1 2 #include <../src/snes/impls/vi/ss/vissimpl.h> /*I "petscsnes.h" I*/ 3 #include <../include/petsc-private/kspimpl.h> 4 #include <../include/petsc-private/matimpl.h> 5 #include <../include/petsc-private/dmimpl.h> 6 7 8 /* 9 SNESVIComputeMeritFunction - Evaluates the merit function for the mixed complementarity problem. 10 11 Input Parameter: 12 . phi - the semismooth function 13 14 Output Parameter: 15 . merit - the merit function 16 . phinorm - ||phi|| 17 18 Notes: 19 The merit function for the mixed complementarity problem is defined as 20 merit = 0.5*phi^T*phi 21 */ 22 #undef __FUNCT__ 23 #define __FUNCT__ "SNESVIComputeMeritFunction" 24 static PetscErrorCode SNESVIComputeMeritFunction(Vec phi, PetscReal *merit,PetscReal *phinorm) 25 { 26 PetscErrorCode ierr; 27 28 PetscFunctionBegin; 29 ierr = VecNormBegin(phi,NORM_2,phinorm);CHKERRQ(ierr); 30 ierr = VecNormEnd(phi,NORM_2,phinorm);CHKERRQ(ierr); 31 32 *merit = 0.5*(*phinorm)*(*phinorm); 33 PetscFunctionReturn(0); 34 } 35 36 PETSC_STATIC_INLINE PetscScalar Phi(PetscScalar a,PetscScalar b) 37 { 38 return a + b - PetscSqrtScalar(a*a + b*b); 39 } 40 41 PETSC_STATIC_INLINE PetscScalar DPhi(PetscScalar a,PetscScalar b) 42 { 43 if ((PetscAbsScalar(a) >= 1.e-6) || (PetscAbsScalar(b) >= 1.e-6)) return 1.0 - a/ PetscSqrtScalar(a*a + b*b); 44 else return .5; 45 } 46 47 /* 48 SNESVIComputeFunction - Reformulates a system of nonlinear equations in mixed complementarity form to a system of nonlinear equations in semismooth form. 49 50 Input Parameters: 51 . snes - the SNES context 52 . x - current iterate 53 . functx - user defined function context 54 55 Output Parameters: 56 . phi - Semismooth function 57 58 */ 59 #undef __FUNCT__ 60 #define __FUNCT__ "SNESVIComputeFunction" 61 static PetscErrorCode SNESVIComputeFunction(SNES snes,Vec X,Vec phi,void *functx) 62 { 63 PetscErrorCode ierr; 64 SNES_VINEWTONSSLS *vi = (SNES_VINEWTONSSLS*)snes->data; 65 Vec Xl = snes->xl,Xu = snes->xu,F = snes->vec_func; 66 PetscScalar *phi_arr,*x_arr,*f_arr,*l,*u; 67 PetscInt i,nlocal; 68 69 PetscFunctionBegin; 70 ierr = (*vi->computeuserfunction)(snes,X,F,functx);CHKERRQ(ierr); 71 ierr = VecGetLocalSize(X,&nlocal);CHKERRQ(ierr); 72 ierr = VecGetArray(X,&x_arr);CHKERRQ(ierr); 73 ierr = VecGetArray(F,&f_arr);CHKERRQ(ierr); 74 ierr = VecGetArray(Xl,&l);CHKERRQ(ierr); 75 ierr = VecGetArray(Xu,&u);CHKERRQ(ierr); 76 ierr = VecGetArray(phi,&phi_arr);CHKERRQ(ierr); 77 78 for (i=0; i < nlocal; i++) { 79 if ((PetscRealPart(l[i]) <= SNES_VI_NINF) && (PetscRealPart(u[i]) >= SNES_VI_INF)) { /* no constraints on variable */ 80 phi_arr[i] = f_arr[i]; 81 } else if (PetscRealPart(l[i]) <= SNES_VI_NINF) { /* upper bound on variable only */ 82 phi_arr[i] = -Phi(u[i] - x_arr[i],-f_arr[i]); 83 } else if (PetscRealPart(u[i]) >= SNES_VI_INF) { /* lower bound on variable only */ 84 phi_arr[i] = Phi(x_arr[i] - l[i],f_arr[i]); 85 } else if (l[i] == u[i]) { 86 phi_arr[i] = l[i] - x_arr[i]; 87 } else { /* both bounds on variable */ 88 phi_arr[i] = Phi(x_arr[i] - l[i],-Phi(u[i] - x_arr[i],-f_arr[i])); 89 } 90 } 91 92 ierr = VecRestoreArray(X,&x_arr);CHKERRQ(ierr); 93 ierr = VecRestoreArray(F,&f_arr);CHKERRQ(ierr); 94 ierr = VecRestoreArray(Xl,&l);CHKERRQ(ierr); 95 ierr = VecRestoreArray(Xu,&u);CHKERRQ(ierr); 96 ierr = VecRestoreArray(phi,&phi_arr);CHKERRQ(ierr); 97 PetscFunctionReturn(0); 98 } 99 100 /* 101 SNESVIComputeBsubdifferentialVectors - Computes the diagonal shift (Da) and row scaling (Db) vectors needed for the 102 the semismooth jacobian. 103 */ 104 #undef __FUNCT__ 105 #define __FUNCT__ "SNESVIComputeBsubdifferentialVectors" 106 PetscErrorCode SNESVIComputeBsubdifferentialVectors(SNES snes,Vec X,Vec F,Mat jac,Vec Da,Vec Db) 107 { 108 PetscErrorCode ierr; 109 PetscScalar *l,*u,*x,*f,*da,*db,da1,da2,db1,db2; 110 PetscInt i,nlocal; 111 112 PetscFunctionBegin; 113 ierr = VecGetArray(X,&x);CHKERRQ(ierr); 114 ierr = VecGetArray(F,&f);CHKERRQ(ierr); 115 ierr = VecGetArray(snes->xl,&l);CHKERRQ(ierr); 116 ierr = VecGetArray(snes->xu,&u);CHKERRQ(ierr); 117 ierr = VecGetArray(Da,&da);CHKERRQ(ierr); 118 ierr = VecGetArray(Db,&db);CHKERRQ(ierr); 119 ierr = VecGetLocalSize(X,&nlocal);CHKERRQ(ierr); 120 121 for (i=0; i< nlocal; i++) { 122 if ((PetscRealPart(l[i]) <= SNES_VI_NINF) && (PetscRealPart(u[i]) >= SNES_VI_INF)) { /* no constraints on variable */ 123 da[i] = 0; 124 db[i] = 1; 125 } else if (PetscRealPart(l[i]) <= SNES_VI_NINF) { /* upper bound on variable only */ 126 da[i] = DPhi(u[i] - x[i], -f[i]); 127 db[i] = DPhi(-f[i],u[i] - x[i]); 128 } else if (PetscRealPart(u[i]) >= SNES_VI_INF) { /* lower bound on variable only */ 129 da[i] = DPhi(x[i] - l[i], f[i]); 130 db[i] = DPhi(f[i],x[i] - l[i]); 131 } else if (l[i] == u[i]) { /* fixed variable */ 132 da[i] = 1; 133 db[i] = 0; 134 } else { /* upper and lower bounds on variable */ 135 da1 = DPhi(x[i] - l[i], -Phi(u[i] - x[i], -f[i])); 136 db1 = DPhi(-Phi(u[i] - x[i], -f[i]),x[i] - l[i]); 137 da2 = DPhi(u[i] - x[i], -f[i]); 138 db2 = DPhi(-f[i],u[i] - x[i]); 139 da[i] = da1 + db1*da2; 140 db[i] = db1*db2; 141 } 142 } 143 144 ierr = VecRestoreArray(X,&x);CHKERRQ(ierr); 145 ierr = VecRestoreArray(F,&f);CHKERRQ(ierr); 146 ierr = VecRestoreArray(snes->xl,&l);CHKERRQ(ierr); 147 ierr = VecRestoreArray(snes->xu,&u);CHKERRQ(ierr); 148 ierr = VecRestoreArray(Da,&da);CHKERRQ(ierr); 149 ierr = VecRestoreArray(Db,&db);CHKERRQ(ierr); 150 PetscFunctionReturn(0); 151 } 152 153 /* 154 SNESVIComputeJacobian - Computes the jacobian of the semismooth function.The Jacobian for the semismooth function is an element of the B-subdifferential of the Fischer-Burmeister function for complementarity problems. 155 156 Input Parameters: 157 . Da - Diagonal shift vector for the semismooth jacobian. 158 . Db - Row scaling vector for the semismooth jacobian. 159 160 Output Parameters: 161 . jac - semismooth jacobian 162 . jac_pre - optional preconditioning matrix 163 164 Notes: 165 The semismooth jacobian matrix is given by 166 jac = Da + Db*jacfun 167 where Db is the row scaling matrix stored as a vector, 168 Da is the diagonal perturbation matrix stored as a vector 169 and jacfun is the jacobian of the original nonlinear function. 170 */ 171 #undef __FUNCT__ 172 #define __FUNCT__ "SNESVIComputeJacobian" 173 PetscErrorCode SNESVIComputeJacobian(Mat jac, Mat jac_pre,Vec Da, Vec Db) 174 { 175 PetscErrorCode ierr; 176 177 /* Do row scaling and add diagonal perturbation */ 178 ierr = MatDiagonalScale(jac,Db,NULL);CHKERRQ(ierr); 179 ierr = MatDiagonalSet(jac,Da,ADD_VALUES);CHKERRQ(ierr); 180 if (jac != jac_pre) { /* If jac and jac_pre are different */ 181 ierr = MatDiagonalScale(jac_pre,Db,NULL);CHKERRQ(ierr); 182 ierr = MatDiagonalSet(jac_pre,Da,ADD_VALUES);CHKERRQ(ierr); 183 } 184 PetscFunctionReturn(0); 185 } 186 187 /* 188 SNESVIComputeMeritFunctionGradient - Computes the gradient of the merit function psi. 189 190 Input Parameters: 191 phi - semismooth function. 192 H - semismooth jacobian 193 194 Output Parameters: 195 dpsi - merit function gradient 196 197 Notes: 198 The merit function gradient is computed as follows 199 dpsi = H^T*phi 200 */ 201 #undef __FUNCT__ 202 #define __FUNCT__ "SNESVIComputeMeritFunctionGradient" 203 PetscErrorCode SNESVIComputeMeritFunctionGradient(Mat H, Vec phi, Vec dpsi) 204 { 205 PetscErrorCode ierr; 206 207 PetscFunctionBegin; 208 ierr = MatMultTranspose(H,phi,dpsi);CHKERRQ(ierr); 209 PetscFunctionReturn(0); 210 } 211 212 213 214 /* 215 SNESSolve_VINEWTONSSLS - Solves the complementarity problem with a semismooth Newton 216 method using a line search. 217 218 Input Parameters: 219 . snes - the SNES context 220 221 Output Parameter: 222 . outits - number of iterations until termination 223 224 Application Interface Routine: SNESSolve() 225 226 Notes: 227 This implements essentially a semismooth Newton method with a 228 line search. The default line search does not do any line seach 229 but rather takes a full newton step. 230 231 Developer Note: the code in this file should be slightly modified so that this routine need not exist and the SNESSolve_NEWTONLS() routine is called directly with the appropriate wrapped function and Jacobian evaluations 232 233 */ 234 #undef __FUNCT__ 235 #define __FUNCT__ "SNESSolve_VINEWTONSSLS" 236 PetscErrorCode SNESSolve_VINEWTONSSLS(SNES snes) 237 { 238 SNES_VINEWTONSSLS *vi = (SNES_VINEWTONSSLS*)snes->data; 239 PetscErrorCode ierr; 240 PetscInt maxits,i,lits; 241 PetscBool lssucceed; 242 MatStructure flg = DIFFERENT_NONZERO_PATTERN; 243 PetscReal gnorm,xnorm=0,ynorm; 244 Vec Y,X,F; 245 KSPConvergedReason kspreason; 246 DM dm; 247 DMSNES sdm; 248 249 PetscFunctionBegin; 250 ierr = SNESGetDM(snes,&dm);CHKERRQ(ierr); 251 ierr = DMGetDMSNES(dm,&sdm);CHKERRQ(ierr); 252 253 vi->computeuserfunction = sdm->ops->computefunction; 254 sdm->ops->computefunction = SNESVIComputeFunction; 255 256 snes->numFailures = 0; 257 snes->numLinearSolveFailures = 0; 258 snes->reason = SNES_CONVERGED_ITERATING; 259 260 maxits = snes->max_its; /* maximum number of iterations */ 261 X = snes->vec_sol; /* solution vector */ 262 F = snes->vec_func; /* residual vector */ 263 Y = snes->work[0]; /* work vectors */ 264 265 ierr = PetscObjectSAWsTakeAccess((PetscObject)snes);CHKERRQ(ierr); 266 snes->iter = 0; 267 snes->norm = 0.0; 268 ierr = PetscObjectSAWsGrantAccess((PetscObject)snes);CHKERRQ(ierr); 269 270 ierr = SNESVIProjectOntoBounds(snes,X);CHKERRQ(ierr); 271 ierr = SNESComputeFunction(snes,X,vi->phi);CHKERRQ(ierr); 272 if (snes->domainerror) { 273 snes->reason = SNES_DIVERGED_FUNCTION_DOMAIN; 274 sdm->ops->computefunction = vi->computeuserfunction; 275 PetscFunctionReturn(0); 276 } 277 /* Compute Merit function */ 278 ierr = SNESVIComputeMeritFunction(vi->phi,&vi->merit,&vi->phinorm);CHKERRQ(ierr); 279 280 ierr = VecNormBegin(X,NORM_2,&xnorm);CHKERRQ(ierr); /* xnorm <- ||x|| */ 281 ierr = VecNormEnd(X,NORM_2,&xnorm);CHKERRQ(ierr); 282 if (PetscIsInfOrNanReal(vi->merit)) SETERRQ(PETSC_COMM_SELF,PETSC_ERR_FP,"User provided compute function generated a Not-a-Number"); 283 284 ierr = PetscObjectSAWsTakeAccess((PetscObject)snes);CHKERRQ(ierr); 285 snes->norm = vi->phinorm; 286 ierr = PetscObjectSAWsGrantAccess((PetscObject)snes);CHKERRQ(ierr); 287 ierr = SNESLogConvergenceHistory(snes,vi->phinorm,0);CHKERRQ(ierr); 288 ierr = SNESMonitor(snes,0,vi->phinorm);CHKERRQ(ierr); 289 290 /* test convergence */ 291 ierr = (*snes->ops->converged)(snes,0,0.0,0.0,vi->phinorm,&snes->reason,snes->cnvP);CHKERRQ(ierr); 292 if (snes->reason) { 293 sdm->ops->computefunction = vi->computeuserfunction; 294 PetscFunctionReturn(0); 295 } 296 297 for (i=0; i<maxits; i++) { 298 299 /* Call general purpose update function */ 300 if (snes->ops->update) { 301 ierr = (*snes->ops->update)(snes, snes->iter);CHKERRQ(ierr); 302 } 303 304 /* Solve J Y = Phi, where J is the semismooth jacobian */ 305 306 /* Get the jacobian -- note that the function must be the original function for snes_fd and snes_fd_color to work for this*/ 307 sdm->ops->computefunction = vi->computeuserfunction; 308 ierr = SNESComputeJacobian(snes,X,&snes->jacobian,&snes->jacobian_pre,&flg);CHKERRQ(ierr); 309 sdm->ops->computefunction = SNESVIComputeFunction; 310 311 /* Get the diagonal shift and row scaling vectors */ 312 ierr = SNESVIComputeBsubdifferentialVectors(snes,X,F,snes->jacobian,vi->Da,vi->Db);CHKERRQ(ierr); 313 /* Compute the semismooth jacobian */ 314 ierr = SNESVIComputeJacobian(snes->jacobian,snes->jacobian_pre,vi->Da,vi->Db);CHKERRQ(ierr); 315 /* Compute the merit function gradient */ 316 ierr = SNESVIComputeMeritFunctionGradient(snes->jacobian,vi->phi,vi->dpsi);CHKERRQ(ierr); 317 ierr = KSPSetOperators(snes->ksp,snes->jacobian,snes->jacobian_pre,flg);CHKERRQ(ierr); 318 ierr = KSPSolve(snes->ksp,vi->phi,Y);CHKERRQ(ierr); 319 ierr = KSPGetConvergedReason(snes->ksp,&kspreason);CHKERRQ(ierr); 320 321 if (kspreason < 0) { 322 if (++snes->numLinearSolveFailures >= snes->maxLinearSolveFailures) { 323 ierr = PetscInfo2(snes,"iter=%D, number linear solve failures %D greater than current SNES allowed, stopping solve\n",snes->iter,snes->numLinearSolveFailures);CHKERRQ(ierr); 324 snes->reason = SNES_DIVERGED_LINEAR_SOLVE; 325 break; 326 } 327 } 328 ierr = KSPGetIterationNumber(snes->ksp,&lits);CHKERRQ(ierr); 329 snes->linear_its += lits; 330 ierr = PetscInfo2(snes,"iter=%D, linear solve iterations=%D\n",snes->iter,lits);CHKERRQ(ierr); 331 /* 332 if (snes->ops->precheck) { 333 PetscBool changed_y = PETSC_FALSE; 334 ierr = (*snes->ops->precheck)(snes,X,Y,snes->precheck,&changed_y);CHKERRQ(ierr); 335 } 336 337 if (PetscLogPrintInfo) { 338 ierr = SNESVICheckResidual_Private(snes,snes->jacobian,F,Y,G,W);CHKERRQ(ierr); 339 } 340 */ 341 /* Compute a (scaled) negative update in the line search routine: 342 Y <- X - lambda*Y 343 and evaluate G = function(Y) (depends on the line search). 344 */ 345 ierr = VecCopy(Y,snes->vec_sol_update);CHKERRQ(ierr); 346 ynorm = 1; gnorm = vi->phinorm; 347 ierr = SNESLineSearchApply(snes->linesearch, X, vi->phi, &gnorm, Y);CHKERRQ(ierr); 348 ierr = SNESLineSearchGetSuccess(snes->linesearch, &lssucceed);CHKERRQ(ierr); 349 ierr = SNESLineSearchGetNorms(snes->linesearch, &xnorm, &gnorm, &ynorm);CHKERRQ(ierr); 350 ierr = PetscInfo4(snes,"fnorm=%18.16e, gnorm=%18.16e, ynorm=%18.16e, lssucceed=%d\n",(double)vi->phinorm,(double)gnorm,(double)ynorm,(int)lssucceed);CHKERRQ(ierr); 351 if (snes->reason == SNES_DIVERGED_FUNCTION_COUNT) break; 352 if (snes->domainerror) { 353 snes->reason = SNES_DIVERGED_FUNCTION_DOMAIN; 354 sdm->ops->computefunction = vi->computeuserfunction; 355 PetscFunctionReturn(0); 356 } 357 if (!lssucceed) { 358 if (++snes->numFailures >= snes->maxFailures) { 359 PetscBool ismin; 360 snes->reason = SNES_DIVERGED_LINE_SEARCH; 361 ierr = SNESVICheckLocalMin_Private(snes,snes->jacobian,vi->phi,X,gnorm,&ismin);CHKERRQ(ierr); 362 if (ismin) snes->reason = SNES_DIVERGED_LOCAL_MIN; 363 break; 364 } 365 } 366 /* Update function and solution vectors */ 367 vi->phinorm = gnorm; 368 vi->merit = 0.5*vi->phinorm*vi->phinorm; 369 /* Monitor convergence */ 370 ierr = PetscObjectSAWsTakeAccess((PetscObject)snes);CHKERRQ(ierr); 371 snes->iter = i+1; 372 snes->norm = vi->phinorm; 373 ierr = PetscObjectSAWsGrantAccess((PetscObject)snes);CHKERRQ(ierr); 374 ierr = SNESLogConvergenceHistory(snes,snes->norm,lits);CHKERRQ(ierr); 375 ierr = SNESMonitor(snes,snes->iter,snes->norm);CHKERRQ(ierr); 376 /* Test for convergence, xnorm = || X || */ 377 if (snes->ops->converged != SNESConvergedSkip) { ierr = VecNorm(X,NORM_2,&xnorm);CHKERRQ(ierr); } 378 ierr = (*snes->ops->converged)(snes,snes->iter,xnorm,ynorm,vi->phinorm,&snes->reason,snes->cnvP);CHKERRQ(ierr); 379 if (snes->reason) break; 380 } 381 if (i == maxits) { 382 ierr = PetscInfo1(snes,"Maximum number of iterations has been reached: %D\n",maxits);CHKERRQ(ierr); 383 if (!snes->reason) snes->reason = SNES_DIVERGED_MAX_IT; 384 } 385 sdm->ops->computefunction = vi->computeuserfunction; 386 PetscFunctionReturn(0); 387 } 388 389 /* -------------------------------------------------------------------------- */ 390 /* 391 SNESSetUp_VINEWTONSSLS - Sets up the internal data structures for the later use 392 of the SNES nonlinear solver. 393 394 Input Parameter: 395 . snes - the SNES context 396 . x - the solution vector 397 398 Application Interface Routine: SNESSetUp() 399 400 Notes: 401 For basic use of the SNES solvers, the user need not explicitly call 402 SNESSetUp(), since these actions will automatically occur during 403 the call to SNESSolve(). 404 */ 405 #undef __FUNCT__ 406 #define __FUNCT__ "SNESSetUp_VINEWTONSSLS" 407 PetscErrorCode SNESSetUp_VINEWTONSSLS(SNES snes) 408 { 409 PetscErrorCode ierr; 410 SNES_VINEWTONSSLS *vi = (SNES_VINEWTONSSLS*) snes->data; 411 412 PetscFunctionBegin; 413 ierr = SNESSetUp_VI(snes);CHKERRQ(ierr); 414 ierr = VecDuplicate(snes->vec_sol, &vi->dpsi);CHKERRQ(ierr); 415 ierr = VecDuplicate(snes->vec_sol, &vi->phi);CHKERRQ(ierr); 416 ierr = VecDuplicate(snes->vec_sol, &vi->Da);CHKERRQ(ierr); 417 ierr = VecDuplicate(snes->vec_sol, &vi->Db);CHKERRQ(ierr); 418 ierr = VecDuplicate(snes->vec_sol, &vi->z);CHKERRQ(ierr); 419 ierr = VecDuplicate(snes->vec_sol, &vi->t);CHKERRQ(ierr); 420 PetscFunctionReturn(0); 421 } 422 /* -------------------------------------------------------------------------- */ 423 #undef __FUNCT__ 424 #define __FUNCT__ "SNESReset_VINEWTONSSLS" 425 PetscErrorCode SNESReset_VINEWTONSSLS(SNES snes) 426 { 427 SNES_VINEWTONSSLS *vi = (SNES_VINEWTONSSLS*) snes->data; 428 PetscErrorCode ierr; 429 430 PetscFunctionBegin; 431 ierr = SNESReset_VI(snes);CHKERRQ(ierr); 432 ierr = VecDestroy(&vi->dpsi);CHKERRQ(ierr); 433 ierr = VecDestroy(&vi->phi);CHKERRQ(ierr); 434 ierr = VecDestroy(&vi->Da);CHKERRQ(ierr); 435 ierr = VecDestroy(&vi->Db);CHKERRQ(ierr); 436 ierr = VecDestroy(&vi->z);CHKERRQ(ierr); 437 ierr = VecDestroy(&vi->t);CHKERRQ(ierr); 438 PetscFunctionReturn(0); 439 } 440 441 /* -------------------------------------------------------------------------- */ 442 /* 443 SNESSetFromOptions_VINEWTONSSLS - Sets various parameters for the SNESVI method. 444 445 Input Parameter: 446 . snes - the SNES context 447 448 Application Interface Routine: SNESSetFromOptions() 449 */ 450 #undef __FUNCT__ 451 #define __FUNCT__ "SNESSetFromOptions_VINEWTONSSLS" 452 static PetscErrorCode SNESSetFromOptions_VINEWTONSSLS(SNES snes) 453 { 454 PetscErrorCode ierr; 455 SNESLineSearch linesearch; 456 457 PetscFunctionBegin; 458 ierr = SNESSetFromOptions_VI(snes);CHKERRQ(ierr); 459 ierr = PetscOptionsHead("SNES semismooth method options");CHKERRQ(ierr); 460 ierr = PetscOptionsTail();CHKERRQ(ierr); 461 /* set up the default line search */ 462 if (!snes->linesearch) { 463 ierr = SNESGetLineSearch(snes, &linesearch);CHKERRQ(ierr); 464 ierr = SNESLineSearchSetType(linesearch, SNESLINESEARCHBT);CHKERRQ(ierr); 465 ierr = SNESLineSearchBTSetAlpha(linesearch, 0.0);CHKERRQ(ierr); 466 } 467 PetscFunctionReturn(0); 468 } 469 470 471 /* -------------------------------------------------------------------------- */ 472 /*MC 473 SNESVINEWTONSSLS - Semi-smooth solver for variational inequalities based on Newton's method 474 475 Options Database: 476 + -snes_vi_type <ss,rs,rsaug> a semi-smooth solver, a reduced space active set method, and a reduced space active set method that does not eliminate the active constraints from the Jacobian instead augments the Jacobian with additional variables that enforce the constraints 477 - -snes_vi_monitor - prints the number of active constraints at each iteration. 478 479 Level: beginner 480 481 References: 482 - T. S. Munson, F. Facchinei, M. C. Ferris, A. Fischer, and C. Kanzow. The semismooth 483 algorithm for large scale complementarity problems. INFORMS Journal on Computing, 13 (2001). 484 485 .seealso: SNESVISetVariableBounds(), SNESVISetComputeVariableBounds(), SNESCreate(), SNES, SNESSetType(), SNESVINEWTONRSLS, SNESVINEWTONSSLS, SNESNEWTONTR, SNESLineSearchSet(), 486 SNESLineSearchSetPostCheck(), SNESLineSearchNo(), SNESLineSearchCubic(), SNESLineSearchQuadratic(), 487 SNESLineSearchSet(), SNESLineSearchNoNorms(), SNESLineSearchSetPreCheck(), SNESLineSearchSetParams(), SNESLineSearchGetParams() 488 489 M*/ 490 #undef __FUNCT__ 491 #define __FUNCT__ "SNESCreate_VINEWTONSSLS" 492 PETSC_EXTERN PetscErrorCode SNESCreate_VINEWTONSSLS(SNES snes) 493 { 494 PetscErrorCode ierr; 495 SNES_VINEWTONSSLS *vi; 496 497 PetscFunctionBegin; 498 snes->ops->reset = SNESReset_VINEWTONSSLS; 499 snes->ops->setup = SNESSetUp_VINEWTONSSLS; 500 snes->ops->solve = SNESSolve_VINEWTONSSLS; 501 snes->ops->destroy = SNESDestroy_VI; 502 snes->ops->setfromoptions = SNESSetFromOptions_VINEWTONSSLS; 503 snes->ops->view = NULL; 504 505 snes->usesksp = PETSC_TRUE; 506 snes->usespc = PETSC_FALSE; 507 508 ierr = PetscNewLog(snes,SNES_VINEWTONSSLS,&vi);CHKERRQ(ierr); 509 snes->data = (void*)vi; 510 511 ierr = PetscObjectComposeFunction((PetscObject)snes,"SNESVISetVariableBounds_C",SNESVISetVariableBounds_VI);CHKERRQ(ierr); 512 ierr = PetscObjectComposeFunction((PetscObject)snes,"SNESVISetComputeVariableBounds_C",SNESVISetComputeVariableBounds_VI);CHKERRQ(ierr); 513 PetscFunctionReturn(0); 514 } 515 516