scipy_minimizer.py 4.44 KB
Newer Older
Martin Reinecke's avatar
Martin Reinecke committed
1 2 3 4 5 6 7 8 9 10 11 12 13
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
Martin Reinecke's avatar
Martin Reinecke committed
14
# Copyright(C) 2013-2018 Max-Planck-Society
Martin Reinecke's avatar
Martin Reinecke committed
15 16 17 18
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik
# and financially supported by the Studienstiftung des deutschen Volkes.

Martin Reinecke's avatar
stage 1  
Martin Reinecke committed
19
from __future__ import division
Martin Reinecke's avatar
Martin Reinecke committed
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
from .minimizer import Minimizer
from ..field import Field
from .. import dobj


class ScipyMinimizer(Minimizer):
    """Scipy-based minimizer

    Parameters
    ----------
    controller : IterationController
        Object that decides when to terminate the minimization.
    method     : str
        The selected Scipy minimization method.
    options    : dictionary
        A set of custom options for the selected minimizer.
    """

    def __init__(self, controller, method, options, need_hessp):
        super(ScipyMinimizer, self).__init__()
        if not dobj.is_numpy():
            raise NotImplementedError
        self._controller = controller
        self._method = method
        self._options = options
        self._need_hessp = need_hessp

    def __call__(self, energy):
Martin Reinecke's avatar
fix  
Martin Reinecke committed
48
        class _MinimizationDone(BaseException):
Martin Reinecke's avatar
Martin Reinecke committed
49 50 51 52 53 54 55 56 57 58 59
            pass

        class _MinHelper(object):
            def __init__(self, controller, energy):
                self._controller = controller
                self._energy = energy
                self._domain = energy.position.domain

            def _update(self, x):
                pos = Field(self._domain, x.reshape(self._domain.shape))
                if (pos.val != self._energy.position.val).any():
Martin Reinecke's avatar
Martin Reinecke committed
60
                    self._energy = self._energy.at(pos.locked_copy())
Martin Reinecke's avatar
Martin Reinecke committed
61 62 63 64 65 66 67 68 69 70
                    status = self._controller.check(self._energy)
                    if status != self._controller.CONTINUE:
                        raise _MinimizationDone

            def fun(self, x):
                self._update(x)
                return self._energy.value

            def jac(self, x):
                self._update(x)
Martin Reinecke's avatar
Martin Reinecke committed
71
                return self._energy.gradient.val.flatten()
Martin Reinecke's avatar
Martin Reinecke committed
72 73 74 75 76

            def hessp(self, x, p):
                self._update(x)
                vec = Field(self._domain, p.reshape(self._domain.shape))
                res = self._energy.curvature(vec)
Martin Reinecke's avatar
Martin Reinecke committed
77
                return res.val.flatten()
Martin Reinecke's avatar
Martin Reinecke committed
78 79 80 81 82 83 84

        import scipy.optimize as opt
        hlp = _MinHelper(self._controller, energy)
        energy = None
        status = self._controller.start(hlp._energy)
        if status != self._controller.CONTINUE:
            return hlp._energy, status
Martin Reinecke's avatar
Martin Reinecke committed
85
        x = hlp._energy.position.val.flatten()
Martin Reinecke's avatar
Martin Reinecke committed
86 87
        try:
            if self._need_hessp:
Martin Reinecke's avatar
Martin Reinecke committed
88 89 90 91
                r = opt.minimize(hlp.fun, x,
                                 method=self._method, jac=hlp.jac,
                                 hessp=hlp.hessp,
                                 options=self._options)
Martin Reinecke's avatar
Martin Reinecke committed
92
            else:
Martin Reinecke's avatar
Martin Reinecke committed
93 94 95
                r = opt.minimize(hlp.fun, x,
                                 method=self._method, jac=hlp.jac,
                                 options=self._options)
Martin Reinecke's avatar
Martin Reinecke committed
96 97 98
        except _MinimizationDone:
            status = self._controller.check(hlp._energy)
            return hlp._energy, self._controller.check(hlp._energy)
Martin Reinecke's avatar
Martin Reinecke committed
99
        if not r.success:
Martin Reinecke's avatar
stage 1  
Martin Reinecke committed
100
            dobj.mprint("Problem in Scipy minimization:", r.message)
Martin Reinecke's avatar
Martin Reinecke committed
101
        else:
Martin Reinecke's avatar
stage 1  
Martin Reinecke committed
102
            dobj.mprint("Problem in Scipy minimization")
Martin Reinecke's avatar
Martin Reinecke committed
103 104 105 106
        return hlp._energy, self._controller.ERROR


def NewtonCG(controller):
Martin Reinecke's avatar
Martin Reinecke committed
107 108 109 110 111 112
    """Returns a ScipyMinimizer object carrying out the Newton-CG algorithm.

    See Also
    --------
    ScipyMinimizer
    """
Martin Reinecke's avatar
Martin Reinecke committed
113 114 115 116 117
    return ScipyMinimizer(controller, "Newton-CG",
                          {"xtol": 1e-20, "maxiter": None}, True)


def L_BFGS_B(controller, maxcor=10):
Martin Reinecke's avatar
Martin Reinecke committed
118 119 120 121 122 123
    """Returns a ScipyMinimizer object carrying out the L-BFGS-B algorithm.

    See Also
    --------
    ScipyMinimizer
    """
Martin Reinecke's avatar
Martin Reinecke committed
124 125 126
    return ScipyMinimizer(controller, "L-BFGS-B",
                          {"ftol": 1e-20, "gtol": 1e-20, "maxcor": maxcor},
                          False)