xref: /petsc/src/tao/unconstrained/impls/lmvm/lmvm.c (revision 3e7ff0edd3573be01c8c0fa32db97c3db8fa5c8d)
1 #include <petsctaolinesearch.h>
2 #include <../src/tao/matrix/lmvmmat.h>
3 #include <../src/tao/unconstrained/impls/lmvm/lmvm.h>
4 
5 #define LMVM_BFGS                0
6 #define LMVM_SCALED_GRADIENT     1
7 #define LMVM_GRADIENT            2
8 
9 #undef __FUNCT__
10 #define __FUNCT__ "TaoSolve_LMVM"
11 static PetscErrorCode TaoSolve_LMVM(Tao tao)
12 {
13   TAO_LMVM                     *lmP = (TAO_LMVM *)tao->data;
14   PetscReal                    f, fold, gdx, gnorm;
15   PetscReal                    step = 1.0;
16   PetscReal                    delta;
17   PetscErrorCode               ierr;
18   PetscInt                     stepType;
19   PetscInt                     iter = 0;
20   TaoConvergedReason           reason = TAO_CONTINUE_ITERATING;
21   TaoLineSearchConvergedReason ls_status = TAOLINESEARCH_CONTINUE_ITERATING;
22 
23   PetscFunctionBegin;
24 
25   if (tao->XL || tao->XU || tao->ops->computebounds) {
26     ierr = PetscPrintf(((PetscObject)tao)->comm,"WARNING: Variable bounds have been set but will be ignored by lmvm algorithm\n");CHKERRQ(ierr);
27   }
28 
29   /*  Check convergence criteria */
30   ierr = TaoComputeObjectiveAndGradient(tao, tao->solution, &f, tao->gradient);CHKERRQ(ierr);
31   ierr = VecNorm(tao->gradient,NORM_2,&gnorm);CHKERRQ(ierr);
32   if (PetscIsInfOrNanReal(f) || PetscIsInfOrNanReal(gnorm)) SETERRQ(PETSC_COMM_SELF,1, "User provided compute function generated Inf or NaN");
33 
34   ierr = TaoMonitor(tao, iter, f, gnorm, 0.0, step, &reason);CHKERRQ(ierr);
35   if (reason != TAO_CONTINUE_ITERATING) PetscFunctionReturn(0);
36 
37   /*  Set initial scaling for the function */
38   if (f != 0.0) {
39     delta = 2.0 * PetscAbsScalar(f) / (gnorm*gnorm);
40   } else {
41     delta = 2.0 / (gnorm*gnorm);
42   }
43   ierr = MatLMVMSetDelta(lmP->M,delta);CHKERRQ(ierr);
44 
45   /*  Set counter for gradient/reset steps */
46   lmP->bfgs = 0;
47   lmP->sgrad = 0;
48   lmP->grad = 0;
49 
50   /*  Have not converged; continue with Newton method */
51   while (reason == TAO_CONTINUE_ITERATING) {
52     /*  Compute direction */
53     ierr = MatLMVMUpdate(lmP->M,tao->solution,tao->gradient);CHKERRQ(ierr);
54     ierr = MatLMVMSolve(lmP->M, tao->gradient, lmP->D);CHKERRQ(ierr);
55     ++lmP->bfgs;
56 
57     /*  Check for success (descent direction) */
58     ierr = VecDot(lmP->D, tao->gradient, &gdx);CHKERRQ(ierr);
59     if ((gdx <= 0.0) || PetscIsInfOrNanReal(gdx)) {
60       /* Step is not descent or direction produced not a number
61          We can assert bfgsUpdates > 1 in this case because
62          the first solve produces the scaled gradient direction,
63          which is guaranteed to be descent
64 
65          Use steepest descent direction (scaled)
66       */
67 
68       ++lmP->grad;
69 
70       if (f != 0.0) {
71         delta = 2.0 * PetscAbsScalar(f) / (gnorm*gnorm);
72       } else {
73         delta = 2.0 / (gnorm*gnorm);
74       }
75       ierr = MatLMVMSetDelta(lmP->M, delta);CHKERRQ(ierr);
76       ierr = MatLMVMReset(lmP->M);CHKERRQ(ierr);
77       ierr = MatLMVMUpdate(lmP->M, tao->solution, tao->gradient);CHKERRQ(ierr);
78       ierr = MatLMVMSolve(lmP->M,tao->gradient, lmP->D);CHKERRQ(ierr);
79 
80       /* On a reset, the direction cannot be not a number; it is a
81          scaled gradient step.  No need to check for this condition. */
82 
83       lmP->bfgs = 1;
84       ++lmP->sgrad;
85       stepType = LMVM_SCALED_GRADIENT;
86     } else {
87       if (1 == lmP->bfgs) {
88         /*  The first BFGS direction is always the scaled gradient */
89         ++lmP->sgrad;
90         stepType = LMVM_SCALED_GRADIENT;
91       } else {
92         ++lmP->bfgs;
93         stepType = LMVM_BFGS;
94       }
95     }
96     ierr = VecScale(lmP->D, -1.0);CHKERRQ(ierr);
97 
98     /*  Perform the linesearch */
99     fold = f;
100     ierr = VecCopy(tao->solution, lmP->Xold);CHKERRQ(ierr);
101     ierr = VecCopy(tao->gradient, lmP->Gold);CHKERRQ(ierr);
102 
103     ierr = TaoLineSearchApply(tao->linesearch, tao->solution, &f, tao->gradient, lmP->D, &step,&ls_status);CHKERRQ(ierr);
104     ierr = TaoAddLineSearchCounts(tao);CHKERRQ(ierr);
105 
106     while (ls_status != TAOLINESEARCH_SUCCESS && ls_status != TAOLINESEARCH_SUCCESS_USER && (stepType != LMVM_GRADIENT)) {
107       /*  Linesearch failed */
108       /*  Reset factors and use scaled gradient step */
109       f = fold;
110       ierr = VecCopy(lmP->Xold, tao->solution);CHKERRQ(ierr);
111       ierr = VecCopy(lmP->Gold, tao->gradient);CHKERRQ(ierr);
112 
113       switch(stepType) {
114       case LMVM_BFGS:
115         /*  Failed to obtain acceptable iterate with BFGS step */
116         /*  Attempt to use the scaled gradient direction */
117 
118         if (f != 0.0) {
119           delta = 2.0 * PetscAbsScalar(f) / (gnorm*gnorm);
120         } else {
121           delta = 2.0 / (gnorm*gnorm);
122         }
123         ierr = MatLMVMSetDelta(lmP->M, delta);CHKERRQ(ierr);
124         ierr = MatLMVMReset(lmP->M);CHKERRQ(ierr);
125         ierr = MatLMVMUpdate(lmP->M, tao->solution, tao->gradient);CHKERRQ(ierr);
126         ierr = MatLMVMSolve(lmP->M, tao->gradient, lmP->D);CHKERRQ(ierr);
127 
128         /* On a reset, the direction cannot be not a number; it is a
129            scaled gradient step.  No need to check for this condition. */
130 
131         lmP->bfgs = 1;
132         ++lmP->sgrad;
133         stepType = LMVM_SCALED_GRADIENT;
134         break;
135 
136       case LMVM_SCALED_GRADIENT:
137         /* The scaled gradient step did not produce a new iterate;
138            attempt to use the gradient direction.
139            Need to make sure we are not using a different diagonal scaling */
140         ierr = MatLMVMSetDelta(lmP->M, 1.0);CHKERRQ(ierr);
141         ierr = MatLMVMReset(lmP->M);CHKERRQ(ierr);
142         ierr = MatLMVMUpdate(lmP->M, tao->solution, tao->gradient);CHKERRQ(ierr);
143         ierr = MatLMVMSolve(lmP->M, tao->gradient, lmP->D);CHKERRQ(ierr);
144 
145         lmP->bfgs = 1;
146         ++lmP->grad;
147         stepType = LMVM_GRADIENT;
148         break;
149       }
150       ierr = VecScale(lmP->D, -1.0);CHKERRQ(ierr);
151 
152       /*  Perform the linesearch */
153       ierr = TaoLineSearchApply(tao->linesearch, tao->solution, &f, tao->gradient, lmP->D, &step, &ls_status);CHKERRQ(ierr);
154       ierr = TaoAddLineSearchCounts(tao);CHKERRQ(ierr);
155     }
156 
157     if (ls_status != TAOLINESEARCH_SUCCESS && ls_status != TAOLINESEARCH_SUCCESS_USER) {
158       /*  Failed to find an improving point */
159       f = fold;
160       ierr = VecCopy(lmP->Xold, tao->solution);CHKERRQ(ierr);
161       ierr = VecCopy(lmP->Gold, tao->gradient);CHKERRQ(ierr);
162       step = 0.0;
163       reason = TAO_DIVERGED_LS_FAILURE;
164       tao->reason = TAO_DIVERGED_LS_FAILURE;
165     }
166     /*  Check for termination */
167     ierr = VecNorm(tao->gradient, NORM_2, &gnorm);CHKERRQ(ierr);
168     iter++;
169     ierr = TaoMonitor(tao,iter,f,gnorm,0.0,step,&reason);CHKERRQ(ierr);
170   }
171   PetscFunctionReturn(0);
172 }
173 
174 #undef __FUNCT__
175 #define __FUNCT__ "TaoSetUp_LMVM"
176 static PetscErrorCode TaoSetUp_LMVM(Tao tao)
177 {
178   TAO_LMVM       *lmP = (TAO_LMVM *)tao->data;
179   PetscInt       n,N;
180   PetscErrorCode ierr;
181 
182   PetscFunctionBegin;
183   /* Existence of tao->solution checked in TaoSetUp() */
184   if (!tao->gradient) {ierr = VecDuplicate(tao->solution,&tao->gradient);CHKERRQ(ierr);  }
185   if (!tao->stepdirection) {ierr = VecDuplicate(tao->solution,&tao->stepdirection);CHKERRQ(ierr);  }
186   if (!lmP->D) {ierr = VecDuplicate(tao->solution,&lmP->D);CHKERRQ(ierr);  }
187   if (!lmP->Xold) {ierr = VecDuplicate(tao->solution,&lmP->Xold);CHKERRQ(ierr);  }
188   if (!lmP->Gold) {ierr = VecDuplicate(tao->solution,&lmP->Gold);CHKERRQ(ierr);  }
189 
190   /*  Create matrix for the limited memory approximation */
191   ierr = VecGetLocalSize(tao->solution,&n);CHKERRQ(ierr);
192   ierr = VecGetSize(tao->solution,&N);CHKERRQ(ierr);
193   ierr = MatCreateLMVM(((PetscObject)tao)->comm,n,N,&lmP->M);CHKERRQ(ierr);
194   ierr = MatLMVMAllocateVectors(lmP->M,tao->solution);CHKERRQ(ierr);
195   PetscFunctionReturn(0);
196 }
197 
198 /* ---------------------------------------------------------- */
199 #undef __FUNCT__
200 #define __FUNCT__ "TaoDestroy_LMVM"
201 static PetscErrorCode TaoDestroy_LMVM(Tao tao)
202 {
203   TAO_LMVM       *lmP = (TAO_LMVM *)tao->data;
204   PetscErrorCode ierr;
205 
206   PetscFunctionBegin;
207   if (tao->setupcalled) {
208     ierr = VecDestroy(&lmP->Xold);CHKERRQ(ierr);
209     ierr = VecDestroy(&lmP->Gold);CHKERRQ(ierr);
210     ierr = VecDestroy(&lmP->D);CHKERRQ(ierr);
211     ierr = MatDestroy(&lmP->M);CHKERRQ(ierr);
212   }
213   ierr = PetscFree(tao->data);CHKERRQ(ierr);
214   PetscFunctionReturn(0);
215 }
216 
217 /*------------------------------------------------------------*/
218 #undef __FUNCT__
219 #define __FUNCT__ "TaoSetFromOptions_LMVM"
220 static PetscErrorCode TaoSetFromOptions_LMVM(Tao tao)
221 {
222   PetscErrorCode ierr;
223 
224   PetscFunctionBegin;
225   ierr = PetscOptionsHead("Limited-memory variable-metric method for unconstrained optimization");CHKERRQ(ierr);
226   ierr = TaoLineSearchSetFromOptions(tao->linesearch);CHKERRQ(ierr);
227   ierr = PetscOptionsTail();CHKERRQ(ierr);
228   PetscFunctionReturn(0);
229   PetscFunctionReturn(0);
230 }
231 
232 /*------------------------------------------------------------*/
233 #undef __FUNCT__
234 #define __FUNCT__ "TaoView_LMVM"
235 static PetscErrorCode TaoView_LMVM(Tao tao, PetscViewer viewer)
236 {
237   TAO_LMVM       *lm = (TAO_LMVM *)tao->data;
238   PetscBool      isascii;
239   PetscErrorCode ierr;
240 
241   PetscFunctionBegin;
242   ierr = PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &isascii);CHKERRQ(ierr);
243   if (isascii) {
244     ierr = PetscViewerASCIIPushTab(viewer);CHKERRQ(ierr);
245     ierr = PetscViewerASCIIPrintf(viewer, "BFGS steps: %D\n", lm->bfgs);CHKERRQ(ierr);
246     ierr = PetscViewerASCIIPrintf(viewer, "Scaled gradient steps: %D\n", lm->sgrad);CHKERRQ(ierr);
247     ierr = PetscViewerASCIIPrintf(viewer, "Gradient steps: %D\n", lm->grad);CHKERRQ(ierr);
248     ierr = PetscViewerASCIIPopTab(viewer);CHKERRQ(ierr);
249   }
250   PetscFunctionReturn(0);
251 }
252 
253 /* ---------------------------------------------------------- */
254 
255 /*MC
256      TAOLMVM - Limited Memory Variable Metric method is a quasi-Newton
257      optimization solver for unconstrained minimization. It solves
258      the Newton step
259               Hkdk = - gk
260 
261      using an approximation Bk in place of Hk, where Bk is composed using
262      the BFGS update formula. A More-Thuente line search is then used
263      to computed the steplength in the dk direction
264   Options Database Keys:
265 +     -tao_lmm_vectors - number of vectors to use for approximation
266 .     -tao_lmm_scale_type - "none","scalar","broyden"
267 .     -tao_lmm_limit_type - "none","average","relative","absolute"
268 .     -tao_lmm_rescale_type - "none","scalar","gl"
269 .     -tao_lmm_limit_mu - mu limiting factor
270 .     -tao_lmm_limit_nu - nu limiting factor
271 .     -tao_lmm_delta_min - minimum delta value
272 .     -tao_lmm_delta_max - maximum delta value
273 .     -tao_lmm_broyden_phi - phi factor for Broyden scaling
274 .     -tao_lmm_scalar_alpha - alpha factor for scalar scaling
275 .     -tao_lmm_rescale_alpha - alpha factor for rescaling diagonal
276 .     -tao_lmm_rescale_beta - beta factor for rescaling diagonal
277 .     -tao_lmm_scalar_history - amount of history for scalar scaling
278 .     -tao_lmm_rescale_history - amount of history for rescaling diagonal
279 -     -tao_lmm_eps - rejection tolerance
280 
281   Level: beginner
282 M*/
283 
284 #undef __FUNCT__
285 #define __FUNCT__ "TaoCreate_LMVM"
286 PETSC_EXTERN PetscErrorCode TaoCreate_LMVM(Tao tao)
287 {
288   TAO_LMVM       *lmP;
289   const char     *morethuente_type = TAOLINESEARCHMT;
290   PetscErrorCode ierr;
291 
292   PetscFunctionBegin;
293   tao->ops->setup = TaoSetUp_LMVM;
294   tao->ops->solve = TaoSolve_LMVM;
295   tao->ops->view = TaoView_LMVM;
296   tao->ops->setfromoptions = TaoSetFromOptions_LMVM;
297   tao->ops->destroy = TaoDestroy_LMVM;
298 
299   ierr = PetscNewLog(tao,&lmP);CHKERRQ(ierr);
300   lmP->D = 0;
301   lmP->M = 0;
302   lmP->Xold = 0;
303   lmP->Gold = 0;
304 
305   tao->data = (void*)lmP;
306   tao->max_it = 2000;
307   tao->max_funcs = 4000;
308   tao->fatol = 1e-4;
309   tao->frtol = 1e-4;
310 
311   ierr = TaoLineSearchCreate(((PetscObject)tao)->comm,&tao->linesearch);CHKERRQ(ierr);
312   ierr = TaoLineSearchSetType(tao->linesearch,morethuente_type);CHKERRQ(ierr);
313   ierr = TaoLineSearchUseTaoRoutines(tao->linesearch,tao);CHKERRQ(ierr);
314   PetscFunctionReturn(0);
315 }
316 
317