Actual source code: taosolver.c
1: #include <petsc/private/taoimpl.h>
2: #include <petsc/private/snesimpl.h>
4: PetscBool TaoRegisterAllCalled = PETSC_FALSE;
5: PetscFunctionList TaoList = NULL;
7: PetscClassId TAO_CLASSID;
9: PetscLogEvent TAO_Solve;
10: PetscLogEvent TAO_ObjectiveEval;
11: PetscLogEvent TAO_GradientEval;
12: PetscLogEvent TAO_ObjGradEval;
13: PetscLogEvent TAO_HessianEval;
14: PetscLogEvent TAO_JacobianEval;
15: PetscLogEvent TAO_ConstraintsEval;
17: const char *TaoSubSetTypes[] = {"subvec", "mask", "matrixfree", "TaoSubSetType", "TAO_SUBSET_", NULL};
19: struct _n_TaoMonitorDrawCtx {
20: PetscViewer viewer;
21: PetscInt howoften; /* when > 0 uses iteration % howoften, when negative only final solution plotted */
22: };
24: static PetscErrorCode KSPPreSolve_TAOEW_Private(KSP ksp, Vec b, Vec x, Tao tao)
25: {
26: SNES snes_ewdummy = tao->snes_ewdummy;
28: PetscFunctionBegin;
29: if (!snes_ewdummy) PetscFunctionReturn(PETSC_SUCCESS);
30: /* populate snes_ewdummy struct values used in KSPPreSolve_SNESEW */
31: snes_ewdummy->vec_func = b;
32: snes_ewdummy->rtol = tao->gttol;
33: snes_ewdummy->iter = tao->niter;
34: PetscCall(VecNorm(b, NORM_2, &snes_ewdummy->norm));
35: PetscCall(KSPPreSolve_SNESEW(ksp, b, x, snes_ewdummy));
36: snes_ewdummy->vec_func = NULL;
37: PetscFunctionReturn(PETSC_SUCCESS);
38: }
40: static PetscErrorCode KSPPostSolve_TAOEW_Private(KSP ksp, Vec b, Vec x, Tao tao)
41: {
42: SNES snes_ewdummy = tao->snes_ewdummy;
44: PetscFunctionBegin;
45: if (!snes_ewdummy) PetscFunctionReturn(PETSC_SUCCESS);
46: PetscCall(KSPPostSolve_SNESEW(ksp, b, x, snes_ewdummy));
47: PetscFunctionReturn(PETSC_SUCCESS);
48: }
50: static PetscErrorCode TaoSetUpEW_Private(Tao tao)
51: {
52: SNESKSPEW *kctx;
53: const char *ewprefix;
55: PetscFunctionBegin;
56: if (!tao->ksp) PetscFunctionReturn(PETSC_SUCCESS);
57: if (tao->ksp_ewconv) {
58: if (!tao->snes_ewdummy) PetscCall(SNESCreate(PetscObjectComm((PetscObject)tao), &tao->snes_ewdummy));
59: tao->snes_ewdummy->ksp_ewconv = PETSC_TRUE;
60: PetscCall(KSPSetPreSolve(tao->ksp, (PetscErrorCode(*)(KSP, Vec, Vec, void *))KSPPreSolve_TAOEW_Private, tao));
61: PetscCall(KSPSetPostSolve(tao->ksp, (PetscErrorCode(*)(KSP, Vec, Vec, void *))KSPPostSolve_TAOEW_Private, tao));
63: PetscCall(KSPGetOptionsPrefix(tao->ksp, &ewprefix));
64: kctx = (SNESKSPEW *)tao->snes_ewdummy->kspconvctx;
65: PetscCall(SNESEWSetFromOptions_Private(kctx, PetscObjectComm((PetscObject)tao), ewprefix));
66: } else PetscCall(SNESDestroy(&tao->snes_ewdummy));
67: PetscFunctionReturn(PETSC_SUCCESS);
68: }
70: /*@
71: TaoCreate - Creates a Tao solver
73: Collective
75: Input Parameter:
76: . comm - MPI communicator
78: Output Parameter:
79: . newtao - the new `Tao` context
81: Options Database Key:
82: . -tao_type - select which method Tao should use
84: Level: beginner
86: .seealso: [](chapter_tao), `Tao`, `TaoSolve()`, `TaoDestroy()`, `TAOSetFromOptions()`, `TAOSetType()`
87: @*/
88: PetscErrorCode TaoCreate(MPI_Comm comm, Tao *newtao)
89: {
90: Tao tao;
92: PetscFunctionBegin;
94: PetscCall(TaoInitializePackage());
95: PetscCall(TaoLineSearchInitializePackage());
96: PetscCall(PetscHeaderCreate(tao, TAO_CLASSID, "Tao", "Optimization solver", "Tao", comm, TaoDestroy, TaoView));
98: /* Set non-NULL defaults */
99: tao->ops->convergencetest = TaoDefaultConvergenceTest;
101: tao->max_it = 10000;
102: tao->max_funcs = -1;
103: #if defined(PETSC_USE_REAL_SINGLE)
104: tao->gatol = 1e-5;
105: tao->grtol = 1e-5;
106: tao->crtol = 1e-5;
107: tao->catol = 1e-5;
108: #else
109: tao->gatol = 1e-8;
110: tao->grtol = 1e-8;
111: tao->crtol = 1e-8;
112: tao->catol = 1e-8;
113: #endif
114: tao->gttol = 0.0;
115: tao->steptol = 0.0;
116: tao->trust0 = PETSC_INFINITY;
117: tao->fmin = PETSC_NINFINITY;
119: tao->hist_reset = PETSC_TRUE;
121: PetscCall(TaoResetStatistics(tao));
122: *newtao = tao;
123: PetscFunctionReturn(PETSC_SUCCESS);
124: }
126: /*@
127: TaoSolve - Solves an optimization problem min F(x) s.t. l <= x <= u
129: Collective
131: Input Parameter:
132: . tao - the `Tao` context
134: Level: beginner
136: Notes:
137: The user must set up the `Tao` object with calls to `TaoSetSolution()`, `TaoSetObjective()`, `TaoSetGradient()`, and (if using 2nd order method) `TaoSetHessian()`.
139: You should call `TaoGetConvergedReason()` or run with `-tao_converged_reason` to determine if the optimization algorithm actually succeeded or
140: why it failed.
142: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoSetObjective()`, `TaoSetGradient()`, `TaoSetHessian()`, `TaoGetConvergedReason()`, `TaoSetUp()`
143: @*/
144: PetscErrorCode TaoSolve(Tao tao)
145: {
146: static PetscBool set = PETSC_FALSE;
148: PetscFunctionBegin;
150: PetscCall(PetscCitationsRegister("@TechReport{tao-user-ref,\n"
151: "title = {Toolkit for Advanced Optimization (TAO) Users Manual},\n"
152: "author = {Todd Munson and Jason Sarich and Stefan Wild and Steve Benson and Lois Curfman McInnes},\n"
153: "Institution = {Argonne National Laboratory},\n"
154: "Year = 2014,\n"
155: "Number = {ANL/MCS-TM-322 - Revision 3.5},\n"
156: "url = {https://www.mcs.anl.gov/research/projects/tao/}\n}\n",
157: &set));
158: tao->header_printed = PETSC_FALSE;
159: PetscCall(TaoSetUp(tao));
160: PetscCall(TaoResetStatistics(tao));
161: if (tao->linesearch) PetscCall(TaoLineSearchReset(tao->linesearch));
163: PetscCall(PetscLogEventBegin(TAO_Solve, tao, 0, 0, 0));
164: PetscTryTypeMethod(tao, solve);
165: PetscCall(PetscLogEventEnd(TAO_Solve, tao, 0, 0, 0));
167: PetscCall(VecViewFromOptions(tao->solution, (PetscObject)tao, "-tao_view_solution"));
169: tao->ntotalits += tao->niter;
171: if (tao->printreason) {
172: PetscViewer viewer = PETSC_VIEWER_STDOUT_(((PetscObject)tao)->comm);
173: PetscCall(PetscViewerASCIIAddTab(viewer, ((PetscObject)tao)->tablevel));
174: if (tao->reason > 0) {
175: PetscCall(PetscViewerASCIIPrintf(viewer, " TAO %s solve converged due to %s iterations %" PetscInt_FMT "\n", ((PetscObject)tao)->prefix ? ((PetscObject)tao)->prefix : "", TaoConvergedReasons[tao->reason], tao->niter));
176: } else {
177: PetscCall(PetscViewerASCIIPrintf(viewer, " TAO %s solve did not converge due to %s iteration %" PetscInt_FMT "\n", ((PetscObject)tao)->prefix ? ((PetscObject)tao)->prefix : "", TaoConvergedReasons[tao->reason], tao->niter));
178: }
179: PetscCall(PetscViewerASCIISubtractTab(viewer, ((PetscObject)tao)->tablevel));
180: }
181: PetscCall(TaoViewFromOptions(tao, NULL, "-tao_view"));
182: PetscFunctionReturn(PETSC_SUCCESS);
183: }
185: /*@
186: TaoSetUp - Sets up the internal data structures for the later use
187: of a Tao solver
189: Collective
191: Input Parameter:
192: . tao - the `Tao` context
194: Level: advanced
196: Note:
197: The user will not need to explicitly call `TaoSetUp()`, as it will
198: automatically be called in `TaoSolve()`. However, if the user
199: desires to call it explicitly, it should come after `TaoCreate()`
200: and any TaoSetSomething() routines, but before `TaoSolve()`.
202: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoSolve()`
203: @*/
204: PetscErrorCode TaoSetUp(Tao tao)
205: {
206: PetscFunctionBegin;
208: if (tao->setupcalled) PetscFunctionReturn(PETSC_SUCCESS);
209: PetscCall(TaoSetUpEW_Private(tao));
210: PetscCheck(tao->solution, PetscObjectComm((PetscObject)tao), PETSC_ERR_ARG_WRONGSTATE, "Must call TaoSetSolution");
211: PetscTryTypeMethod(tao, setup);
212: tao->setupcalled = PETSC_TRUE;
213: PetscFunctionReturn(PETSC_SUCCESS);
214: }
216: /*@C
217: TaoDestroy - Destroys the `Tao` context that was created with `TaoCreate()`
219: Collective
221: Input Parameter:
222: . tao - the `Tao` context
224: Level: beginner
226: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoSolve()`
227: @*/
228: PetscErrorCode TaoDestroy(Tao *tao)
229: {
230: PetscFunctionBegin;
231: if (!*tao) PetscFunctionReturn(PETSC_SUCCESS);
233: if (--((PetscObject)*tao)->refct > 0) {
234: *tao = NULL;
235: PetscFunctionReturn(PETSC_SUCCESS);
236: }
238: if ((*tao)->ops->destroy) PetscCall((*((*tao))->ops->destroy)(*tao));
239: PetscCall(KSPDestroy(&(*tao)->ksp));
240: PetscCall(SNESDestroy(&(*tao)->snes_ewdummy));
241: PetscCall(TaoLineSearchDestroy(&(*tao)->linesearch));
243: if ((*tao)->ops->convergencedestroy) {
244: PetscCall((*(*tao)->ops->convergencedestroy)((*tao)->cnvP));
245: if ((*tao)->jacobian_state_inv) PetscCall(MatDestroy(&(*tao)->jacobian_state_inv));
246: }
247: PetscCall(VecDestroy(&(*tao)->solution));
248: PetscCall(VecDestroy(&(*tao)->gradient));
249: PetscCall(VecDestroy(&(*tao)->ls_res));
251: if ((*tao)->gradient_norm) {
252: PetscCall(PetscObjectDereference((PetscObject)(*tao)->gradient_norm));
253: PetscCall(VecDestroy(&(*tao)->gradient_norm_tmp));
254: }
256: PetscCall(VecDestroy(&(*tao)->XL));
257: PetscCall(VecDestroy(&(*tao)->XU));
258: PetscCall(VecDestroy(&(*tao)->IL));
259: PetscCall(VecDestroy(&(*tao)->IU));
260: PetscCall(VecDestroy(&(*tao)->DE));
261: PetscCall(VecDestroy(&(*tao)->DI));
262: PetscCall(VecDestroy(&(*tao)->constraints));
263: PetscCall(VecDestroy(&(*tao)->constraints_equality));
264: PetscCall(VecDestroy(&(*tao)->constraints_inequality));
265: PetscCall(VecDestroy(&(*tao)->stepdirection));
266: PetscCall(MatDestroy(&(*tao)->hessian_pre));
267: PetscCall(MatDestroy(&(*tao)->hessian));
268: PetscCall(MatDestroy(&(*tao)->ls_jac));
269: PetscCall(MatDestroy(&(*tao)->ls_jac_pre));
270: PetscCall(MatDestroy(&(*tao)->jacobian_pre));
271: PetscCall(MatDestroy(&(*tao)->jacobian));
272: PetscCall(MatDestroy(&(*tao)->jacobian_state_pre));
273: PetscCall(MatDestroy(&(*tao)->jacobian_state));
274: PetscCall(MatDestroy(&(*tao)->jacobian_state_inv));
275: PetscCall(MatDestroy(&(*tao)->jacobian_design));
276: PetscCall(MatDestroy(&(*tao)->jacobian_equality));
277: PetscCall(MatDestroy(&(*tao)->jacobian_equality_pre));
278: PetscCall(MatDestroy(&(*tao)->jacobian_inequality));
279: PetscCall(MatDestroy(&(*tao)->jacobian_inequality_pre));
280: PetscCall(ISDestroy(&(*tao)->state_is));
281: PetscCall(ISDestroy(&(*tao)->design_is));
282: PetscCall(VecDestroy(&(*tao)->res_weights_v));
283: PetscCall(TaoCancelMonitors(*tao));
284: if ((*tao)->hist_malloc) PetscCall(PetscFree4((*tao)->hist_obj, (*tao)->hist_resid, (*tao)->hist_cnorm, (*tao)->hist_lits));
285: if ((*tao)->res_weights_n) {
286: PetscCall(PetscFree((*tao)->res_weights_rows));
287: PetscCall(PetscFree((*tao)->res_weights_cols));
288: PetscCall(PetscFree((*tao)->res_weights_w));
289: }
290: PetscCall(PetscHeaderDestroy(tao));
291: PetscFunctionReturn(PETSC_SUCCESS);
292: }
294: /*@
295: TaoKSPSetUseEW - Sets `SNES` use Eisenstat-Walker method for computing relative tolerance for linear solvers.
297: Logically Collective
299: Input Parameters:
300: + tao - Tao context
301: - flag - `PETSC_TRUE` or `PETSC_FALSE`
303: Level: advanced
305: Note:
306: See `SNESKSPSetUseEW()` for customization details.
308: Reference:
309: . * - S. C. Eisenstat and H. F. Walker, "Choosing the forcing terms in an inexact Newton method", SISC 17 (1), pp.16-32, 1996.
311: .seealso: [](chapter_tao), `Tao`, `SNESKSPSetUseEW()`
312: @*/
313: PetscErrorCode TaoKSPSetUseEW(Tao tao, PetscBool flag)
314: {
315: PetscFunctionBegin;
318: tao->ksp_ewconv = flag;
319: PetscFunctionReturn(PETSC_SUCCESS);
320: }
322: /*@
323: TaoSetFromOptions - Sets various Tao parameters from the options database
325: Collective
327: Input Parameter:
328: . tao - the `Tao` solver context
330: Options Database Keys:
331: + -tao_type <type> - The algorithm that Tao uses (lmvm, nls, etc.)
332: . -tao_gatol <gatol> - absolute error tolerance for ||gradient||
333: . -tao_grtol <grtol> - relative error tolerance for ||gradient||
334: . -tao_gttol <gttol> - reduction of ||gradient|| relative to initial gradient
335: . -tao_max_it <max> - sets maximum number of iterations
336: . -tao_max_funcs <max> - sets maximum number of function evaluations
337: . -tao_fmin <fmin> - stop if function value reaches fmin
338: . -tao_steptol <tol> - stop if trust region radius less than <tol>
339: . -tao_trust0 <t> - initial trust region radius
340: . -tao_monitor - prints function value and residual norm at each iteration
341: . -tao_smonitor - same as tao_monitor, but truncates very small values
342: . -tao_cmonitor - prints function value, residual, and constraint norm at each iteration
343: . -tao_view_solution - prints solution vector at each iteration
344: . -tao_view_ls_residual - prints least-squares residual vector at each iteration
345: . -tao_view_stepdirection - prints step direction vector at each iteration
346: . -tao_view_gradient - prints gradient vector at each iteration
347: . -tao_draw_solution - graphically view solution vector at each iteration
348: . -tao_draw_step - graphically view step vector at each iteration
349: . -tao_draw_gradient - graphically view gradient at each iteration
350: . -tao_fd_gradient - use gradient computed with finite differences
351: . -tao_fd_hessian - use hessian computed with finite differences
352: . -tao_mf_hessian - use matrix-free hessian computed with finite differences
353: . -tao_cancelmonitors - cancels all monitors (except those set with command line)
354: . -tao_view - prints information about the Tao after solving
355: - -tao_converged_reason - prints the reason Tao stopped iterating
357: Level: beginner
359: Note:
360: To see all options, run your program with the `-help` option or consult the
361: user's manual. Should be called after `TaoCreate()` but before `TaoSolve()`
363: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoSolve()`
364: @*/
365: PetscErrorCode TaoSetFromOptions(Tao tao)
366: {
367: TaoType default_type = TAOLMVM;
368: char type[256], monfilename[PETSC_MAX_PATH_LEN];
369: PetscViewer monviewer;
370: PetscBool flg;
371: MPI_Comm comm;
373: PetscFunctionBegin;
375: PetscCall(PetscObjectGetComm((PetscObject)tao, &comm));
377: if (((PetscObject)tao)->type_name) default_type = ((PetscObject)tao)->type_name;
379: PetscObjectOptionsBegin((PetscObject)tao);
380: /* Check for type from options */
381: PetscCall(PetscOptionsFList("-tao_type", "Tao Solver type", "TaoSetType", TaoList, default_type, type, 256, &flg));
382: if (flg) {
383: PetscCall(TaoSetType(tao, type));
384: } else if (!((PetscObject)tao)->type_name) {
385: PetscCall(TaoSetType(tao, default_type));
386: }
388: /* Tao solvers do not set the prefix, set it here if not yet done
389: We do it after SetType since solver may have been changed */
390: if (tao->linesearch) {
391: const char *prefix;
392: PetscCall(TaoLineSearchGetOptionsPrefix(tao->linesearch, &prefix));
393: if (!prefix) PetscCall(TaoLineSearchSetOptionsPrefix(tao->linesearch, ((PetscObject)(tao))->prefix));
394: }
396: PetscCall(PetscOptionsReal("-tao_catol", "Stop if constraints violations within", "TaoSetConstraintTolerances", tao->catol, &tao->catol, &flg));
397: if (flg) tao->catol_changed = PETSC_TRUE;
398: PetscCall(PetscOptionsReal("-tao_crtol", "Stop if relative constraint violations within", "TaoSetConstraintTolerances", tao->crtol, &tao->crtol, &flg));
399: if (flg) tao->crtol_changed = PETSC_TRUE;
400: PetscCall(PetscOptionsReal("-tao_gatol", "Stop if norm of gradient less than", "TaoSetTolerances", tao->gatol, &tao->gatol, &flg));
401: if (flg) tao->gatol_changed = PETSC_TRUE;
402: PetscCall(PetscOptionsReal("-tao_grtol", "Stop if norm of gradient divided by the function value is less than", "TaoSetTolerances", tao->grtol, &tao->grtol, &flg));
403: if (flg) tao->grtol_changed = PETSC_TRUE;
404: PetscCall(PetscOptionsReal("-tao_gttol", "Stop if the norm of the gradient is less than the norm of the initial gradient times tol", "TaoSetTolerances", tao->gttol, &tao->gttol, &flg));
405: if (flg) tao->gttol_changed = PETSC_TRUE;
406: PetscCall(PetscOptionsInt("-tao_max_it", "Stop if iteration number exceeds", "TaoSetMaximumIterations", tao->max_it, &tao->max_it, &flg));
407: if (flg) tao->max_it_changed = PETSC_TRUE;
408: PetscCall(PetscOptionsInt("-tao_max_funcs", "Stop if number of function evaluations exceeds", "TaoSetMaximumFunctionEvaluations", tao->max_funcs, &tao->max_funcs, &flg));
409: if (flg) tao->max_funcs_changed = PETSC_TRUE;
410: PetscCall(PetscOptionsReal("-tao_fmin", "Stop if function less than", "TaoSetFunctionLowerBound", tao->fmin, &tao->fmin, &flg));
411: if (flg) tao->fmin_changed = PETSC_TRUE;
412: PetscCall(PetscOptionsReal("-tao_steptol", "Stop if step size or trust region radius less than", "", tao->steptol, &tao->steptol, &flg));
413: if (flg) tao->steptol_changed = PETSC_TRUE;
414: PetscCall(PetscOptionsReal("-tao_trust0", "Initial trust region radius", "TaoSetTrustRegionRadius", tao->trust0, &tao->trust0, &flg));
415: if (flg) tao->trust0_changed = PETSC_TRUE;
416: PetscCall(PetscOptionsString("-tao_view_solution", "view solution vector after each evaluation", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
417: if (flg) {
418: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
419: PetscCall(TaoSetMonitor(tao, TaoSolutionMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
420: }
422: PetscCall(PetscOptionsBool("-tao_converged_reason", "Print reason for Tao converged", "TaoSolve", tao->printreason, &tao->printreason, NULL));
423: PetscCall(PetscOptionsString("-tao_view_gradient", "view gradient vector after each evaluation", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
424: if (flg) {
425: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
426: PetscCall(TaoSetMonitor(tao, TaoGradientMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
427: }
429: PetscCall(PetscOptionsString("-tao_view_stepdirection", "view step direction vector after each iteration", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
430: if (flg) {
431: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
432: PetscCall(TaoSetMonitor(tao, TaoStepDirectionMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
433: }
435: PetscCall(PetscOptionsString("-tao_view_residual", "view least-squares residual vector after each evaluation", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
436: if (flg) {
437: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
438: PetscCall(TaoSetMonitor(tao, TaoResidualMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
439: }
441: PetscCall(PetscOptionsString("-tao_monitor", "Use the default convergence monitor", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
442: if (flg) {
443: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
444: PetscCall(TaoSetMonitor(tao, TaoMonitorDefault, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
445: }
447: PetscCall(PetscOptionsString("-tao_gmonitor", "Use the convergence monitor with extra globalization info", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
448: if (flg) {
449: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
450: PetscCall(TaoSetMonitor(tao, TaoDefaultGMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
451: }
453: PetscCall(PetscOptionsString("-tao_smonitor", "Use the short convergence monitor", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
454: if (flg) {
455: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
456: PetscCall(TaoSetMonitor(tao, TaoDefaultSMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
457: }
459: PetscCall(PetscOptionsString("-tao_cmonitor", "Use the default convergence monitor with constraint norm", "TaoSetMonitor", "stdout", monfilename, sizeof(monfilename), &flg));
460: if (flg) {
461: PetscCall(PetscViewerASCIIOpen(comm, monfilename, &monviewer));
462: PetscCall(TaoSetMonitor(tao, TaoDefaultCMonitor, monviewer, (PetscErrorCode(*)(void **))PetscViewerDestroy));
463: }
465: flg = PETSC_FALSE;
466: PetscCall(PetscOptionsBool("-tao_cancelmonitors", "cancel all monitors and call any registered destroy routines", "TaoCancelMonitors", flg, &flg, NULL));
467: if (flg) PetscCall(TaoCancelMonitors(tao));
469: flg = PETSC_FALSE;
470: PetscCall(PetscOptionsBool("-tao_draw_solution", "Plot solution vector at each iteration", "TaoSetMonitor", flg, &flg, NULL));
471: if (flg) {
472: TaoMonitorDrawCtx drawctx;
473: PetscInt howoften = 1;
474: PetscCall(TaoMonitorDrawCtxCreate(PetscObjectComm((PetscObject)tao), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &drawctx));
475: PetscCall(TaoSetMonitor(tao, TaoDrawSolutionMonitor, drawctx, (PetscErrorCode(*)(void **))TaoMonitorDrawCtxDestroy));
476: }
478: flg = PETSC_FALSE;
479: PetscCall(PetscOptionsBool("-tao_draw_step", "plots step direction at each iteration", "TaoSetMonitor", flg, &flg, NULL));
480: if (flg) PetscCall(TaoSetMonitor(tao, TaoDrawStepMonitor, NULL, NULL));
482: flg = PETSC_FALSE;
483: PetscCall(PetscOptionsBool("-tao_draw_gradient", "plots gradient at each iteration", "TaoSetMonitor", flg, &flg, NULL));
484: if (flg) {
485: TaoMonitorDrawCtx drawctx;
486: PetscInt howoften = 1;
487: PetscCall(TaoMonitorDrawCtxCreate(PetscObjectComm((PetscObject)tao), NULL, NULL, PETSC_DECIDE, PETSC_DECIDE, 300, 300, howoften, &drawctx));
488: PetscCall(TaoSetMonitor(tao, TaoDrawGradientMonitor, drawctx, (PetscErrorCode(*)(void **))TaoMonitorDrawCtxDestroy));
489: }
490: flg = PETSC_FALSE;
491: PetscCall(PetscOptionsBool("-tao_fd_gradient", "compute gradient using finite differences", "TaoDefaultComputeGradient", flg, &flg, NULL));
492: if (flg) PetscCall(TaoSetGradient(tao, NULL, TaoDefaultComputeGradient, NULL));
493: flg = PETSC_FALSE;
494: PetscCall(PetscOptionsBool("-tao_fd_hessian", "compute hessian using finite differences", "TaoDefaultComputeHessian", flg, &flg, NULL));
495: if (flg) {
496: Mat H;
498: PetscCall(MatCreate(PetscObjectComm((PetscObject)tao), &H));
499: PetscCall(MatSetType(H, MATAIJ));
500: PetscCall(TaoSetHessian(tao, H, H, TaoDefaultComputeHessian, NULL));
501: PetscCall(MatDestroy(&H));
502: }
503: flg = PETSC_FALSE;
504: PetscCall(PetscOptionsBool("-tao_mf_hessian", "compute matrix-free hessian using finite differences", "TaoDefaultComputeHessianMFFD", flg, &flg, NULL));
505: if (flg) {
506: Mat H;
508: PetscCall(MatCreate(PetscObjectComm((PetscObject)tao), &H));
509: PetscCall(TaoSetHessian(tao, H, H, TaoDefaultComputeHessianMFFD, NULL));
510: PetscCall(MatDestroy(&H));
511: }
512: flg = PETSC_FALSE;
513: PetscCall(PetscOptionsBool("-tao_recycle_history", "enable recycling/re-using information from the previous TaoSolve() call for some algorithms", "TaoSetRecycleHistory", flg, &flg, NULL));
514: if (flg) PetscCall(TaoSetRecycleHistory(tao, PETSC_TRUE));
515: PetscCall(PetscOptionsEnum("-tao_subset_type", "subset type", "", TaoSubSetTypes, (PetscEnum)tao->subset_type, (PetscEnum *)&tao->subset_type, NULL));
517: if (tao->ksp) {
518: PetscCall(PetscOptionsBool("-tao_ksp_ew", "Use Eisentat-Walker linear system convergence test", "TaoKSPSetUseEW", tao->ksp_ewconv, &tao->ksp_ewconv, NULL));
519: PetscCall(TaoKSPSetUseEW(tao, tao->ksp_ewconv));
520: }
522: PetscTryTypeMethod(tao, setfromoptions, PetscOptionsObject);
524: /* process any options handlers added with PetscObjectAddOptionsHandler() */
525: PetscCall(PetscObjectProcessOptionsHandlers((PetscObject)tao, PetscOptionsObject));
526: PetscOptionsEnd();
528: if (tao->linesearch) PetscCall(TaoLineSearchSetFromOptions(tao->linesearch));
529: PetscFunctionReturn(PETSC_SUCCESS);
530: }
532: /*@C
533: TaoViewFromOptions - View a `Tao` object based on values in the options database
535: Collective
537: Input Parameters:
538: + A - the `Tao` context
539: . obj - Optional object that provides the prefix for the options database
540: - name - command line option
542: Level: intermediate
544: .seealso: [](chapter_tao), `Tao`, `TaoView`, `PetscObjectViewFromOptions()`, `TaoCreate()`
545: @*/
546: PetscErrorCode TaoViewFromOptions(Tao A, PetscObject obj, const char name[])
547: {
548: PetscFunctionBegin;
550: PetscCall(PetscObjectViewFromOptions((PetscObject)A, obj, name));
551: PetscFunctionReturn(PETSC_SUCCESS);
552: }
554: /*@C
555: TaoView - Prints information about the `Tao` object
557: Collective
559: InputParameters:
560: + tao - the `Tao` context
561: - viewer - visualization context
563: Options Database Key:
564: . -tao_view - Calls `TaoView()` at the end of `TaoSolve()`
566: Level: beginner
568: Notes:
569: The available visualization contexts include
570: + `PETSC_VIEWER_STDOUT_SELF` - standard output (default)
571: - `PETSC_VIEWER_STDOUT_WORLD` - synchronized standard
572: output where only the first processor opens
573: the file. All other processors send their
574: data to the first processor to print.
576: .seealso: [](chapter_tao), `Tao`, `PetscViewerASCIIOpen()`
577: @*/
578: PetscErrorCode TaoView(Tao tao, PetscViewer viewer)
579: {
580: PetscBool isascii, isstring;
581: TaoType type;
583: PetscFunctionBegin;
585: if (!viewer) PetscCall(PetscViewerASCIIGetStdout(((PetscObject)tao)->comm, &viewer));
587: PetscCheckSameComm(tao, 1, viewer, 2);
589: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &isascii));
590: PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERSTRING, &isstring));
591: if (isascii) {
592: PetscCall(PetscObjectPrintClassNamePrefixType((PetscObject)tao, viewer));
594: if (tao->ops->view) {
595: PetscCall(PetscViewerASCIIPushTab(viewer));
596: PetscUseTypeMethod(tao, view, viewer);
597: PetscCall(PetscViewerASCIIPopTab(viewer));
598: }
599: if (tao->linesearch) {
600: PetscCall(PetscViewerASCIIPushTab(viewer));
601: PetscCall(TaoLineSearchView(tao->linesearch, viewer));
602: PetscCall(PetscViewerASCIIPopTab(viewer));
603: }
604: if (tao->ksp) {
605: PetscCall(PetscViewerASCIIPushTab(viewer));
606: PetscCall(KSPView(tao->ksp, viewer));
607: PetscCall(PetscViewerASCIIPrintf(viewer, "total KSP iterations: %" PetscInt_FMT "\n", tao->ksp_tot_its));
608: PetscCall(PetscViewerASCIIPopTab(viewer));
609: }
611: PetscCall(PetscViewerASCIIPushTab(viewer));
613: if (tao->XL || tao->XU) PetscCall(PetscViewerASCIIPrintf(viewer, "Active Set subset type: %s\n", TaoSubSetTypes[tao->subset_type]));
615: PetscCall(PetscViewerASCIIPrintf(viewer, "convergence tolerances: gatol=%g,", (double)tao->gatol));
616: PetscCall(PetscViewerASCIIPrintf(viewer, " steptol=%g,", (double)tao->steptol));
617: PetscCall(PetscViewerASCIIPrintf(viewer, " gttol=%g\n", (double)tao->gttol));
618: PetscCall(PetscViewerASCIIPrintf(viewer, "Residual in Function/Gradient:=%g\n", (double)tao->residual));
620: if (tao->constrained) {
621: PetscCall(PetscViewerASCIIPrintf(viewer, "convergence tolerances:"));
622: PetscCall(PetscViewerASCIIPrintf(viewer, " catol=%g,", (double)tao->catol));
623: PetscCall(PetscViewerASCIIPrintf(viewer, " crtol=%g\n", (double)tao->crtol));
624: PetscCall(PetscViewerASCIIPrintf(viewer, "Residual in Constraints:=%g\n", (double)tao->cnorm));
625: }
627: if (tao->trust < tao->steptol) {
628: PetscCall(PetscViewerASCIIPrintf(viewer, "convergence tolerances: steptol=%g\n", (double)tao->steptol));
629: PetscCall(PetscViewerASCIIPrintf(viewer, "Final trust region radius:=%g\n", (double)tao->trust));
630: }
632: if (tao->fmin > -1.e25) PetscCall(PetscViewerASCIIPrintf(viewer, "convergence tolerances: function minimum=%g\n", (double)tao->fmin));
633: PetscCall(PetscViewerASCIIPrintf(viewer, "Objective value=%g\n", (double)tao->fc));
635: PetscCall(PetscViewerASCIIPrintf(viewer, "total number of iterations=%" PetscInt_FMT ", ", tao->niter));
636: PetscCall(PetscViewerASCIIPrintf(viewer, " (max: %" PetscInt_FMT ")\n", tao->max_it));
638: if (tao->nfuncs > 0) {
639: PetscCall(PetscViewerASCIIPrintf(viewer, "total number of function evaluations=%" PetscInt_FMT ",", tao->nfuncs));
640: PetscCall(PetscViewerASCIIPrintf(viewer, " max: %" PetscInt_FMT "\n", tao->max_funcs));
641: }
642: if (tao->ngrads > 0) {
643: PetscCall(PetscViewerASCIIPrintf(viewer, "total number of gradient evaluations=%" PetscInt_FMT ",", tao->ngrads));
644: PetscCall(PetscViewerASCIIPrintf(viewer, " max: %" PetscInt_FMT "\n", tao->max_funcs));
645: }
646: if (tao->nfuncgrads > 0) {
647: PetscCall(PetscViewerASCIIPrintf(viewer, "total number of function/gradient evaluations=%" PetscInt_FMT ",", tao->nfuncgrads));
648: PetscCall(PetscViewerASCIIPrintf(viewer, " (max: %" PetscInt_FMT ")\n", tao->max_funcs));
649: }
650: if (tao->nhess > 0) PetscCall(PetscViewerASCIIPrintf(viewer, "total number of Hessian evaluations=%" PetscInt_FMT "\n", tao->nhess));
651: if (tao->nconstraints > 0) PetscCall(PetscViewerASCIIPrintf(viewer, "total number of constraint function evaluations=%" PetscInt_FMT "\n", tao->nconstraints));
652: if (tao->njac > 0) PetscCall(PetscViewerASCIIPrintf(viewer, "total number of Jacobian evaluations=%" PetscInt_FMT "\n", tao->njac));
654: if (tao->reason > 0) {
655: PetscCall(PetscViewerASCIIPrintf(viewer, "Solution converged: "));
656: switch (tao->reason) {
657: case TAO_CONVERGED_GATOL:
658: PetscCall(PetscViewerASCIIPrintf(viewer, " ||g(X)|| <= gatol\n"));
659: break;
660: case TAO_CONVERGED_GRTOL:
661: PetscCall(PetscViewerASCIIPrintf(viewer, " ||g(X)||/|f(X)| <= grtol\n"));
662: break;
663: case TAO_CONVERGED_GTTOL:
664: PetscCall(PetscViewerASCIIPrintf(viewer, " ||g(X)||/||g(X0)|| <= gttol\n"));
665: break;
666: case TAO_CONVERGED_STEPTOL:
667: PetscCall(PetscViewerASCIIPrintf(viewer, " Steptol -- step size small\n"));
668: break;
669: case TAO_CONVERGED_MINF:
670: PetscCall(PetscViewerASCIIPrintf(viewer, " Minf -- f < fmin\n"));
671: break;
672: case TAO_CONVERGED_USER:
673: PetscCall(PetscViewerASCIIPrintf(viewer, " User Terminated\n"));
674: break;
675: default:
676: PetscCall(PetscViewerASCIIPrintf(viewer, "\n"));
677: break;
678: }
679: } else {
680: PetscCall(PetscViewerASCIIPrintf(viewer, "Solver terminated: %d", tao->reason));
681: switch (tao->reason) {
682: case TAO_DIVERGED_MAXITS:
683: PetscCall(PetscViewerASCIIPrintf(viewer, " Maximum Iterations\n"));
684: break;
685: case TAO_DIVERGED_NAN:
686: PetscCall(PetscViewerASCIIPrintf(viewer, " NAN or Inf encountered\n"));
687: break;
688: case TAO_DIVERGED_MAXFCN:
689: PetscCall(PetscViewerASCIIPrintf(viewer, " Maximum Function Evaluations\n"));
690: break;
691: case TAO_DIVERGED_LS_FAILURE:
692: PetscCall(PetscViewerASCIIPrintf(viewer, " Line Search Failure\n"));
693: break;
694: case TAO_DIVERGED_TR_REDUCTION:
695: PetscCall(PetscViewerASCIIPrintf(viewer, " Trust Region too small\n"));
696: break;
697: case TAO_DIVERGED_USER:
698: PetscCall(PetscViewerASCIIPrintf(viewer, " User Terminated\n"));
699: break;
700: default:
701: PetscCall(PetscViewerASCIIPrintf(viewer, "\n"));
702: break;
703: }
704: }
705: PetscCall(PetscViewerASCIIPopTab(viewer));
706: } else if (isstring) {
707: PetscCall(TaoGetType(tao, &type));
708: PetscCall(PetscViewerStringSPrintf(viewer, " %-3.3s", type));
709: }
710: PetscFunctionReturn(PETSC_SUCCESS);
711: }
713: /*@
714: TaoSetRecycleHistory - Sets the boolean flag to enable/disable re-using
715: iterate information from the previous `TaoSolve()`. This feature is disabled by
716: default.
718: Logically Collective
720: Input Parameters:
721: + tao - the `Tao` context
722: - recycle - boolean flag
724: Options Database Key:
725: . -tao_recycle_history <true,false> - reuse the history
727: Level: intermediate
729: Notes:
730: For conjugate gradient methods (`TAOBNCG`), this re-uses the latest search direction
731: from the previous `TaoSolve()` call when computing the first search direction in a
732: new solution. By default, CG methods set the first search direction to the
733: negative gradient.
735: For quasi-Newton family of methods (`TAOBQNLS`, `TAOBQNKLS`, `TAOBQNKTR`, `TAOBQNKTL`), this re-uses
736: the accumulated quasi-Newton Hessian approximation from the previous `TaoSolve()`
737: call. By default, QN family of methods reset the initial Hessian approximation to
738: the identity matrix.
740: For any other algorithm, this setting has no effect.
742: .seealso: [](chapter_tao), `Tao`, `TaoGetRecycleHistory()`, `TAOBNCG`, `TAOBQNLS`, `TAOBQNKLS`, `TAOBQNKTR`, `TAOBQNKTL`
743: @*/
744: PetscErrorCode TaoSetRecycleHistory(Tao tao, PetscBool recycle)
745: {
746: PetscFunctionBegin;
749: tao->recycle = recycle;
750: PetscFunctionReturn(PETSC_SUCCESS);
751: }
753: /*@
754: TaoGetRecycleHistory - Retrieve the boolean flag for re-using iterate information
755: from the previous `TaoSolve()`. This feature is disabled by default.
757: Logically Collective
759: Input Parameter:
760: . tao - the `Tao` context
762: Output Parameter:
763: . recycle - boolean flag
765: Level: intermediate
767: .seealso: [](chapter_tao), `Tao`, `TaoSetRecycleHistory()`, `TAOBNCG`, `TAOBQNLS`, `TAOBQNKLS`, `TAOBQNKTR`, `TAOBQNKTL`
768: @*/
769: PetscErrorCode TaoGetRecycleHistory(Tao tao, PetscBool *recycle)
770: {
771: PetscFunctionBegin;
774: *recycle = tao->recycle;
775: PetscFunctionReturn(PETSC_SUCCESS);
776: }
778: /*@
779: TaoSetTolerances - Sets parameters used in `TaoSolve()` convergence tests
781: Logically Collective
783: Input Parameters:
784: + tao - the `Tao` context
785: . gatol - stop if norm of gradient is less than this
786: . grtol - stop if relative norm of gradient is less than this
787: - gttol - stop if norm of gradient is reduced by this factor
789: Options Database Keys:
790: + -tao_gatol <gatol> - Sets gatol
791: . -tao_grtol <grtol> - Sets grtol
792: - -tao_gttol <gttol> - Sets gttol
794: Stopping Criteria:
795: .vb
796: ||g(X)|| <= gatol
797: ||g(X)|| / |f(X)| <= grtol
798: ||g(X)|| / ||g(X0)|| <= gttol
799: .ve
801: Level: beginner
803: Note:
804: Use `PETSC_DEFAULT` to leave one or more tolerances unchanged.
806: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`, `TaoGetTolerances()`
807: @*/
808: PetscErrorCode TaoSetTolerances(Tao tao, PetscReal gatol, PetscReal grtol, PetscReal gttol)
809: {
810: PetscFunctionBegin;
816: if (gatol != PETSC_DEFAULT) {
817: if (gatol < 0) {
818: PetscCall(PetscInfo(tao, "Tried to set negative gatol -- ignored.\n"));
819: } else {
820: tao->gatol = PetscMax(0, gatol);
821: tao->gatol_changed = PETSC_TRUE;
822: }
823: }
825: if (grtol != PETSC_DEFAULT) {
826: if (grtol < 0) {
827: PetscCall(PetscInfo(tao, "Tried to set negative grtol -- ignored.\n"));
828: } else {
829: tao->grtol = PetscMax(0, grtol);
830: tao->grtol_changed = PETSC_TRUE;
831: }
832: }
834: if (gttol != PETSC_DEFAULT) {
835: if (gttol < 0) {
836: PetscCall(PetscInfo(tao, "Tried to set negative gttol -- ignored.\n"));
837: } else {
838: tao->gttol = PetscMax(0, gttol);
839: tao->gttol_changed = PETSC_TRUE;
840: }
841: }
842: PetscFunctionReturn(PETSC_SUCCESS);
843: }
845: /*@
846: TaoSetConstraintTolerances - Sets constraint tolerance parameters used in `TaoSolve()` convergence tests
848: Logically Collective
850: Input Parameters:
851: + tao - the `Tao` context
852: . catol - absolute constraint tolerance, constraint norm must be less than `catol` for used for gatol convergence criteria
853: - crtol - relative constraint tolerance, constraint norm must be less than `crtol` for used for gatol, gttol convergence criteria
855: Options Database Keys:
856: + -tao_catol <catol> - Sets catol
857: - -tao_crtol <crtol> - Sets crtol
859: Level: intermediate
861: Notes:
862: Use `PETSC_DEFAULT` to leave any tolerance unchanged.
864: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`, `TaoGetTolerances()`, `TaoGetConstraintTolerances()`, `TaoSetTolerances()`
865: @*/
866: PetscErrorCode TaoSetConstraintTolerances(Tao tao, PetscReal catol, PetscReal crtol)
867: {
868: PetscFunctionBegin;
873: if (catol != PETSC_DEFAULT) {
874: if (catol < 0) {
875: PetscCall(PetscInfo(tao, "Tried to set negative catol -- ignored.\n"));
876: } else {
877: tao->catol = PetscMax(0, catol);
878: tao->catol_changed = PETSC_TRUE;
879: }
880: }
882: if (crtol != PETSC_DEFAULT) {
883: if (crtol < 0) {
884: PetscCall(PetscInfo(tao, "Tried to set negative crtol -- ignored.\n"));
885: } else {
886: tao->crtol = PetscMax(0, crtol);
887: tao->crtol_changed = PETSC_TRUE;
888: }
889: }
890: PetscFunctionReturn(PETSC_SUCCESS);
891: }
893: /*@
894: TaoGetConstraintTolerances - Gets constraint tolerance parameters used in `TaoSolve()` convergence tests
896: Not Collective
898: Input Parameter:
899: . tao - the `Tao` context
901: Output Parameters:
902: + catol - absolute constraint tolerance, constraint norm must be less than `catol` for used for gatol convergence criteria
903: - crtol - relative constraint tolerance, constraint norm must be less than `crtol` for used for gatol, gttol convergence criteria
905: Level: intermediate
907: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReasons`,`TaoGetTolerances()`, `TaoSetTolerances()`, `TaoSetConstraintTolerances()`
908: @*/
909: PetscErrorCode TaoGetConstraintTolerances(Tao tao, PetscReal *catol, PetscReal *crtol)
910: {
911: PetscFunctionBegin;
913: if (catol) *catol = tao->catol;
914: if (crtol) *crtol = tao->crtol;
915: PetscFunctionReturn(PETSC_SUCCESS);
916: }
918: /*@
919: TaoSetFunctionLowerBound - Sets a bound on the solution objective value.
920: When an approximate solution with an objective value below this number
921: has been found, the solver will terminate.
923: Logically Collective
925: Input Parameters:
926: + tao - the Tao solver context
927: - fmin - the tolerance
929: Options Database Key:
930: . -tao_fmin <fmin> - sets the minimum function value
932: Level: intermediate
934: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`, `TaoSetTolerances()`
935: @*/
936: PetscErrorCode TaoSetFunctionLowerBound(Tao tao, PetscReal fmin)
937: {
938: PetscFunctionBegin;
941: tao->fmin = fmin;
942: tao->fmin_changed = PETSC_TRUE;
943: PetscFunctionReturn(PETSC_SUCCESS);
944: }
946: /*@
947: TaoGetFunctionLowerBound - Gets the bound on the solution objective value.
948: When an approximate solution with an objective value below this number
949: has been found, the solver will terminate.
951: Not Collective
953: Input Parameter:
954: . tao - the `Tao` solver context
956: OutputParameter:
957: . fmin - the minimum function value
959: Level: intermediate
961: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`, `TaoSetFunctionLowerBound()`
962: @*/
963: PetscErrorCode TaoGetFunctionLowerBound(Tao tao, PetscReal *fmin)
964: {
965: PetscFunctionBegin;
968: *fmin = tao->fmin;
969: PetscFunctionReturn(PETSC_SUCCESS);
970: }
972: /*@
973: TaoSetMaximumFunctionEvaluations - Sets a maximum number of function evaluations allowed for a `TaoSolve()`.
975: Logically Collective
977: Input Parameters:
978: + tao - the `Tao` solver context
979: - nfcn - the maximum number of function evaluations (>=0)
981: Options Database Key:
982: . -tao_max_funcs <nfcn> - sets the maximum number of function evaluations
984: Level: intermediate
986: .seealso: [](chapter_tao), `Tao`, `TaoSetTolerances()`, `TaoSetMaximumIterations()`
987: @*/
988: PetscErrorCode TaoSetMaximumFunctionEvaluations(Tao tao, PetscInt nfcn)
989: {
990: PetscFunctionBegin;
993: if (nfcn >= 0) {
994: tao->max_funcs = PetscMax(0, nfcn);
995: } else {
996: tao->max_funcs = -1;
997: }
998: tao->max_funcs_changed = PETSC_TRUE;
999: PetscFunctionReturn(PETSC_SUCCESS);
1000: }
1002: /*@
1003: TaoGetMaximumFunctionEvaluations - Gets a maximum number of function evaluations allowed for a `TaoSolve()`
1005: Logically Collective
1007: Input Parameter:
1008: . tao - the `Tao` solver context
1010: Output Parameter:
1011: . nfcn - the maximum number of function evaluations
1013: Level: intermediate
1015: .seealso: [](chapter_tao), `Tao`, `TaoSetMaximumFunctionEvaluations()`, `TaoGetMaximumIterations()`
1016: @*/
1017: PetscErrorCode TaoGetMaximumFunctionEvaluations(Tao tao, PetscInt *nfcn)
1018: {
1019: PetscFunctionBegin;
1022: *nfcn = tao->max_funcs;
1023: PetscFunctionReturn(PETSC_SUCCESS);
1024: }
1026: /*@
1027: TaoGetCurrentFunctionEvaluations - Get current number of function evaluations used by a `Tao` object
1029: Not Collective
1031: Input Parameter:
1032: . tao - the `Tao` solver context
1034: Output Parameter:
1035: . nfuncs - the current number of function evaluations (maximum between gradient and function evaluations)
1037: Level: intermediate
1039: .seealso: [](chapter_tao), `Tao`, `TaoSetMaximumFunctionEvaluations()`, `TaoGetMaximumFunctionEvaluations()`, `TaoGetMaximumIterations()`
1040: @*/
1041: PetscErrorCode TaoGetCurrentFunctionEvaluations(Tao tao, PetscInt *nfuncs)
1042: {
1043: PetscFunctionBegin;
1046: *nfuncs = PetscMax(tao->nfuncs, tao->nfuncgrads);
1047: PetscFunctionReturn(PETSC_SUCCESS);
1048: }
1050: /*@
1051: TaoSetMaximumIterations - Sets a maximum number of iterates to be used in `TaoSolve()`
1053: Logically Collective
1055: Input Parameters:
1056: + tao - the `Tao` solver context
1057: - maxits - the maximum number of iterates (>=0)
1059: Options Database Key:
1060: . -tao_max_it <its> - sets the maximum number of iterations
1062: Level: intermediate
1064: .seealso: [](chapter_tao), `Tao`, `TaoSetTolerances()`, `TaoSetMaximumFunctionEvaluations()`
1065: @*/
1066: PetscErrorCode TaoSetMaximumIterations(Tao tao, PetscInt maxits)
1067: {
1068: PetscFunctionBegin;
1071: tao->max_it = PetscMax(0, maxits);
1072: tao->max_it_changed = PETSC_TRUE;
1073: PetscFunctionReturn(PETSC_SUCCESS);
1074: }
1076: /*@
1077: TaoGetMaximumIterations - Gets a maximum number of iterates that will be used
1079: Not Collective
1081: Input Parameter:
1082: . tao - the `Tao` solver context
1084: Output Parameter:
1085: . maxits - the maximum number of iterates
1087: Level: intermediate
1089: .seealso: [](chapter_tao), `Tao`, `TaoSetMaximumIterations()`, `TaoGetMaximumFunctionEvaluations()`
1090: @*/
1091: PetscErrorCode TaoGetMaximumIterations(Tao tao, PetscInt *maxits)
1092: {
1093: PetscFunctionBegin;
1096: *maxits = tao->max_it;
1097: PetscFunctionReturn(PETSC_SUCCESS);
1098: }
1100: /*@
1101: TaoSetInitialTrustRegionRadius - Sets the initial trust region radius.
1103: Logically Collective
1105: Input Parameters:
1106: + tao - a `Tao` optimization solver
1107: - radius - the trust region radius
1109: Options Database Key:
1110: . -tao_trust0 <t0> - sets initial trust region radius
1112: Level: intermediate
1114: .seealso: [](chapter_tao), `Tao`, `TaoGetTrustRegionRadius()`, `TaoSetTrustRegionTolerance()`, `TAONTR`
1115: @*/
1116: PetscErrorCode TaoSetInitialTrustRegionRadius(Tao tao, PetscReal radius)
1117: {
1118: PetscFunctionBegin;
1121: tao->trust0 = PetscMax(0.0, radius);
1122: tao->trust0_changed = PETSC_TRUE;
1123: PetscFunctionReturn(PETSC_SUCCESS);
1124: }
1126: /*@
1127: TaoGetInitialTrustRegionRadius - Gets the initial trust region radius.
1129: Not Collective
1131: Input Parameter:
1132: . tao - a `Tao` optimization solver
1134: Output Parameter:
1135: . radius - the trust region radius
1137: Level: intermediate
1139: .seealso: [](chapter_tao), `Tao`, `TaoSetInitialTrustRegionRadius()`, `TaoGetCurrentTrustRegionRadius()`, `TAONTR`
1140: @*/
1141: PetscErrorCode TaoGetInitialTrustRegionRadius(Tao tao, PetscReal *radius)
1142: {
1143: PetscFunctionBegin;
1146: *radius = tao->trust0;
1147: PetscFunctionReturn(PETSC_SUCCESS);
1148: }
1150: /*@
1151: TaoGetCurrentTrustRegionRadius - Gets the current trust region radius.
1153: Not Collective
1155: Input Parameter:
1156: . tao - a `Tao` optimization solver
1158: Output Parameter:
1159: . radius - the trust region radius
1161: Level: intermediate
1163: .seealso: [](chapter_tao), `Tao`, `TaoSetInitialTrustRegionRadius()`, `TaoGetInitialTrustRegionRadius()`, `TAONTR`
1164: @*/
1165: PetscErrorCode TaoGetCurrentTrustRegionRadius(Tao tao, PetscReal *radius)
1166: {
1167: PetscFunctionBegin;
1170: *radius = tao->trust;
1171: PetscFunctionReturn(PETSC_SUCCESS);
1172: }
1174: /*@
1175: TaoGetTolerances - gets the current values of some tolerances used for the convergence testing of `TaoSolve()`
1177: Not Collective
1179: Input Parameter:
1180: . tao - the `Tao` context
1182: Output Parameters:
1183: + gatol - stop if norm of gradient is less than this
1184: . grtol - stop if relative norm of gradient is less than this
1185: - gttol - stop if norm of gradient is reduced by a this factor
1187: Level: intermediate
1189: Note:
1190: `NULL` can be used as an argument if not all tolerances values are needed
1192: .seealso: [](chapter_tao), `Tao`, `TaoSetTolerances()`
1193: @*/
1194: PetscErrorCode TaoGetTolerances(Tao tao, PetscReal *gatol, PetscReal *grtol, PetscReal *gttol)
1195: {
1196: PetscFunctionBegin;
1198: if (gatol) *gatol = tao->gatol;
1199: if (grtol) *grtol = tao->grtol;
1200: if (gttol) *gttol = tao->gttol;
1201: PetscFunctionReturn(PETSC_SUCCESS);
1202: }
1204: /*@
1205: TaoGetKSP - Gets the linear solver used by the optimization solver.
1207: Not Collective
1209: Input Parameter:
1210: . tao - the `Tao` solver
1212: Output Parameter:
1213: . ksp - the `KSP` linear solver used in the optimization solver
1215: Level: intermediate
1217: .seealso: [](chapter_tao), `Tao`, `KSP`
1218: @*/
1219: PetscErrorCode TaoGetKSP(Tao tao, KSP *ksp)
1220: {
1221: PetscFunctionBegin;
1224: *ksp = tao->ksp;
1225: PetscFunctionReturn(PETSC_SUCCESS);
1226: }
1228: /*@
1229: TaoGetLinearSolveIterations - Gets the total number of linear iterations
1230: used by the `Tao` solver
1232: Not Collective
1234: Input Parameter:
1235: . tao - the `Tao` context
1237: Output Parameter:
1238: . lits - number of linear iterations
1240: Level: intermediate
1242: Note:
1243: This counter is reset to zero for each successive call to `TaoSolve()`
1245: .seealso: [](chapter_tao), `Tao`, `TaoGetKSP()`
1246: @*/
1247: PetscErrorCode TaoGetLinearSolveIterations(Tao tao, PetscInt *lits)
1248: {
1249: PetscFunctionBegin;
1252: *lits = tao->ksp_tot_its;
1253: PetscFunctionReturn(PETSC_SUCCESS);
1254: }
1256: /*@
1257: TaoGetLineSearch - Gets the line search used by the optimization solver.
1259: Not Collective
1261: Input Parameter:
1262: . tao - the `Tao` solver
1264: Output Parameter:
1265: . ls - the line search used in the optimization solver
1267: Level: intermediate
1269: .seealso: [](chapter_tao), `Tao`, `TaoLineSearch`, `TaoLineSearchType`
1270: @*/
1271: PetscErrorCode TaoGetLineSearch(Tao tao, TaoLineSearch *ls)
1272: {
1273: PetscFunctionBegin;
1276: *ls = tao->linesearch;
1277: PetscFunctionReturn(PETSC_SUCCESS);
1278: }
1280: /*@
1281: TaoAddLineSearchCounts - Adds the number of function evaluations spent
1282: in the line search to the running total.
1284: Input Parameters:
1285: + tao - the `Tao` solver
1286: - ls - the line search used in the optimization solver
1288: Level: developer
1290: .seealso: [](chapter_tao), `Tao`, `TaoGetLineSearch()`, `TaoLineSearchApply()`
1291: @*/
1292: PetscErrorCode TaoAddLineSearchCounts(Tao tao)
1293: {
1294: PetscBool flg;
1295: PetscInt nfeval, ngeval, nfgeval;
1297: PetscFunctionBegin;
1299: if (tao->linesearch) {
1300: PetscCall(TaoLineSearchIsUsingTaoRoutines(tao->linesearch, &flg));
1301: if (!flg) {
1302: PetscCall(TaoLineSearchGetNumberFunctionEvaluations(tao->linesearch, &nfeval, &ngeval, &nfgeval));
1303: tao->nfuncs += nfeval;
1304: tao->ngrads += ngeval;
1305: tao->nfuncgrads += nfgeval;
1306: }
1307: }
1308: PetscFunctionReturn(PETSC_SUCCESS);
1309: }
1311: /*@
1312: TaoGetSolution - Returns the vector with the current solution from the `Tao` object
1314: Not Collective
1316: Input Parameter:
1317: . tao - the `Tao` context
1319: Output Parameter:
1320: . X - the current solution
1322: Level: intermediate
1324: Note:
1325: The returned vector will be the same object that was passed into `TaoSetSolution()`
1327: .seealso: [](chapter_tao), `Tao`, `TaoSetSolution()`, `TaoSolve()`
1328: @*/
1329: PetscErrorCode TaoGetSolution(Tao tao, Vec *X)
1330: {
1331: PetscFunctionBegin;
1334: *X = tao->solution;
1335: PetscFunctionReturn(PETSC_SUCCESS);
1336: }
1338: /*@
1339: TaoResetStatistics - Initialize the statistics collected by the `Tao` object.
1340: These statistics include the iteration number, residual norms, and convergence status.
1341: This routine gets called before solving each optimization problem.
1343: Collective
1345: Input Parameter:
1346: . solver - the `Tao` context
1348: Level: developer
1350: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoSolve()`
1351: @*/
1352: PetscErrorCode TaoResetStatistics(Tao tao)
1353: {
1354: PetscFunctionBegin;
1356: tao->niter = 0;
1357: tao->nfuncs = 0;
1358: tao->nfuncgrads = 0;
1359: tao->ngrads = 0;
1360: tao->nhess = 0;
1361: tao->njac = 0;
1362: tao->nconstraints = 0;
1363: tao->ksp_its = 0;
1364: tao->ksp_tot_its = 0;
1365: tao->reason = TAO_CONTINUE_ITERATING;
1366: tao->residual = 0.0;
1367: tao->cnorm = 0.0;
1368: tao->step = 0.0;
1369: tao->lsflag = PETSC_FALSE;
1370: if (tao->hist_reset) tao->hist_len = 0;
1371: PetscFunctionReturn(PETSC_SUCCESS);
1372: }
1374: /*@C
1375: TaoSetUpdate - Sets the general-purpose update function called
1376: at the beginning of every iteration of the optimization algorithm. Specifically
1377: it is called at the top of every iteration, after the new solution and the gradient
1378: is determined, but before the Hessian is computed (if applicable).
1380: Logically Collective
1382: Input Parameters:
1383: + tao - The tao solver context
1384: - func - The function
1386: Calling sequence of func:
1387: $ func (Tao tao, PetscInt step);
1389: . step - The current step of the iteration
1391: Level: advanced
1393: .seealso: [](chapter_tao), `Tao`, `TaoSolve()`
1394: @*/
1395: PetscErrorCode TaoSetUpdate(Tao tao, PetscErrorCode (*func)(Tao, PetscInt, void *), void *ctx)
1396: {
1397: PetscFunctionBegin;
1399: tao->ops->update = func;
1400: tao->user_update = ctx;
1401: PetscFunctionReturn(PETSC_SUCCESS);
1402: }
1404: /*@C
1405: TaoSetConvergenceTest - Sets the function that is to be used to test
1406: for convergence o fthe iterative minimization solution. The new convergence
1407: testing routine will replace Tao's default convergence test.
1409: Logically Collective
1411: Input Parameters:
1412: + tao - the `Tao` object
1413: . conv - the routine to test for convergence
1414: - ctx - [optional] context for private data for the convergence routine
1415: (may be `NULL`)
1417: Calling sequence of conv:
1418: $ PetscErrorCode conv(Tao tao, void *ctx)
1420: + tao - the `Tao` object
1421: - ctx - [optional] convergence context
1423: Level: advanced
1425: Note:
1426: The new convergence testing routine should call `TaoSetConvergedReason()`.
1428: .seealso: [](chapter_tao), `Tao`, `TaoSolve()`, `TaoSetConvergedReason()`, `TaoGetSolutionStatus()`, `TaoGetTolerances()`, `TaoSetMonitor`
1429: @*/
1430: PetscErrorCode TaoSetConvergenceTest(Tao tao, PetscErrorCode (*conv)(Tao, void *), void *ctx)
1431: {
1432: PetscFunctionBegin;
1434: tao->ops->convergencetest = conv;
1435: tao->cnvP = ctx;
1436: PetscFunctionReturn(PETSC_SUCCESS);
1437: }
1439: /*@C
1440: TaoSetMonitor - Sets an additional function that is to be used at every
1441: iteration of the solver to display the iteration's
1442: progress.
1444: Logically Collective
1446: Input Parameters:
1447: + tao - the `Tao` solver context
1448: . mymonitor - monitoring routine
1449: - mctx - [optional] user-defined context for private data for the
1450: monitor routine (may be `NULL`)
1452: Calling sequence of mymonitor:
1453: .vb
1454: PetscErrorCode mymonitor(Tao tao,void *mctx)
1455: .ve
1457: + tao - the `Tao` solver context
1458: - mctx - [optional] monitoring context
1460: Options Database Keys:
1461: + -tao_monitor - sets the default monitor `TaoMonitorDefault()`
1462: . -tao_smonitor - sets short monitor
1463: . -tao_cmonitor - same as smonitor plus constraint norm
1464: . -tao_view_solution - view solution at each iteration
1465: . -tao_view_gradient - view gradient at each iteration
1466: . -tao_view_ls_residual - view least-squares residual vector at each iteration
1467: - -tao_cancelmonitors - cancels all monitors that have been hardwired into a code by calls to TaoSetMonitor(), but does not cancel those set via the options database.
1469: Level: intermediate
1471: Notes:
1472: Several different monitoring routines may be set by calling
1473: `TaoSetMonitor()` multiple times; all will be called in the
1474: order in which they were set.
1476: Fortran Note:
1477: Only one monitor function may be set
1479: .seealso: [](chapter_tao), `Tao`, `TaoSolve()`, `TaoMonitorDefault()`, `TaoCancelMonitors()`, `TaoSetDestroyRoutine()`, `TaoView()`
1480: @*/
1481: PetscErrorCode TaoSetMonitor(Tao tao, PetscErrorCode (*func)(Tao, void *), void *ctx, PetscErrorCode (*dest)(void **))
1482: {
1483: PetscInt i;
1484: PetscBool identical;
1486: PetscFunctionBegin;
1488: PetscCheck(tao->numbermonitors < MAXTAOMONITORS, PetscObjectComm((PetscObject)tao), PETSC_ERR_SUP, "Cannot attach another monitor -- max=%d", MAXTAOMONITORS);
1490: for (i = 0; i < tao->numbermonitors; i++) {
1491: PetscCall(PetscMonitorCompare((PetscErrorCode(*)(void))func, ctx, dest, (PetscErrorCode(*)(void))tao->monitor[i], tao->monitorcontext[i], tao->monitordestroy[i], &identical));
1492: if (identical) PetscFunctionReturn(PETSC_SUCCESS);
1493: }
1494: tao->monitor[tao->numbermonitors] = func;
1495: tao->monitorcontext[tao->numbermonitors] = (void *)ctx;
1496: tao->monitordestroy[tao->numbermonitors] = dest;
1497: ++tao->numbermonitors;
1498: PetscFunctionReturn(PETSC_SUCCESS);
1499: }
1501: /*@
1502: TaoCancelMonitors - Clears all the monitor functions for a `Tao` object.
1504: Logically Collective
1506: Input Parameter:
1507: . tao - the `Tao` solver context
1509: Options Database Key:
1510: . -tao_cancelmonitors - cancels all monitors that have been hardwired
1511: into a code by calls to `TaoSetMonitor()`, but does not cancel those
1512: set via the options database
1514: Level: advanced
1516: Note:
1517: There is no way to clear one specific monitor from a `Tao` object.
1519: .seealso: [](chapter_tao), `Tao`, `TaoMonitorDefault()`, `TaoSetMonitor()`
1520: @*/
1521: PetscErrorCode TaoCancelMonitors(Tao tao)
1522: {
1523: PetscInt i;
1525: PetscFunctionBegin;
1527: for (i = 0; i < tao->numbermonitors; i++) {
1528: if (tao->monitordestroy[i]) PetscCall((*tao->monitordestroy[i])(&tao->monitorcontext[i]));
1529: }
1530: tao->numbermonitors = 0;
1531: PetscFunctionReturn(PETSC_SUCCESS);
1532: }
1534: /*@
1535: TaoMonitorDefault - Default routine for monitoring progress of `TaoSolve()`
1537: Collective
1539: Input Parameters:
1540: + tao - the `Tao` context
1541: - ctx - `PetscViewer` context or `NULL`
1543: Options Database Key:
1544: . -tao_monitor - turn on default monitoring
1546: Level: advanced
1548: Note:
1549: This monitor prints the function value and gradient
1550: norm at each iteration.
1552: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1553: @*/
1554: PetscErrorCode TaoMonitorDefault(Tao tao, void *ctx)
1555: {
1556: PetscInt its, tabs;
1557: PetscReal fct, gnorm;
1558: PetscViewer viewer = (PetscViewer)ctx;
1560: PetscFunctionBegin;
1563: its = tao->niter;
1564: fct = tao->fc;
1565: gnorm = tao->residual;
1566: PetscCall(PetscViewerASCIIGetTab(viewer, &tabs));
1567: PetscCall(PetscViewerASCIISetTab(viewer, ((PetscObject)tao)->tablevel));
1568: if (its == 0 && ((PetscObject)tao)->prefix && !tao->header_printed) {
1569: PetscCall(PetscViewerASCIIPrintf(viewer, " Iteration information for %s solve.\n", ((PetscObject)tao)->prefix));
1570: tao->header_printed = PETSC_TRUE;
1571: }
1572: PetscCall(PetscViewerASCIIPrintf(viewer, "%3" PetscInt_FMT " TAO,", its));
1573: PetscCall(PetscViewerASCIIPrintf(viewer, " Function value: %g,", (double)fct));
1574: if (gnorm >= PETSC_INFINITY) {
1575: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: Inf \n"));
1576: } else {
1577: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: %g \n", (double)gnorm));
1578: }
1579: PetscCall(PetscViewerASCIISetTab(viewer, tabs));
1580: PetscFunctionReturn(PETSC_SUCCESS);
1581: }
1583: /*@
1584: TaoDefaultGMonitor - Default routine for monitoring progress of `TaoSolve()` with extra detail on the globalization method.
1586: Collective
1588: Input Parameters:
1589: + tao - the `Tao` context
1590: - ctx - `PetscViewer` context or `NULL`
1592: Options Database Key:
1593: . -tao_gmonitor - turn on monitoring with globalization information
1595: Level: advanced
1597: Note:
1598: This monitor prints the function value and gradient norm at each
1599: iteration, as well as the step size and trust radius. Note that the
1600: step size and trust radius may be the same for some algorithms.
1602: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1603: @*/
1604: PetscErrorCode TaoDefaultGMonitor(Tao tao, void *ctx)
1605: {
1606: PetscInt its, tabs;
1607: PetscReal fct, gnorm, stp, tr;
1608: PetscViewer viewer = (PetscViewer)ctx;
1610: PetscFunctionBegin;
1613: its = tao->niter;
1614: fct = tao->fc;
1615: gnorm = tao->residual;
1616: stp = tao->step;
1617: tr = tao->trust;
1618: PetscCall(PetscViewerASCIIGetTab(viewer, &tabs));
1619: PetscCall(PetscViewerASCIISetTab(viewer, ((PetscObject)tao)->tablevel));
1620: if (its == 0 && ((PetscObject)tao)->prefix && !tao->header_printed) {
1621: PetscCall(PetscViewerASCIIPrintf(viewer, " Iteration information for %s solve.\n", ((PetscObject)tao)->prefix));
1622: tao->header_printed = PETSC_TRUE;
1623: }
1624: PetscCall(PetscViewerASCIIPrintf(viewer, "%3" PetscInt_FMT " TAO,", its));
1625: PetscCall(PetscViewerASCIIPrintf(viewer, " Function value: %g,", (double)fct));
1626: if (gnorm >= PETSC_INFINITY) {
1627: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: Inf,"));
1628: } else {
1629: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: %g,", (double)gnorm));
1630: }
1631: PetscCall(PetscViewerASCIIPrintf(viewer, " Step: %g, Trust: %g\n", (double)stp, (double)tr));
1632: PetscCall(PetscViewerASCIISetTab(viewer, tabs));
1633: PetscFunctionReturn(PETSC_SUCCESS);
1634: }
1636: /*@
1637: TaoDefaultSMonitor - Default routine for monitoring progress of `TaoSolve()`
1639: Collective
1641: Input Parameters:
1642: + tao - the `Tao` context
1643: - ctx - `PetscViewer` context of type `PETSCVIEWERASCII`
1645: Options Database Key:
1646: . -tao_smonitor - turn on default short monitoring
1648: Level: advanced
1650: Note:
1651: Same as `TaoMonitorDefault()` except
1652: it prints fewer digits of the residual as the residual gets smaller.
1653: This is because the later digits are meaningless and are often
1654: different on different machines; by using this routine different
1655: machines will usually generate the same output.
1657: .seealso: [](chapter_tao), `Tao`, `TaoMonitorDefault()`, `TaoSetMonitor()`
1658: @*/
1659: PetscErrorCode TaoDefaultSMonitor(Tao tao, void *ctx)
1660: {
1661: PetscInt its, tabs;
1662: PetscReal fct, gnorm;
1663: PetscViewer viewer = (PetscViewer)ctx;
1665: PetscFunctionBegin;
1668: its = tao->niter;
1669: fct = tao->fc;
1670: gnorm = tao->residual;
1671: PetscCall(PetscViewerASCIIGetTab(viewer, &tabs));
1672: PetscCall(PetscViewerASCIISetTab(viewer, ((PetscObject)tao)->tablevel));
1673: PetscCall(PetscViewerASCIIPrintf(viewer, "iter = %3" PetscInt_FMT ",", its));
1674: PetscCall(PetscViewerASCIIPrintf(viewer, " Function value %g,", (double)fct));
1675: if (gnorm >= PETSC_INFINITY) {
1676: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: Inf \n"));
1677: } else if (gnorm > 1.e-6) {
1678: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: %g \n", (double)gnorm));
1679: } else if (gnorm > 1.e-11) {
1680: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: < 1.0e-6 \n"));
1681: } else {
1682: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: < 1.0e-11 \n"));
1683: }
1684: PetscCall(PetscViewerASCIISetTab(viewer, tabs));
1685: PetscFunctionReturn(PETSC_SUCCESS);
1686: }
1688: /*@
1689: TaoDefaultCMonitor - same as `TaoMonitorDefault()` except
1690: it prints the norm of the constraint function.
1692: Collective
1694: Input Parameters:
1695: + tao - the `Tao` context
1696: - ctx - `PetscViewer` context or `NULL`
1698: Options Database Key:
1699: . -tao_cmonitor - monitor the constraints
1701: Level: advanced
1703: .seealso: [](chapter_tao), `Tao`, `TaoMonitorDefault()`, `TaoSetMonitor()`
1704: @*/
1705: PetscErrorCode TaoDefaultCMonitor(Tao tao, void *ctx)
1706: {
1707: PetscInt its, tabs;
1708: PetscReal fct, gnorm;
1709: PetscViewer viewer = (PetscViewer)ctx;
1711: PetscFunctionBegin;
1714: its = tao->niter;
1715: fct = tao->fc;
1716: gnorm = tao->residual;
1717: PetscCall(PetscViewerASCIIGetTab(viewer, &tabs));
1718: PetscCall(PetscViewerASCIISetTab(viewer, ((PetscObject)tao)->tablevel));
1719: PetscCall(PetscViewerASCIIPrintf(viewer, "iter = %" PetscInt_FMT ",", its));
1720: PetscCall(PetscViewerASCIIPrintf(viewer, " Function value: %g,", (double)fct));
1721: PetscCall(PetscViewerASCIIPrintf(viewer, " Residual: %g ", (double)gnorm));
1722: PetscCall(PetscViewerASCIIPrintf(viewer, " Constraint: %g \n", (double)tao->cnorm));
1723: PetscCall(PetscViewerASCIISetTab(viewer, tabs));
1724: PetscFunctionReturn(PETSC_SUCCESS);
1725: }
1727: /*@C
1728: TaoSolutionMonitor - Views the solution at each iteration of `TaoSolve()`
1730: Collective
1732: Input Parameters:
1733: + tao - the `Tao` context
1734: - ctx - `PetscViewer` context or `NULL`
1736: Options Database Key:
1737: . -tao_view_solution - view the solution
1739: Level: advanced
1741: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1742: @*/
1743: PetscErrorCode TaoSolutionMonitor(Tao tao, void *ctx)
1744: {
1745: PetscViewer viewer = (PetscViewer)ctx;
1747: PetscFunctionBegin;
1750: PetscCall(VecView(tao->solution, viewer));
1751: PetscFunctionReturn(PETSC_SUCCESS);
1752: }
1754: /*@C
1755: TaoGradientMonitor - Views the gradient at each iteration of `TaoSolve()`
1757: Collective
1759: Input Parameters:
1760: + tao - the `Tao` context
1761: - ctx - `PetscViewer` context or `NULL`
1763: Options Database Key:
1764: . -tao_view_gradient - view the gradient at each iteration
1766: Level: advanced
1768: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1769: @*/
1770: PetscErrorCode TaoGradientMonitor(Tao tao, void *ctx)
1771: {
1772: PetscViewer viewer = (PetscViewer)ctx;
1774: PetscFunctionBegin;
1777: PetscCall(VecView(tao->gradient, viewer));
1778: PetscFunctionReturn(PETSC_SUCCESS);
1779: }
1781: /*@C
1782: TaoStepDirectionMonitor - Views the step-direction at each iteration of `TaoSolve()`
1784: Collective
1786: Input Parameters:
1787: + tao - the `Tao` context
1788: - ctx - `PetscViewer` context or `NULL`
1790: Options Database Key:
1791: . -tao_view_stepdirection - view the step direction vector at each iteration
1793: Level: advanced
1795: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1796: @*/
1797: PetscErrorCode TaoStepDirectionMonitor(Tao tao, void *ctx)
1798: {
1799: PetscViewer viewer = (PetscViewer)ctx;
1801: PetscFunctionBegin;
1804: PetscCall(VecView(tao->stepdirection, viewer));
1805: PetscFunctionReturn(PETSC_SUCCESS);
1806: }
1808: /*@C
1809: TaoDrawSolutionMonitor - Plots the solution at each iteration of `TaoSolve()`
1811: Collective
1813: Input Parameters:
1814: + tao - the `Tao` context
1815: - ctx - `TaoMonitorDraw` context
1817: Options Database Key:
1818: . -tao_draw_solution - draw the solution at each iteration
1820: Level: advanced
1822: .seealso: [](chapter_tao), `Tao`, `TaoSolutionMonitor()`, `TaoSetMonitor()`, `TaoDrawGradientMonitor`, `TaoMonitorDraw`
1823: @*/
1824: PetscErrorCode TaoDrawSolutionMonitor(Tao tao, void *ctx)
1825: {
1826: TaoMonitorDrawCtx ictx = (TaoMonitorDrawCtx)ctx;
1828: PetscFunctionBegin;
1830: if (!(((ictx->howoften > 0) && (!(tao->niter % ictx->howoften))) || ((ictx->howoften == -1) && tao->reason))) PetscFunctionReturn(PETSC_SUCCESS);
1831: PetscCall(VecView(tao->solution, ictx->viewer));
1832: PetscFunctionReturn(PETSC_SUCCESS);
1833: }
1835: /*@C
1836: TaoDrawGradientMonitor - Plots the gradient at each iteration of `TaoSolve()`
1838: Collective
1840: Input Parameters:
1841: + tao - the `Tao` context
1842: - ctx - `PetscViewer` context
1844: Options Database Key:
1845: . -tao_draw_gradient - draw the gradient at each iteration
1847: Level: advanced
1849: .seealso: [](chapter_tao), `Tao`, `TaoGradientMonitor()`, `TaoSetMonitor()`, `TaoDrawSolutionMonitor`
1850: @*/
1851: PetscErrorCode TaoDrawGradientMonitor(Tao tao, void *ctx)
1852: {
1853: TaoMonitorDrawCtx ictx = (TaoMonitorDrawCtx)ctx;
1855: PetscFunctionBegin;
1857: if (!(((ictx->howoften > 0) && (!(tao->niter % ictx->howoften))) || ((ictx->howoften == -1) && tao->reason))) PetscFunctionReturn(PETSC_SUCCESS);
1858: PetscCall(VecView(tao->gradient, ictx->viewer));
1859: PetscFunctionReturn(PETSC_SUCCESS);
1860: }
1862: /*@C
1863: TaoDrawStepMonitor - Plots the step direction at each iteration of `TaoSolve()`
1865: Collective
1867: Input Parameters:
1868: + tao - the `Tao` context
1869: - ctx - the `PetscViewer` context
1871: Options Database Key:
1872: . -tao_draw_step - draw the step direction at each iteration
1874: Level: advanced
1876: .seealso: [](chapter_tao), `Tao`, `TaoSetMonitor()`, `TaoDrawSolutionMonitor`
1877: @*/
1878: PetscErrorCode TaoDrawStepMonitor(Tao tao, void *ctx)
1879: {
1880: PetscViewer viewer = (PetscViewer)ctx;
1882: PetscFunctionBegin;
1885: PetscCall(VecView(tao->stepdirection, viewer));
1886: PetscFunctionReturn(PETSC_SUCCESS);
1887: }
1889: /*@C
1890: TaoResidualMonitor - Views the least-squares residual at each iteration of `TaoSolve()`
1892: Collective
1894: Input Parameters:
1895: + tao - the `Tao` context
1896: - ctx - the `PetscViewer` context or `NULL`
1898: Options Database Key:
1899: . -tao_view_ls_residual - view the residual at each iteration
1901: Level: advanced
1903: .seealso: [](chapter_tao), `Tao`, `TaoDefaultSMonitor()`, `TaoSetMonitor()`
1904: @*/
1905: PetscErrorCode TaoResidualMonitor(Tao tao, void *ctx)
1906: {
1907: PetscViewer viewer = (PetscViewer)ctx;
1909: PetscFunctionBegin;
1912: PetscCall(VecView(tao->ls_res, viewer));
1913: PetscFunctionReturn(PETSC_SUCCESS);
1914: }
1916: /*@
1917: TaoDefaultConvergenceTest - Determines whether the solver should continue iterating
1918: or terminate.
1920: Collective
1922: Input Parameters:
1923: + tao - the `Tao` context
1924: - dummy - unused dummy context
1926: Output Parameter:
1927: . reason - for terminating
1929: Level: developer
1931: Notes:
1932: This routine checks the residual in the optimality conditions, the
1933: relative residual in the optimity conditions, the number of function
1934: evaluations, and the function value to test convergence. Some
1935: solvers may use different convergence routines.
1937: .seealso: [](chapter_tao), `Tao`, `TaoSetTolerances()`, `TaoGetConvergedReason()`, `TaoSetConvergedReason()`
1938: @*/
1939: PetscErrorCode TaoDefaultConvergenceTest(Tao tao, void *dummy)
1940: {
1941: PetscInt niter = tao->niter, nfuncs = PetscMax(tao->nfuncs, tao->nfuncgrads);
1942: PetscInt max_funcs = tao->max_funcs;
1943: PetscReal gnorm = tao->residual, gnorm0 = tao->gnorm0;
1944: PetscReal f = tao->fc, steptol = tao->steptol, trradius = tao->step;
1945: PetscReal gatol = tao->gatol, grtol = tao->grtol, gttol = tao->gttol;
1946: PetscReal catol = tao->catol, crtol = tao->crtol;
1947: PetscReal fmin = tao->fmin, cnorm = tao->cnorm;
1948: TaoConvergedReason reason = tao->reason;
1950: PetscFunctionBegin;
1952: if (reason != TAO_CONTINUE_ITERATING) PetscFunctionReturn(PETSC_SUCCESS);
1954: if (PetscIsInfOrNanReal(f)) {
1955: PetscCall(PetscInfo(tao, "Failed to converged, function value is Inf or NaN\n"));
1956: reason = TAO_DIVERGED_NAN;
1957: } else if (f <= fmin && cnorm <= catol) {
1958: PetscCall(PetscInfo(tao, "Converged due to function value %g < minimum function value %g\n", (double)f, (double)fmin));
1959: reason = TAO_CONVERGED_MINF;
1960: } else if (gnorm <= gatol && cnorm <= catol) {
1961: PetscCall(PetscInfo(tao, "Converged due to residual norm ||g(X)||=%g < %g\n", (double)gnorm, (double)gatol));
1962: reason = TAO_CONVERGED_GATOL;
1963: } else if (f != 0 && PetscAbsReal(gnorm / f) <= grtol && cnorm <= crtol) {
1964: PetscCall(PetscInfo(tao, "Converged due to residual ||g(X)||/|f(X)| =%g < %g\n", (double)(gnorm / f), (double)grtol));
1965: reason = TAO_CONVERGED_GRTOL;
1966: } else if (gnorm0 != 0 && ((gttol == 0 && gnorm == 0) || gnorm / gnorm0 < gttol) && cnorm <= crtol) {
1967: PetscCall(PetscInfo(tao, "Converged due to relative residual norm ||g(X)||/||g(X0)|| = %g < %g\n", (double)(gnorm / gnorm0), (double)gttol));
1968: reason = TAO_CONVERGED_GTTOL;
1969: } else if (max_funcs >= 0 && nfuncs > max_funcs) {
1970: PetscCall(PetscInfo(tao, "Exceeded maximum number of function evaluations: %" PetscInt_FMT " > %" PetscInt_FMT "\n", nfuncs, max_funcs));
1971: reason = TAO_DIVERGED_MAXFCN;
1972: } else if (tao->lsflag != 0) {
1973: PetscCall(PetscInfo(tao, "Tao Line Search failure.\n"));
1974: reason = TAO_DIVERGED_LS_FAILURE;
1975: } else if (trradius < steptol && niter > 0) {
1976: PetscCall(PetscInfo(tao, "Trust region/step size too small: %g < %g\n", (double)trradius, (double)steptol));
1977: reason = TAO_CONVERGED_STEPTOL;
1978: } else if (niter >= tao->max_it) {
1979: PetscCall(PetscInfo(tao, "Exceeded maximum number of iterations: %" PetscInt_FMT " > %" PetscInt_FMT "\n", niter, tao->max_it));
1980: reason = TAO_DIVERGED_MAXITS;
1981: } else {
1982: reason = TAO_CONTINUE_ITERATING;
1983: }
1984: tao->reason = reason;
1985: PetscFunctionReturn(PETSC_SUCCESS);
1986: }
1988: /*@C
1989: TaoSetOptionsPrefix - Sets the prefix used for searching for all
1990: Tao options in the database.
1992: Logically Collective
1994: Input Parameters:
1995: + tao - the `Tao` context
1996: - prefix - the prefix string to prepend to all Tao option requests
1998: Notes:
1999: A hyphen (-) must NOT be given at the beginning of the prefix name.
2000: The first character of all runtime options is AUTOMATICALLY the hyphen.
2002: For example, to distinguish between the runtime options for two
2003: different Tao solvers, one could call
2004: .vb
2005: TaoSetOptionsPrefix(tao1,"sys1_")
2006: TaoSetOptionsPrefix(tao2,"sys2_")
2007: .ve
2009: This would enable use of different options for each system, such as
2010: .vb
2011: -sys1_tao_method blmvm -sys1_tao_grtol 1.e-3
2012: -sys2_tao_method lmvm -sys2_tao_grtol 1.e-4
2013: .ve
2015: Level: advanced
2017: .seealso: [](chapter_tao), `Tao`, `TaoSetFromOptions()`, `TaoAppendOptionsPrefix()`, `TaoGetOptionsPrefix()`
2018: @*/
2019: PetscErrorCode TaoSetOptionsPrefix(Tao tao, const char p[])
2020: {
2021: PetscFunctionBegin;
2023: PetscCall(PetscObjectSetOptionsPrefix((PetscObject)tao, p));
2024: if (tao->linesearch) PetscCall(TaoLineSearchSetOptionsPrefix(tao->linesearch, p));
2025: if (tao->ksp) PetscCall(KSPSetOptionsPrefix(tao->ksp, p));
2026: PetscFunctionReturn(PETSC_SUCCESS);
2027: }
2029: /*@C
2030: TaoAppendOptionsPrefix - Appends to the prefix used for searching for all Tao options in the database.
2032: Logically Collective
2034: Input Parameters:
2035: + tao - the `Tao` solver context
2036: - prefix - the prefix string to prepend to all `Tao` option requests
2038: Note:
2039: A hyphen (-) must NOT be given at the beginning of the prefix name.
2040: The first character of all runtime options is automatically the hyphen.
2042: Level: advanced
2044: .seealso: [](chapter_tao), `Tao`, `TaoSetFromOptions()`, `TaoSetOptionsPrefix()`, `TaoGetOptionsPrefix()`
2045: @*/
2046: PetscErrorCode TaoAppendOptionsPrefix(Tao tao, const char p[])
2047: {
2048: PetscFunctionBegin;
2050: PetscCall(PetscObjectAppendOptionsPrefix((PetscObject)tao, p));
2051: if (tao->linesearch) PetscCall(PetscObjectAppendOptionsPrefix((PetscObject)tao->linesearch, p));
2052: if (tao->ksp) PetscCall(KSPAppendOptionsPrefix(tao->ksp, p));
2053: PetscFunctionReturn(PETSC_SUCCESS);
2054: }
2056: /*@C
2057: TaoGetOptionsPrefix - Gets the prefix used for searching for all
2058: Tao options in the database
2060: Not Collective
2062: Input Parameter:
2063: . tao - the `Tao` context
2065: Output Parameter:
2066: . prefix - pointer to the prefix string used is returned
2068: Fortran Note:
2069: Pass in a string 'prefix' of sufficient length to hold the prefix.
2071: Level: advanced
2073: .seealso: [](chapter_tao), `Tao`, `TaoSetFromOptions()`, `TaoSetOptionsPrefix()`, `TaoAppendOptionsPrefix()`
2074: @*/
2075: PetscErrorCode TaoGetOptionsPrefix(Tao tao, const char *p[])
2076: {
2077: PetscFunctionBegin;
2079: PetscCall(PetscObjectGetOptionsPrefix((PetscObject)tao, p));
2080: PetscFunctionReturn(PETSC_SUCCESS);
2081: }
2083: /*@C
2084: TaoSetType - Sets the `TaoType` for the minimization solver.
2086: Collective
2088: Input Parameters:
2089: + solver - the `Tao` solver context
2090: - type - a known method
2092: Options Database Key:
2093: . -tao_type <type> - Sets the method; use -help for a list
2094: of available methods (for instance, "-tao_type lmvm" or "-tao_type tron")
2096: Level: intermediate
2098: .seealso: [](chapter_tao), `Tao`, `TaoCreate()`, `TaoGetType()`, `TaoType`
2099: @*/
2100: PetscErrorCode TaoSetType(Tao tao, TaoType type)
2101: {
2102: PetscErrorCode (*create_xxx)(Tao);
2103: PetscBool issame;
2105: PetscFunctionBegin;
2108: PetscCall(PetscObjectTypeCompare((PetscObject)tao, type, &issame));
2109: if (issame) PetscFunctionReturn(PETSC_SUCCESS);
2111: PetscCall(PetscFunctionListFind(TaoList, type, (void (**)(void)) & create_xxx));
2112: PetscCheck(create_xxx, PetscObjectComm((PetscObject)tao), PETSC_ERR_ARG_UNKNOWN_TYPE, "Unable to find requested Tao type %s", type);
2114: /* Destroy the existing solver information */
2115: PetscTryTypeMethod(tao, destroy);
2116: PetscCall(KSPDestroy(&tao->ksp));
2117: PetscCall(TaoLineSearchDestroy(&tao->linesearch));
2118: tao->ops->setup = NULL;
2119: tao->ops->solve = NULL;
2120: tao->ops->view = NULL;
2121: tao->ops->setfromoptions = NULL;
2122: tao->ops->destroy = NULL;
2124: tao->setupcalled = PETSC_FALSE;
2126: PetscCall((*create_xxx)(tao));
2127: PetscCall(PetscObjectChangeTypeName((PetscObject)tao, type));
2128: PetscFunctionReturn(PETSC_SUCCESS);
2129: }
2131: /*@C
2132: TaoRegister - Adds a method to the Tao package for minimization.
2134: Synopsis:
2135: TaoRegister(char *name_solver,char *path,char *name_Create,PetscErrorCode (*routine_Create)(Tao))
2137: Not collective
2139: Input Parameters:
2140: + sname - name of a new user-defined solver
2141: - func - routine to Create method context
2143: Sample usage:
2144: .vb
2145: TaoRegister("my_solver",MySolverCreate);
2146: .ve
2148: Then, your solver can be chosen with the procedural interface via
2149: $ TaoSetType(tao,"my_solver")
2150: or at runtime via the option
2151: $ -tao_type my_solver
2153: Level: advanced
2155: Note:
2156: `TaoRegister()` may be called multiple times to add several user-defined solvers.
2158: .seealso: [](chapter_tao), `Tao`, `TaoSetType()`, `TaoRegisterAll()`, `TaoRegisterDestroy()`
2159: @*/
2160: PetscErrorCode TaoRegister(const char sname[], PetscErrorCode (*func)(Tao))
2161: {
2162: PetscFunctionBegin;
2163: PetscCall(TaoInitializePackage());
2164: PetscCall(PetscFunctionListAdd(&TaoList, sname, (void (*)(void))func));
2165: PetscFunctionReturn(PETSC_SUCCESS);
2166: }
2168: /*@C
2169: TaoRegisterDestroy - Frees the list of minimization solvers that were
2170: registered by `TaoRegister()`.
2172: Not Collective
2174: Level: advanced
2176: .seealso: [](chapter_tao), `Tao`, `TaoRegisterAll()`, `TaoRegister()`
2177: @*/
2178: PetscErrorCode TaoRegisterDestroy(void)
2179: {
2180: PetscFunctionBegin;
2181: PetscCall(PetscFunctionListDestroy(&TaoList));
2182: TaoRegisterAllCalled = PETSC_FALSE;
2183: PetscFunctionReturn(PETSC_SUCCESS);
2184: }
2186: /*@
2187: TaoGetIterationNumber - Gets the number of `TaoSolve()` iterations completed
2188: at this time.
2190: Not Collective
2192: Input Parameter:
2193: . tao - the `Tao` context
2195: Output Parameter:
2196: . iter - iteration number
2198: Notes:
2199: For example, during the computation of iteration 2 this would return 1.
2201: Level: intermediate
2203: .seealso: [](chapter_tao), `Tao`, `TaoGetLinearSolveIterations()`, `TaoGetResidualNorm()`, `TaoGetObjective()`
2204: @*/
2205: PetscErrorCode TaoGetIterationNumber(Tao tao, PetscInt *iter)
2206: {
2207: PetscFunctionBegin;
2210: *iter = tao->niter;
2211: PetscFunctionReturn(PETSC_SUCCESS);
2212: }
2214: /*@
2215: TaoGetResidualNorm - Gets the current value of the norm of the residual (gradient)
2216: at this time.
2218: Not Collective
2220: Input Parameter:
2221: . tao - the `Tao` context
2223: Output Parameter:
2224: . value - the current value
2226: Level: intermediate
2228: Developer Note:
2229: This is the 2-norm of the residual, we cannot use `TaoGetGradientNorm()` because that has
2230: a different meaning. For some reason `Tao` sometimes calls the gradient the residual.
2232: .seealso: [](chapter_tao), `Tao`, `TaoGetLinearSolveIterations()`, `TaoGetIterationNumber()`, `TaoGetObjective()`
2233: @*/
2234: PetscErrorCode TaoGetResidualNorm(Tao tao, PetscReal *value)
2235: {
2236: PetscFunctionBegin;
2239: *value = tao->residual;
2240: PetscFunctionReturn(PETSC_SUCCESS);
2241: }
2243: /*@
2244: TaoSetIterationNumber - Sets the current iteration number.
2246: Logically Collective
2248: Input Parameters:
2249: + tao - the `Tao` context
2250: - iter - iteration number
2252: Level: developer
2254: .seealso: [](chapter_tao), `Tao`, `TaoGetLinearSolveIterations()`
2255: @*/
2256: PetscErrorCode TaoSetIterationNumber(Tao tao, PetscInt iter)
2257: {
2258: PetscFunctionBegin;
2261: PetscCall(PetscObjectSAWsTakeAccess((PetscObject)tao));
2262: tao->niter = iter;
2263: PetscCall(PetscObjectSAWsGrantAccess((PetscObject)tao));
2264: PetscFunctionReturn(PETSC_SUCCESS);
2265: }
2267: /*@
2268: TaoGetTotalIterationNumber - Gets the total number of `TaoSolve()` iterations
2269: completed. This number keeps accumulating if multiple solves
2270: are called with the `Tao` object.
2272: Not Collective
2274: Input Parameter:
2275: . tao - the `Tao` context
2277: Output Parameter:
2278: . iter - number of iterations
2280: Level: intermediate
2282: Notes:
2283: The total iteration count is updated after each solve, if there is a current
2284: `TaoSolve()` in progress then those iterations are not included in the count
2286: .seealso: [](chapter_tao), `Tao`, `TaoGetLinearSolveIterations()`
2287: @*/
2288: PetscErrorCode TaoGetTotalIterationNumber(Tao tao, PetscInt *iter)
2289: {
2290: PetscFunctionBegin;
2293: *iter = tao->ntotalits;
2294: PetscFunctionReturn(PETSC_SUCCESS);
2295: }
2297: /*@
2298: TaoSetTotalIterationNumber - Sets the current total iteration number.
2300: Logically Collective
2302: Input Parameters:
2303: + tao - the `Tao` context
2304: - iter - the iteration number
2306: Level: developer
2308: .seealso: [](chapter_tao), `Tao`, `TaoGetLinearSolveIterations()`
2309: @*/
2310: PetscErrorCode TaoSetTotalIterationNumber(Tao tao, PetscInt iter)
2311: {
2312: PetscFunctionBegin;
2315: PetscCall(PetscObjectSAWsTakeAccess((PetscObject)tao));
2316: tao->ntotalits = iter;
2317: PetscCall(PetscObjectSAWsGrantAccess((PetscObject)tao));
2318: PetscFunctionReturn(PETSC_SUCCESS);
2319: }
2321: /*@
2322: TaoSetConvergedReason - Sets the termination flag on a `Tao` object
2324: Logically Collective
2326: Input Parameters:
2327: + tao - the `Tao` context
2328: - reason - the `TaoConvergedReason`
2330: Level: intermediate
2332: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`
2333: @*/
2334: PetscErrorCode TaoSetConvergedReason(Tao tao, TaoConvergedReason reason)
2335: {
2336: PetscFunctionBegin;
2339: tao->reason = reason;
2340: PetscFunctionReturn(PETSC_SUCCESS);
2341: }
2343: /*@
2344: TaoGetConvergedReason - Gets the reason the `TaoSolve()` was stopped.
2346: Not Collective
2348: Input Parameter:
2349: . tao - the `Tao` solver context
2351: Output Parameter:
2352: . reason - value of `TaoConvergedReason`
2354: Level: intermediate
2356: .seealso: [](chapter_tao), `Tao`, `TaoConvergedReason`, `TaoSetConvergenceTest()`, `TaoSetTolerances()`
2357: @*/
2358: PetscErrorCode TaoGetConvergedReason(Tao tao, TaoConvergedReason *reason)
2359: {
2360: PetscFunctionBegin;
2363: *reason = tao->reason;
2364: PetscFunctionReturn(PETSC_SUCCESS);
2365: }
2367: /*@
2368: TaoGetSolutionStatus - Get the current iterate, objective value,
2369: residual, infeasibility, and termination from a `Tao` object
2371: Not Collective
2373: Input Parameter:
2374: . tao - the `Tao` context
2376: Output Parameters:
2377: + iterate - the current iterate number (>=0)
2378: . f - the current function value
2379: . gnorm - the square of the gradient norm, duality gap, or other measure indicating distance from optimality.
2380: . cnorm - the infeasibility of the current solution with regard to the constraints.
2381: . xdiff - the step length or trust region radius of the most recent iterate.
2382: - reason - The termination reason, which can equal `TAO_CONTINUE_ITERATING`
2384: Level: intermediate
2386: Notes:
2387: Tao returns the values set by the solvers in the routine `TaoMonitor()`.
2389: If any of the output arguments are set to `NULL`, no corresponding value will be returned.
2391: .seealso: [](chapter_tao), `TaoMonitor()`, `TaoGetConvergedReason()`
2392: @*/
2393: PetscErrorCode TaoGetSolutionStatus(Tao tao, PetscInt *its, PetscReal *f, PetscReal *gnorm, PetscReal *cnorm, PetscReal *xdiff, TaoConvergedReason *reason)
2394: {
2395: PetscFunctionBegin;
2397: if (its) *its = tao->niter;
2398: if (f) *f = tao->fc;
2399: if (gnorm) *gnorm = tao->residual;
2400: if (cnorm) *cnorm = tao->cnorm;
2401: if (reason) *reason = tao->reason;
2402: if (xdiff) *xdiff = tao->step;
2403: PetscFunctionReturn(PETSC_SUCCESS);
2404: }
2406: /*@C
2407: TaoGetType - Gets the current `TaoType` being used in the `Tao` object
2409: Not Collective
2411: Input Parameter:
2412: . tao - the `Tao` solver context
2414: Output Parameter:
2415: . type - the `TaoType`
2417: Level: intermediate
2419: .seealso: [](chapter_tao), `Tao`, `TaoType`, `TaoSetType()`
2420: @*/
2421: PetscErrorCode TaoGetType(Tao tao, TaoType *type)
2422: {
2423: PetscFunctionBegin;
2426: *type = ((PetscObject)tao)->type_name;
2427: PetscFunctionReturn(PETSC_SUCCESS);
2428: }
2430: /*@C
2431: TaoMonitor - Monitor the solver and the current solution. This
2432: routine will record the iteration number and residual statistics,
2433: and call any monitors specified by the user.
2435: Input Parameters:
2436: + tao - the `Tao` context
2437: . its - the current iterate number (>=0)
2438: . f - the current objective function value
2439: . res - the gradient norm, square root of the duality gap, or other measure indicating distince from optimality. This measure will be recorded and
2440: used for some termination tests.
2441: . cnorm - the infeasibility of the current solution with regard to the constraints.
2442: - steplength - multiple of the step direction added to the previous iterate.
2444: Output Parameter:
2445: . reason - The termination reason, which can equal `TAO_CONTINUE_ITERATING`
2447: Options Database Key:
2448: . -tao_monitor - Use the default monitor, which prints statistics to standard output
2450: Level: developer
2452: .seealso: [](chapter_tao), `Tao`, `TaoGetConvergedReason()`, `TaoMonitorDefault()`, `TaoSetMonitor()`
2453: @*/
2454: PetscErrorCode TaoMonitor(Tao tao, PetscInt its, PetscReal f, PetscReal res, PetscReal cnorm, PetscReal steplength)
2455: {
2456: PetscInt i;
2458: PetscFunctionBegin;
2460: tao->fc = f;
2461: tao->residual = res;
2462: tao->cnorm = cnorm;
2463: tao->step = steplength;
2464: if (!its) {
2465: tao->cnorm0 = cnorm;
2466: tao->gnorm0 = res;
2467: }
2468: PetscCheck(!PetscIsInfOrNanReal(f) && !PetscIsInfOrNanReal(res), PetscObjectComm((PetscObject)tao), PETSC_ERR_USER, "User provided compute function generated Inf or NaN");
2469: for (i = 0; i < tao->numbermonitors; i++) PetscCall((*tao->monitor[i])(tao, tao->monitorcontext[i]));
2470: PetscFunctionReturn(PETSC_SUCCESS);
2471: }
2473: /*@
2474: TaoSetConvergenceHistory - Sets the array used to hold the convergence history.
2476: Logically Collective
2478: Input Parameters:
2479: + tao - the `Tao` solver context
2480: . obj - array to hold objective value history
2481: . resid - array to hold residual history
2482: . cnorm - array to hold constraint violation history
2483: . lits - integer array holds the number of linear iterations for each Tao iteration
2484: . na - size of `obj`, `resid`, and `cnorm`
2485: - reset - `PETSC_TRUE` indicates each new minimization resets the history counter to zero,
2486: else it continues storing new values for new minimizations after the old ones
2488: Level: intermediate
2490: Notes:
2491: If set, `Tao` will fill the given arrays with the indicated
2492: information at each iteration. If 'obj','resid','cnorm','lits' are
2493: *all* `NULL` then space (using size `na`, or 1000 if na is `PETSC_DECIDE` or
2494: `PETSC_DEFAULT`) is allocated for the history.
2495: If not all are `NULL`, then only the non-`NULL` information categories
2496: will be stored, the others will be ignored.
2498: Any convergence information after iteration number 'na' will not be stored.
2500: This routine is useful, e.g., when running a code for purposes
2501: of accurate performance monitoring, when no I/O should be done
2502: during the section of code that is being timed.
2504: .seealso: [](chapter_tao), `TaoGetConvergenceHistory()`
2505: @*/
2506: PetscErrorCode TaoSetConvergenceHistory(Tao tao, PetscReal obj[], PetscReal resid[], PetscReal cnorm[], PetscInt lits[], PetscInt na, PetscBool reset)
2507: {
2508: PetscFunctionBegin;
2515: if (na == PETSC_DECIDE || na == PETSC_DEFAULT) na = 1000;
2516: if (!obj && !resid && !cnorm && !lits) {
2517: PetscCall(PetscCalloc4(na, &obj, na, &resid, na, &cnorm, na, &lits));
2518: tao->hist_malloc = PETSC_TRUE;
2519: }
2521: tao->hist_obj = obj;
2522: tao->hist_resid = resid;
2523: tao->hist_cnorm = cnorm;
2524: tao->hist_lits = lits;
2525: tao->hist_max = na;
2526: tao->hist_reset = reset;
2527: tao->hist_len = 0;
2528: PetscFunctionReturn(PETSC_SUCCESS);
2529: }
2531: /*@C
2532: TaoGetConvergenceHistory - Gets the arrays used that hold the convergence history.
2534: Collective
2536: Input Parameter:
2537: . tao - the `Tao` context
2539: Output Parameters:
2540: + obj - array used to hold objective value history
2541: . resid - array used to hold residual history
2542: . cnorm - array used to hold constraint violation history
2543: . lits - integer array used to hold linear solver iteration count
2544: - nhist - size of `obj`, `resid`, `cnorm`, and `lits`
2546: Level: advanced
2548: Notes:
2549: This routine must be preceded by calls to `TaoSetConvergenceHistory()`
2550: and `TaoSolve()`, otherwise it returns useless information.
2552: This routine is useful, e.g., when running a code for purposes
2553: of accurate performance monitoring, when no I/O should be done
2554: during the section of code that is being timed.
2556: Fortran Note:
2557: The calling sequence is
2558: .vb
2559: call TaoGetConvergenceHistory(Tao tao, PetscInt nhist, PetscErrorCode ierr)
2560: .ve
2562: .seealso: [](chapter_tao), `Tao`, `TaoSolve()`, `TaoSetConvergenceHistory()`
2563: @*/
2564: PetscErrorCode TaoGetConvergenceHistory(Tao tao, PetscReal **obj, PetscReal **resid, PetscReal **cnorm, PetscInt **lits, PetscInt *nhist)
2565: {
2566: PetscFunctionBegin;
2568: if (obj) *obj = tao->hist_obj;
2569: if (cnorm) *cnorm = tao->hist_cnorm;
2570: if (resid) *resid = tao->hist_resid;
2571: if (lits) *lits = tao->hist_lits;
2572: if (nhist) *nhist = tao->hist_len;
2573: PetscFunctionReturn(PETSC_SUCCESS);
2574: }
2576: /*@
2577: TaoSetApplicationContext - Sets the optional user-defined context for a `Tao` solver.
2579: Logically Collective
2581: Input Parameters:
2582: + tao - the `Tao` context
2583: - usrP - optional user context
2585: Level: intermediate
2587: .seealso: [](chapter_tao), `Tao`, `TaoGetApplicationContext()`, `TaoSetApplicationContext()`
2588: @*/
2589: PetscErrorCode TaoSetApplicationContext(Tao tao, void *usrP)
2590: {
2591: PetscFunctionBegin;
2593: tao->user = usrP;
2594: PetscFunctionReturn(PETSC_SUCCESS);
2595: }
2597: /*@
2598: TaoGetApplicationContext - Gets the user-defined context for a `Tao` solver
2600: Not Collective
2602: Input Parameter:
2603: . tao - the `Tao` context
2605: Output Parameter:
2606: . usrP - user context
2608: Level: intermediate
2610: .seealso: [](chapter_tao), `Tao`, `TaoSetApplicationContext()`
2611: @*/
2612: PetscErrorCode TaoGetApplicationContext(Tao tao, void *usrP)
2613: {
2614: PetscFunctionBegin;
2617: *(void **)usrP = tao->user;
2618: PetscFunctionReturn(PETSC_SUCCESS);
2619: }
2621: /*@
2622: TaoSetGradientNorm - Sets the matrix used to define the norm that measures the size of the gradient.
2624: Collective
2626: Input Parameters:
2627: + tao - the `Tao` context
2628: - M - matrix that defines the norm
2630: Level: beginner
2632: .seealso: [](chapter_tao), `Tao`, `TaoGetGradientNorm()`, `TaoGradientNorm()`
2633: @*/
2634: PetscErrorCode TaoSetGradientNorm(Tao tao, Mat M)
2635: {
2636: PetscFunctionBegin;
2639: PetscCall(PetscObjectReference((PetscObject)M));
2640: PetscCall(MatDestroy(&tao->gradient_norm));
2641: PetscCall(VecDestroy(&tao->gradient_norm_tmp));
2642: tao->gradient_norm = M;
2643: PetscCall(MatCreateVecs(M, NULL, &tao->gradient_norm_tmp));
2644: PetscFunctionReturn(PETSC_SUCCESS);
2645: }
2647: /*@
2648: TaoGetGradientNorm - Returns the matrix used to define the norm used for measuring the size of the gradient.
2650: Not Collective
2652: Input Parameter:
2653: . tao - the `Tao` context
2655: Output Parameter:
2656: . M - gradient norm
2658: Level: beginner
2660: .seealso: [](chapter_tao), `Tao`, `TaoSetGradientNorm()`, `TaoGradientNorm()`
2661: @*/
2662: PetscErrorCode TaoGetGradientNorm(Tao tao, Mat *M)
2663: {
2664: PetscFunctionBegin;
2667: *M = tao->gradient_norm;
2668: PetscFunctionReturn(PETSC_SUCCESS);
2669: }
2671: /*@C
2672: TaoGradientNorm - Compute the norm using the `NormType`, the user has selected
2674: Collective
2676: Input Parameters:
2677: + tao - the `Tao` context
2678: . gradient - the gradient to be computed
2679: - norm - the norm type
2681: Output Parameter:
2682: . gnorm - the gradient norm
2684: Level: advanced
2686: .seealso: [](chapter_tao), `Tao`, `TaoSetGradientNorm()`, `TaoGetGradientNorm()`
2687: @*/
2688: PetscErrorCode TaoGradientNorm(Tao tao, Vec gradient, NormType type, PetscReal *gnorm)
2689: {
2690: PetscFunctionBegin;
2695: if (tao->gradient_norm) {
2696: PetscScalar gnorms;
2698: PetscCheck(type == NORM_2, PetscObjectComm((PetscObject)gradient), PETSC_ERR_ARG_WRONG, "Norm type must be NORM_2 if an inner product for the gradient norm is set.");
2699: PetscCall(MatMult(tao->gradient_norm, gradient, tao->gradient_norm_tmp));
2700: PetscCall(VecDot(gradient, tao->gradient_norm_tmp, &gnorms));
2701: *gnorm = PetscRealPart(PetscSqrtScalar(gnorms));
2702: } else {
2703: PetscCall(VecNorm(gradient, type, gnorm));
2704: }
2705: PetscFunctionReturn(PETSC_SUCCESS);
2706: }
2708: /*@C
2709: TaoMonitorDrawCtxCreate - Creates the monitor context for `TaoMonitorDrawSolution()`
2711: Collective
2713: Output Parameter:
2714: . ctx - the monitor context
2716: Options Database Key:
2717: . -tao_draw_solution_initial - show initial guess as well as current solution
2719: Level: intermediate
2721: .seealso: [](chapter_tao), `Tao`, `TaoMonitorSet()`, `TaoMonitorDefault()`, `VecView()`, `TaoMonitorDrawCtx()`
2722: @*/
2723: PetscErrorCode TaoMonitorDrawCtxCreate(MPI_Comm comm, const char host[], const char label[], int x, int y, int m, int n, PetscInt howoften, TaoMonitorDrawCtx *ctx)
2724: {
2725: PetscFunctionBegin;
2726: PetscCall(PetscNew(ctx));
2727: PetscCall(PetscViewerDrawOpen(comm, host, label, x, y, m, n, &(*ctx)->viewer));
2728: PetscCall(PetscViewerSetFromOptions((*ctx)->viewer));
2729: (*ctx)->howoften = howoften;
2730: PetscFunctionReturn(PETSC_SUCCESS);
2731: }
2733: /*@C
2734: TaoMonitorDrawCtxDestroy - Destroys the monitor context for `TaoMonitorDrawSolution()`
2736: Collective
2738: Input Parameter:
2739: . ctx - the monitor context
2741: Level: intermediate
2743: .seealso: [](chapter_tao), `Tao`, `TaoMonitorSet()`, `TaoMonitorDefault()`, `VecView()`, `TaoMonitorDrawSolution()`
2744: @*/
2745: PetscErrorCode TaoMonitorDrawCtxDestroy(TaoMonitorDrawCtx *ictx)
2746: {
2747: PetscFunctionBegin;
2748: PetscCall(PetscViewerDestroy(&(*ictx)->viewer));
2749: PetscCall(PetscFree(*ictx));
2750: PetscFunctionReturn(PETSC_SUCCESS);
2751: }