      /// 
  44:      ///</summary>
  45:      public class DGGLSE
  46:      {
  47:      
  48:   
  49:          #region Dependencies
  50:          
  51:          DAXPY _daxpy; DCOPY _dcopy; DGEMV _dgemv; DGGRQF _dggrqf; DORMQR _dormqr; DORMRQ _dormrq; DTRMV _dtrmv; DTRTRS _dtrtrs; 
  52:          XERBLA _xerbla;ILAENV _ilaenv; 
  53:   
  54:          #endregion
  55:   
  56:   
  57:          #region Fields
  58:          
  59:          const double ONE = 1.0E+0; bool LQUERY = false; int LOPT = 0; int LWKMIN = 0; int LWKOPT = 0; int MN = 0; int NB = 0; 
  60:          int NB1 = 0;int NB2 = 0; int NB3 = 0; int NB4 = 0; int NR = 0; 
  61:   
  62:          #endregion
  63:   
  64:          public DGGLSE(DAXPY daxpy, DCOPY dcopy, DGEMV dgemv, DGGRQF dggrqf, DORMQR dormqr, DORMRQ dormrq, DTRMV dtrmv, DTRTRS dtrtrs, XERBLA xerbla, ILAENV ilaenv)
  65:          {
  66:      
  67:   
  68:              #region Set Dependencies
  69:              
  70:              this._daxpy = daxpy; this._dcopy = dcopy; this._dgemv = dgemv; this._dggrqf = dggrqf; this._dormqr = dormqr; 
  71:              this._dormrq = dormrq;this._dtrmv = dtrmv; this._dtrtrs = dtrtrs; this._xerbla = xerbla; this._ilaenv = ilaenv; 
  72:   
  73:              #endregion
  74:   
  75:          }
  76:      
  77:          public DGGLSE()
  78:          {
  79:      
  80:   
  81:              #region Dependencies (Initialization)
  82:              
  83:              DAXPY daxpy = new DAXPY();
  84:              DCOPY dcopy = new DCOPY();
  85:              LSAME lsame = new LSAME();
  86:              XERBLA xerbla = new XERBLA();
  87:              DLAMC3 dlamc3 = new DLAMC3();
  88:              DLAPY2 dlapy2 = new DLAPY2();
  89:              DNRM2 dnrm2 = new DNRM2();
  90:              DSCAL dscal = new DSCAL();
  91:              IEEECK ieeeck = new IEEECK();
  92:              IPARMQ iparmq = new IPARMQ();
  93:              DGEMV dgemv = new DGEMV(lsame, xerbla);
  94:              DGER dger = new DGER(xerbla);
  95:              DLARF dlarf = new DLARF(dgemv, dger, lsame);
  96:              DLAMC1 dlamc1 = new DLAMC1(dlamc3);
  97:              DLAMC4 dlamc4 = new DLAMC4(dlamc3);
  98:              DLAMC5 dlamc5 = new DLAMC5(dlamc3);
  99:              DLAMC2 dlamc2 = new DLAMC2(dlamc3, dlamc1, dlamc4, dlamc5);
 100:              DLAMCH dlamch = new DLAMCH(lsame, dlamc2);
 101:              DLARFG dlarfg = new DLARFG(dlamch, dlapy2, dnrm2, dscal);
 102:              DGEQR2 dgeqr2 = new DGEQR2(dlarf, dlarfg, xerbla);
 103:              DGEMM dgemm = new DGEMM(lsame, xerbla);
 104:              DTRMM dtrmm = new DTRMM(lsame, xerbla);
 105:              DLARFB dlarfb = new DLARFB(lsame, dcopy, dgemm, dtrmm);
 106:              DTRMV dtrmv = new DTRMV(lsame, xerbla);
 107:              DLARFT dlarft = new DLARFT(dgemv, dtrmv, lsame);
 108:              ILAENV ilaenv = new ILAENV(ieeeck, iparmq);
 109:              DGEQRF dgeqrf = new DGEQRF(dgeqr2, dlarfb, dlarft, xerbla, ilaenv);
 110:              DGERQ2 dgerq2 = new DGERQ2(dlarf, dlarfg, xerbla);
 111:              DGERQF dgerqf = new DGERQF(dgerq2, dlarfb, dlarft, xerbla, ilaenv);
 112:              DORMR2 dormr2 = new DORMR2(lsame, dlarf, xerbla);
 113:              DORMRQ dormrq = new DORMRQ(lsame, ilaenv, dlarfb, dlarft, dormr2, xerbla);
 114:              DGGRQF dggrqf = new DGGRQF(dgeqrf, dgerqf, dormrq, xerbla, ilaenv);
 115:              DORM2R dorm2r = new DORM2R(lsame, dlarf, xerbla);
 116:              DORMQR dormqr = new DORMQR(lsame, ilaenv, dlarfb, dlarft, dorm2r, xerbla);
 117:              DTRSM dtrsm = new DTRSM(lsame, xerbla);
 118:              DTRTRS dtrtrs = new DTRTRS(lsame, dtrsm, xerbla);
 119:   
 120:              #endregion
 121:   
 122:   
 123:              #region Set Dependencies
 124:              
 125:              this._daxpy = daxpy; this._dcopy = dcopy; this._dgemv = dgemv; this._dggrqf = dggrqf; this._dormqr = dormqr; 
 126:              this._dormrq = dormrq;this._dtrmv = dtrmv; this._dtrtrs = dtrtrs; this._xerbla = xerbla; this._ilaenv = ilaenv; 
 127:   
 128:              #endregion
 129:   
 130:          }
 131:          /// <summary>
 132:          /// Purpose
 133:          /// =======
 134:          /// 
 135:          /// DGGLSE solves the linear equality-constrained least squares (LSE)
 136:          /// problem:
 137:          /// 
 138:          /// minimize || c - A*x ||_2   subject to   B*x = d
 139:          /// 
 140:          /// where A is an M-by-N matrix, B is a P-by-N matrix, c is a given
 141:          /// M-vector, and d is a given P-vector. It is assumed that
 142:          /// P .LE. N .LE. M+P, and
 143:          /// 
 144:          /// rank(B) = P and  rank( (A) ) = N.
 145:          /// ( (B) )
 146:          /// 
 147:          /// These conditions ensure that the LSE problem has a unique solution,
 148:          /// which is obtained using a generalized RQ factorization of the
 149:          /// matrices (B, A) given by
 150:          /// 
 151:          /// B = (0 R)*Q,   A = Z*T*Q.
 152:          /// 
 153:          ///</summary>
 154:          /// <param name="M">
 155:          /// (input) INTEGER
 156:          /// The number of rows of the matrix A.  M .GE. 0.
 157:          ///</param>
 158:          /// <param name="N">
 159:          /// (input) INTEGER
 160:          /// The number of columns of the matrices A and B. N .GE. 0.
 161:          ///</param>
 162:          /// <param name="P">
 163:          /// (input) INTEGER
 164:          /// The number of rows of the matrix B. 0 .LE. P .LE. N .LE. M+P.
 165:          ///</param>
 166:          /// <param name="A">
 167:          /// (input/output) DOUBLE PRECISION array, dimension (LDA,N)
 168:          /// On entry, the M-by-N matrix A.
 169:          /// On exit, the elements on and above the diagonal of the array
 170:          /// contain the min(M,N)-by-N upper trapezoidal matrix T.
 171:          ///</param>
 172:          /// <param name="LDA">
 173:          /// (input) INTEGER
 174:          /// The leading dimension of the array A. LDA .GE. max(1,M).
 175:          ///</param>
 176:          /// <param name="B">
 177:          /// = (0 R)*Q,   A = Z*T*Q.
 178:          ///</param>
 179:          /// <param name="LDB">
 180:          /// (input) INTEGER
 181:          /// The leading dimension of the array B. LDB .GE. max(1,P).
 182:          ///</param>
 183:          /// <param name="C">
 184:          /// (input/output) DOUBLE PRECISION array, dimension (M)
 185:          /// On entry, C contains the right hand side vector for the
 186:          /// least squares part of the LSE problem.
 187:          /// On exit, the residual sum of squares for the solution
 188:          /// is given by the sum of squares of elements N-P+1 to M of
 189:          /// vector C.
 190:          ///</param>
 191:          /// <param name="D">
 192:          /// (input/output) DOUBLE PRECISION array, dimension (P)
 193:          /// On entry, D contains the right hand side vector for the
 194:          /// constrained equation.
 195:          /// On exit, D is destroyed.
 196:          ///</param>
 197:          /// <param name="X">
 198:          /// (output) DOUBLE PRECISION array, dimension (N)
 199:          /// On exit, X is the solution of the LSE problem.
 200:          ///</param>
 201:          /// <param name="WORK">
 202:          /// (workspace/output) DOUBLE PRECISION array, dimension (MAX(1,LWORK))
 203:          /// On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
 204:          ///</param>
 205:          /// <param name="LWORK">
 206:          /// (input) INTEGER
 207:          /// The dimension of the array WORK. LWORK .GE. max(1,M+N+P).
 208:          /// For optimum performance LWORK .GE. P+min(M,N)+max(M,N)*NB,
 209:          /// where NB is an upper bound for the optimal blocksizes for
 210:          /// DGEQRF, SGERQF, DORMQR and SORMRQ.
 211:          /// 
 212:          /// If LWORK = -1, then a workspace query is assumed; the routine
 213:          /// only calculates the optimal size of the WORK array, returns
 214:          /// this value as the first entry of the WORK array, and no error
 215:          /// message related to LWORK is issued by XERBLA.
 216:          ///</param>
 217:          /// <param name="INFO">
 218:          /// (output) INTEGER
 219:          /// = 0:  successful exit.
 220:          /// .LT. 0:  if INFO = -i, the i-th argument had an illegal value.
 221:          /// = 1:  the upper triangular factor R associated with B in the
 222:          /// generalized RQ factorization of the pair (B, A) is
 223:          /// singular, so that rank(B) .LT. P; the least squares
 224:          /// solution could not be computed.
 225:          /// = 2:  the (N-P) by (N-P) part of the upper trapezoidal factor
 226:          /// T associated with A in the generalized RQ factorization
 227:          /// of the pair (B, A) is singular, so that
 228:          /// rank( (A) ) .LT. N; the least squares solution could not
 229:          /// ( (B) )
 230:          /// be computed.
 231:          ///</param>
 232:          public void Run(int M, int N, int P, ref double[] A, int offset_a, int LDA, ref double[] B, int offset_b
 233:                           , int LDB, ref double[] C, int offset_c, ref double[] D, int offset_d, ref double[] X, int offset_x, ref double[] WORK, int offset_work, int LWORK
 234:                           , ref int INFO)
 235:          {
 236:   
 237:              #region Array Index Correction
 238:              
 239:               int o_a = -1 - LDA + offset_a;  int o_b = -1 - LDB + offset_b;  int o_c = -1 + offset_c;  int o_d = -1 + offset_d; 
 240:               int o_x = -1 + offset_x; int o_work = -1 + offset_work; 
 241:   
 242:              #endregion
 243:   
 244:   
 245:              #region Prolog
 246:              
 247:              // *
 248:              // *  -- LAPACK driver routine (version 3.1) --
 249:              // *     Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
 250:              // *     November 2006
 251:              // *
 252:              // *     .. Scalar Arguments ..
 253:              // *     ..
 254:              // *     .. Array Arguments ..
 255:              // *     ..
 256:              // *
 257:              // *  Purpose
 258:              // *  =======
 259:              // *
 260:              // *  DGGLSE solves the linear equality-constrained least squares (LSE)
 261:              // *  problem:
 262:              // *
 263:              // *          minimize || c - A*x ||_2   subject to   B*x = d
 264:              // *
 265:              // *  where A is an M-by-N matrix, B is a P-by-N matrix, c is a given
 266:              // *  M-vector, and d is a given P-vector. It is assumed that
 267:              // *  P <= N <= M+P, and
 268:              // *
 269:              // *           rank(B) = P and  rank( (A) ) = N.
 270:              // *                                ( (B) )
 271:              // *
 272:              // *  These conditions ensure that the LSE problem has a unique solution,
 273:              // *  which is obtained using a generalized RQ factorization of the
 274:              // *  matrices (B, A) given by
 275:              // *
 276:              // *     B = (0 R)*Q,   A = Z*T*Q.
 277:              // *
 278:              // *  Arguments
 279:              // *  =========
 280:              // *
 281:              // *  M       (input) INTEGER
 282:              // *          The number of rows of the matrix A.  M >= 0.
 283:              // *
 284:              // *  N       (input) INTEGER
 285:              // *          The number of columns of the matrices A and B. N >= 0.
 286:              // *
 287:              // *  P       (input) INTEGER
 288:              // *          The number of rows of the matrix B. 0 <= P <= N <= M+P.
 289:              // *
 290:              // *  A       (input/output) DOUBLE PRECISION array, dimension (LDA,N)
 291:              // *          On entry, the M-by-N matrix A.
 292:              // *          On exit, the elements on and above the diagonal of the array
 293:              // *          contain the min(M,N)-by-N upper trapezoidal matrix T.
 294:              // *
 295:              // *  LDA     (input) INTEGER
 296:              // *          The leading dimension of the array A. LDA >= max(1,M).
 297:              // *
 298:              // *  B       (input/output) DOUBLE PRECISION array, dimension (LDB,N)
 299:              // *          On entry, the P-by-N matrix B.
 300:              // *          On exit, the upper triangle of the subarray B(1:P,N-P+1:N)
 301:              // *          contains the P-by-P upper triangular matrix R.
 302:              // *
 303:              // *  LDB     (input) INTEGER
 304:              // *          The leading dimension of the array B. LDB >= max(1,P).
 305:              // *
 306:              // *  C       (input/output) DOUBLE PRECISION array, dimension (M)
 307:              // *          On entry, C contains the right hand side vector for the
 308:              // *          least squares part of the LSE problem.
 309:              // *          On exit, the residual sum of squares for the solution
 310:              // *          is given by the sum of squares of elements N-P+1 to M of
 311:              // *          vector C.
 312:              // *
 313:              // *  D       (input/output) DOUBLE PRECISION array, dimension (P)
 314:              // *          On entry, D contains the right hand side vector for the
 315:              // *          constrained equation.
 316:              // *          On exit, D is destroyed.
 317:              // *
 318:              // *  X       (output) DOUBLE PRECISION array, dimension (N)
 319:              // *          On exit, X is the solution of the LSE problem.
 320:              // *
 321:              // *  WORK    (workspace/output) DOUBLE PRECISION array, dimension (MAX(1,LWORK))
 322:              // *          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
 323:              // *
 324:              // *  LWORK   (input) INTEGER
 325:              // *          The dimension of the array WORK. LWORK >= max(1,M+N+P).
 326:              // *          For optimum performance LWORK >= P+min(M,N)+max(M,N)*NB,
 327:              // *          where NB is an upper bound for the optimal blocksizes for
 328:              // *          DGEQRF, SGERQF, DORMQR and SORMRQ.
 329:              // *
 330:              // *          If LWORK = -1, then a workspace query is assumed; the routine
 331:              // *          only calculates the optimal size of the WORK array, returns
 332:              // *          this value as the first entry of the WORK array, and no error
 333:              // *          message related to LWORK is issued by XERBLA.
 334:              // *
 335:              // *  INFO    (output) INTEGER
 336:              // *          = 0:  successful exit.
 337:              // *          < 0:  if INFO = -i, the i-th argument had an illegal value.
 338:              // *          = 1:  the upper triangular factor R associated with B in the
 339:              // *                generalized RQ factorization of the pair (B, A) is
 340:              // *                singular, so that rank(B) < P; the least squares
 341:              // *                solution could not be computed.
 342:              // *          = 2:  the (N-P) by (N-P) part of the upper trapezoidal factor
 343:              // *                T associated with A in the generalized RQ factorization
 344:              // *                of the pair (B, A) is singular, so that
 345:              // *                rank( (A) ) < N; the least squares solution could not
 346:              // *                    ( (B) )
 347:              // *                be computed.
 348:              // *
 349:              // *  =====================================================================
 350:              // *
 351:              // *     .. Parameters ..
 352:              // *     ..
 353:              // *     .. Local Scalars ..
 354:              // *     ..
 355:              // *     .. External Subroutines ..
 356:              // *     ..
 357:              // *     .. External Functions ..
 358:              // *     ..
 359:              // *     .. Intrinsic Functions ..
 360:              //      INTRINSIC          INT, MAX, MIN;
 361:              // *     ..
 362:              // *     .. Executable Statements ..
 363:              // *
 364:              // *     Test the input parameters
 365:              // *
 366:   
 367:              #endregion
 368:   
 369:   
 370:              #region Body
 371:              
 372:              INFO = 0;
 373:              MN = Math.Min(M, N);
 374:              LQUERY = (LWORK ==  - 1);
 375:              if (M < 0)
 376:              {
 377:                  INFO =  - 1;
 378:              }
 379:              else
 380:              {
 381:                  if (N < 0)
 382:                  {
 383:                      INFO =  - 2;
 384:                  }
 385:                  else
 386:                  {
 387:                      if (P < 0 || P > N || P < N - M)
 388:                      {
 389:                          INFO =  - 3;
 390:                      }
 391:                      else
 392:                      {
 393:                          if (LDA < Math.Max(1, M))
 394:                          {
 395:                              INFO =  - 5;
 396:                          }
 397:                          else
 398:                          {
 399:                              if (LDB < Math.Max(1, P))
 400:                              {
 401:                                  INFO =  - 7;
 402:                              }
 403:                          }
 404:                      }
 405:                  }
 406:              }
 407:              // *
 408:              // *     Calculate workspace
 409:              // *
 410:              if (INFO == 0)
 411:              {
 412:                  if (N == 0)
 413:                  {
 414:                      LWKMIN = 1;
 415:                      LWKOPT = 1;
 416:                  }
 417:                  else
 418:                  {
 419:                      NB1 = this._ilaenv.Run(1, "DGEQRF", " ", M, N,  - 1,  - 1);
 420:                      NB2 = this._ilaenv.Run(1, "DGERQF", " ", M, N,  - 1,  - 1);
 421:                      NB3 = this._ilaenv.Run(1, "DORMQR", " ", M, N, P,  - 1);
 422:                      NB4 = this._ilaenv.Run(1, "DORMRQ", " ", M, N, P,  - 1);
 423:                      NB = Math.Max(NB1, Math.Max(NB2, Math.Max(NB3, NB4)));
 424:                      LWKMIN = M + N + P;
 425:                      LWKOPT = P + MN + Math.Max(M, N) * NB;
 426:                  }
 427:                  WORK[1 + o_work] = LWKOPT;
 428:                  // *
 429:                  if (LWORK < LWKMIN && !LQUERY)
 430:                  {
 431:                      INFO =  - 12;
 432:                  }
 433:              }
 434:              // *
 435:              if (INFO != 0)
 436:              {
 437:                  this._xerbla.Run("DGGLSE",  - INFO);
 438:                  return;
 439:              }
 440:              else
 441:              {
 442:                  if (LQUERY)
 443:                  {
 444:                      return;
 445:                  }
 446:              }
 447:              // *
 448:              // *     Quick return if possible
 449:              // *
 450:              if (N == 0) return;
 451:              // *
 452:              // *     Compute the GRQ factorization of matrices B and A:
 453:              // *
 454:              // *            B*Q' = (  0  T12 ) P   Z'*A*Q' = ( R11 R12 ) N-P
 455:              // *                     N-P  P                  (  0  R22 ) M+P-N
 456:              // *                                               N-P  P
 457:              // *
 458:              // *     where T12 and R11 are upper triangular, and Q and Z are
 459:              // *     orthogonal.
 460:              // *
 461:              this._dggrqf.Run(P, M, N, ref B, offset_b, LDB, ref WORK, offset_work
 462:                               , ref A, offset_a, LDA, ref WORK, P + 1 + o_work, ref WORK, P + MN + 1 + o_work, LWORK - P - MN, ref INFO);
 463:              LOPT = (int)WORK[P + MN + 1 + o_work];
 464:              // *
 465:              // *     Update c = Z'*c = ( c1 ) N-P
 466:              // *                       ( c2 ) M+P-N
 467:              // *
 468:              this._dormqr.Run("Left", "Transpose", M, 1, MN, ref A, offset_a
 469:                               , LDA, WORK, P + 1 + o_work, ref C, offset_c, Math.Max(1, M), ref WORK, P + MN + 1 + o_work, LWORK - P - MN
 470:                               , ref INFO);
 471:              LOPT = Math.Max(LOPT, Convert.ToInt32(Math.Truncate(WORK[P + MN + 1 + o_work])));
 472:              // *
 473:              // *     Solve T12*x2 = d for x2
 474:              // *
 475:              if (P > 0)
 476:              {
 477:                  this._dtrtrs.Run("Upper", "No transpose", "Non-unit", P, 1, B, 1+(N - P + 1) * LDB + o_b
 478:                                   , LDB, ref D, offset_d, P, ref INFO);
 479:                  // *
 480:                  if (INFO > 0)
 481:                  {
 482:                      INFO = 1;
 483:                      return;
 484:                  }
 485:                  // *
 486:                  // *        Put the solution in X
 487:                  // *
 488:                  this._dcopy.Run(P, D, offset_d, 1, ref X, N - P + 1 + o_x, 1);
 489:                  // *
 490:                  // *        Update c1
 491:                  // *
 492:                  this._dgemv.Run("No transpose", N - P, P,  - ONE, A, 1+(N - P + 1) * LDA + o_a, LDA
 493:                                  , D, offset_d, 1, ONE, ref C, offset_c, 1);
 494:              }
 495:              // *
 496:              // *     Solve R11*x1 = c1 for x1
 497:              // *
 498:              if (N > P)
 499:              {
 500:                  this._dtrtrs.Run("Upper", "No transpose", "Non-unit", N - P, 1, A, offset_a
 501:                                   , LDA, ref C, offset_c, N - P, ref INFO);
 502:                  // *
 503:                  if (INFO > 0)
 504:                  {
 505:                      INFO = 2;
 506:                      return;
 507:                  }
 508:                  // *
 509:                  // *        Put the solutions in X
 510:                  // *
 511:                  this._dcopy.Run(N - P, C, offset_c, 1, ref X, offset_x, 1);
 512:              }
 513:              // *
 514:              // *     Compute the residual vector:
 515:              // *
 516:              if (M < N)
 517:              {
 518:                  NR = M + P - N;
 519:                  if (NR > 0)
 520:                  {
 521:                      this._dgemv.Run("No transpose", NR, N - M,  - ONE, A, N - P + 1+(M + 1) * LDA + o_a, LDA
 522:                                      , D, NR + 1 + o_d, 1, ONE, ref C, N - P + 1 + o_c, 1);
 523:                  }
 524:              }
 525:              else
 526:              {
 527:                  NR = P;
 528:              }
 529:              if (NR > 0)
 530:              {
 531:                  this._dtrmv.Run("Upper", "No transpose", "Non unit", NR, A, N - P + 1+(N - P + 1) * LDA + o_a, LDA
 532:                                  , ref D, offset_d, 1);
 533:                  this._daxpy.Run(NR,  - ONE, D, offset_d, 1, ref C, N - P + 1 + o_c, 1);
 534:              }
 535:              // *
 536:              // *     Backward transformation x = Q'*x
 537:              // *
 538:              this._dormrq.Run("Left", "Transpose", N, 1, P, ref B, offset_b
 539:                               , LDB, WORK, 1 + o_work, ref X, offset_x, N, ref WORK, P + MN + 1 + o_work, LWORK - P - MN
 540:                               , ref INFO);
 541:              WORK[1 + o_work] = P + MN + Math.Max(LOPT, Convert.ToInt32(Math.Truncate(WORK[P + MN + 1 + o_work])));
 542:              // *
 543:              return;
 544:              // *
 545:              // *     End of DGGLSE
 546:              // *
 547:   
 548:              #endregion
 549:   
 550:          }
 551:      }
 552:  }