Actual source code: owarmijo.h

petsc-3.10.5 2019-03-28
Report Typos and Errors
  1: #ifndef __TAOLINESEARCH_OWARMIJO_H

  4: /* Context for an Armijo (nonmonotone) linesearch for orthant wise unconstrained
  5:  minimization.

  7:  Given a function f, the current iterate x, and a descent direction d:
  8:  Find the smallest i in 0, 1, 2, ..., such that:

 10:     f(x + (beta**i)d) <= f(x) + (sigma*beta**i)<grad f(x),d>

 12:  The nonmonotone modification of this linesearch replaces the f(x) term
 13:  with a reference value, R, and seeks to find the smallest i such that:

 15:     f(x + (beta**i)d) <= R + (sigma*beta**i)<grad f(x),d>

 17:  This modification does effect neither the convergence nor rate of
 18:  convergence of an algorithm when R is chosen appropriately.  Essentially,
 19:  R must decrease on average in some sense.  The benefit of a nonmonotone
 20:  linesearch is that local minimizers can be avoided (by allowing increase
 21:  in function value), and typically, fewer iterations are performed in
 22:  the main code.

 24:  The reference value is chosen based upon some historical information
 25:  consisting of function values for previous iterates.  The amount of
 26:  historical information used is determined by the memory size where the
 27:  memory is used to store the previous function values.  The memory is
 28:  initialized to alpha*f(x^0) for some alpha >= 1, with alpha=1 signifying
 29:  that we always force decrease from the initial point.

 31:  The reference value can be the maximum value in the memory or can be
 32:  chosen to provide some mean descent.  Elements are removed from the
 33:  memory with a replacement policy that either removes the oldest
 34:  value in the memory (FIFO), or the largest value in the memory (MRU).

 36:  Additionally, we can add a watchdog strategy to the search, which
 37:  essentially accepts small directions and only checks the nonmonotonic
 38:  descent criteria every m-steps.  This strategy is NOT implemented in
 39:  the code.

 41:  Finally, care must be taken when steepest descent directions are used.
 42:  For example, when the Newton direction is not not satisfy a sufficient
 43:  descent criteria.  The code will apply the same test regardless of
 44:  the direction.  This type of search may not be appropriate for all
 45:  algorithms.  For example, when a gradient direction is used, we may
 46:  want to revert to the best point found and reset the memory so that
 47:  we stay in an appropriate level set after using a gradient steps.
 48:  This type of search is currently NOT supported by the code.

 50:  References:
 51:   Armijo, "Minimization of Functions Having Lipschitz Continuous
 52:     First-Partial Derivatives," Pacific Journal of Mathematics, volume 16,
 53:     pages 1-3, 1966.
 54:   Ferris and Lucidi, "Nonmonotone Stabilization Methods for Nonlinear
 55:     Equations," Journal of Optimization Theory and Applications, volume 81,
 56:     pages 53-71, 1994.
 57:   Grippo, Lampariello, and Lucidi, "A Nonmonotone Line Search Technique
 58:     for Newton's Method," SIAM Journal on Numerical Analysis, volume 23,
 59:     pages 707-716, 1986.
 60:   Grippo, Lampariello, and Lucidi, "A Class of Nonmonotone Stabilization
 61:     Methods in Unconstrained Optimization," Numerische Mathematik, volume 59,
 62:   pages 779-805, 1991. */
 63:  #include <petsc/private/taolinesearchimpl.h>
 64: typedef struct {
 65:   PetscReal *memory;

 67:   PetscReal alpha;                      /* Initial reference factor >= 1 */
 68:   PetscReal beta;                       /* Steplength determination < 1 */
 69:   PetscReal beta_inf;                   /* Steplength determination < 1 */
 70:   PetscReal sigma;                      /* Acceptance criteria < 1) */
 71:   PetscReal minimumStep;                /* Minimum step size */
 72:   PetscReal lastReference;              /* Reference value of last iteration */

 74:   PetscInt memorySize;                  /* Number of functions kept in memory */
 75:   PetscInt current;                     /* Current element for FIFO */
 76:   PetscInt referencePolicy;             /* Integer for reference calculation rule */
 77:   PetscInt replacementPolicy;           /* Policy for replacing values in memory */

 79:   PetscBool nondescending;
 80:   PetscBool memorySetup;

 82:   Vec x;        /* Maintain reference to variable vector to check for changes */
 83:   Vec work;
 84: } TaoLineSearch_OWARMIJO;

 86: static PetscErrorCode ProjWork_OWLQN(Vec w,Vec x,Vec gv,PetscReal *gdx);

 88: #endif