xref: /petsc/src/tao/bound/impls/bnk/bntr.c (revision 2f75a4aaab0e3f692f7125f85b5f5852e7ceb6cb)
1 #include <../src/tao/bound/impls/bnk/bnk.h>
2 #include <petscksp.h>
3 
4 /*
5  Implements Newton's Method with a trust region approach for solving
6  bound constrained minimization problems.
7 
8  The linear system solve has to be done with a conjugate gradient method.
9 */
10 
11 static PetscErrorCode TaoSolve_BNTR(Tao tao)
12 {
13   PetscErrorCode               ierr;
14   TAO_BNK                      *bnk = (TAO_BNK *)tao->data;
15 
16   PetscReal                    oldTrust, prered, actred, stepNorm, steplen;
17   PetscBool                    stepAccepted = PETSC_TRUE;
18   PetscInt                     stepType;
19 
20   PetscFunctionBegin;
21   /*   Project the current point onto the feasible set */
22   ierr = TaoComputeVariableBounds(tao);CHKERRQ(ierr);
23   ierr = TaoLineSearchSetVariableBounds(tao->linesearch,tao->XL,tao->XU);CHKERRQ(ierr);
24 
25   /* Project the initial point onto the feasible region */
26   ierr = VecMedian(tao->XL,tao->solution,tao->XU,tao->solution);CHKERRQ(ierr);
27 
28   /* Check convergence criteria */
29   ierr = TaoComputeObjectiveAndGradient(tao, tao->solution, &bnk->f, bnk->unprojected_gradient);CHKERRQ(ierr);
30   ierr = VecBoundGradientProjection(bnk->unprojected_gradient,tao->solution,tao->XL,tao->XU,tao->gradient);CHKERRQ(ierr);
31   ierr = TaoGradientNorm(tao, tao->gradient,NORM_2,&bnk->gnorm);CHKERRQ(ierr);
32   if (PetscIsInfOrNanReal(bnk->f) || PetscIsInfOrNanReal(bnk->gnorm)) SETERRQ(PETSC_COMM_SELF,1, "User provided compute function generated Inf or NaN");
33 
34   tao->reason = TAO_CONTINUE_ITERATING;
35   ierr = TaoLogConvergenceHistory(tao,bnk->f,bnk->gnorm,0.0,tao->ksp_its);CHKERRQ(ierr);
36   ierr = TaoMonitor(tao,tao->niter,bnk->f,bnk->gnorm,0.0,tao->trust);CHKERRQ(ierr);
37   ierr = (*tao->ops->convergencetest)(tao,tao->cnvP);CHKERRQ(ierr);
38   if (tao->reason != TAO_CONTINUE_ITERATING) PetscFunctionReturn(0);
39 
40   /* Initialize the preconditioner and trust radius */
41   ierr = TaoBNKInitialize(tao);CHKERRQ(ierr);
42 
43   /* Have not converged; continue with Newton method */
44   while (tao->reason == TAO_CONTINUE_ITERATING) {
45 
46     if (stepAccepted) {
47       tao->niter++;
48       tao->ksp_its=0;
49       /* Compute the Hessian */
50       ierr = TaoComputeHessian(tao,tao->solution,tao->hessian,tao->hessian_pre);CHKERRQ(ierr);
51       /* Update the BFGS preconditioner */
52       if (BNK_PC_BFGS == bnk->pc_type) {
53         if (BFGS_SCALE_PHESS == bnk->bfgs_scale_type) {
54           /* Obtain diagonal for the bfgs preconditioner  */
55           ierr = MatGetDiagonal(tao->hessian, bnk->Diag);CHKERRQ(ierr);
56           ierr = VecAbs(bnk->Diag);CHKERRQ(ierr);
57           ierr = VecReciprocal(bnk->Diag);CHKERRQ(ierr);
58           ierr = MatLMVMSetScale(bnk->M,bnk->Diag);CHKERRQ(ierr);
59         }
60         /* Update the limited memory preconditioner and get existing # of updates */
61         ierr = MatLMVMUpdate(bnk->M, tao->solution, bnk->unprojected_gradient);CHKERRQ(ierr);
62       }
63     }
64 
65     /* Use the common BNK kernel to compute the Newton step (for inactive variables only) */
66     ierr = TaoBNKComputeStep(tao, PETSC_FALSE, &stepType);CHKERRQ(ierr);
67 
68     /* Store current solution before it changes */
69     oldTrust = tao->trust;
70     bnk->fold = bnk->f;
71     ierr = VecCopy(tao->solution, bnk->Xold);CHKERRQ(ierr);
72     ierr = VecCopy(tao->gradient, bnk->Gold);CHKERRQ(ierr);
73     ierr = VecCopy(bnk->unprojected_gradient, bnk->unprojected_gradient_old);CHKERRQ(ierr);
74 
75     /* Temporarily accept the step and project it into the bounds */
76     ierr = VecAXPY(tao->solution, 1.0, tao->stepdirection);CHKERRQ(ierr);
77     ierr = VecMedian(tao->XL, tao->solution, tao->XU, tao->solution);CHKERRQ(ierr);
78 
79     /* Check if the projection changed the step direction */
80     ierr = VecCopy(tao->solution, tao->stepdirection);CHKERRQ(ierr);
81     ierr = VecAXPY(tao->stepdirection, -1.0, bnk->Xold);CHKERRQ(ierr);
82     ierr = VecNorm(tao->stepdirection, NORM_2, &stepNorm);CHKERRQ(ierr);
83     if (stepNorm != bnk->dnorm) {
84       /* Projection changed the step, so we have to recompute predicted reduction.
85          However, we deliberately do not change the step norm and the trust radius
86          in order for the safeguard to more closely mimic a piece-wise linesearch
87          along the bounds. */
88       ierr = MatMult(tao->hessian, tao->stepdirection, bnk->Xwork);CHKERRQ(ierr);
89       ierr = VecAYPX(bnk->Xwork, -0.5, tao->gradient);CHKERRQ(ierr);
90       ierr = VecDot(bnk->Xwork, tao->stepdirection, &prered);
91     } else {
92       /* Step did not change, so we can just recover the pre-computed prediction */
93       ierr = KSPCGGetObjFcn(tao->ksp, &prered);CHKERRQ(ierr);
94     }
95     prered = -prered;
96 
97     /* Compute the actual reduction and update the trust radius */
98     ierr = TaoComputeObjective(tao, tao->solution, &bnk->f);CHKERRQ(ierr);
99     actred = bnk->fold - bnk->f;
100     ierr = TaoBNKUpdateTrustRadius(tao, prered, actred, stepType, &stepAccepted);CHKERRQ(ierr);
101 
102     if (stepAccepted) {
103       /* Step is good, evaluate the gradient and the hessian */
104       steplen = 1.0;
105       ierr = TaoComputeGradient(tao, tao->solution, bnk->unprojected_gradient);CHKERRQ(ierr);
106       ierr = VecBoundGradientProjection(bnk->unprojected_gradient,tao->solution,tao->XL,tao->XU,tao->gradient);CHKERRQ(ierr);
107     } else {
108       /* Step is bad, revert old solution and re-solve with new radius*/
109       steplen = 0.0;
110       bnk->f = bnk->fold;
111       ierr = VecCopy(bnk->Xold, tao->solution);CHKERRQ(ierr);
112       ierr = VecCopy(bnk->Gold, tao->gradient);CHKERRQ(ierr);
113       ierr = VecCopy(bnk->unprojected_gradient_old, bnk->unprojected_gradient);CHKERRQ(ierr);
114       if (oldTrust == tao->trust) {
115         /* Can't change the radius anymore so just terminate */
116         tao->reason = TAO_DIVERGED_TR_REDUCTION;
117       }
118     }
119 
120     /*  Check for termination */
121     ierr = TaoGradientNorm(tao, tao->gradient, NORM_2, &bnk->gnorm);CHKERRQ(ierr);
122     if (PetscIsInfOrNanReal(bnk->gnorm)) SETERRQ(PETSC_COMM_SELF,1,"User provided compute function generated Not-a-Number");
123     ierr = TaoLogConvergenceHistory(tao,bnk->f,bnk->gnorm,0.0,tao->ksp_its);CHKERRQ(ierr);
124     ierr = TaoMonitor(tao,tao->niter,bnk->f,bnk->gnorm,0.0,steplen);CHKERRQ(ierr);
125     ierr = (*tao->ops->convergencetest)(tao,tao->cnvP);CHKERRQ(ierr);
126   }
127   PetscFunctionReturn(0);
128 }
129 
130 /*------------------------------------------------------------*/
131 
132 PETSC_EXTERN PetscErrorCode TaoCreate_BNTR(Tao tao)
133 {
134   TAO_BNK        *bnk;
135   PetscErrorCode ierr;
136 
137   PetscFunctionBegin;
138   ierr = TaoCreate_BNK(tao);CHKERRQ(ierr);
139   tao->ops->solve=TaoSolve_BNTR;
140 
141   bnk = (TAO_BNK *)tao->data;
142   bnk->update_type = BNK_UPDATE_REDUCTION; /* trust region updates based on predicted/actual reduction */
143   bnk->sval = 0.0; /* disable Hessian shifting */
144   PetscFunctionReturn(0);
145 }