invertible_operator_mixin.py 4.2 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
# 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/>.
Theo Steininger's avatar
Theo Steininger committed
13
14
15
16
17
#
# Copyright(C) 2013-2017 Max-Planck-Society
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik
# and financially supported by the Studienstiftung des deutschen Volkes.
18

Martin Reinecke's avatar
Martin Reinecke committed
19
from builtins import object
20
21
from ...minimization import ConjugateGradient
from ...field import Field
22
23
24


class InvertibleOperatorMixin(object):
Theo Steininger's avatar
Theo Steininger committed
25
    """ Mixin class to invert implicit defined operators.
26

Theo Steininger's avatar
Theo Steininger committed
27
28
29
30
31
32
    To invert the application of a given implicitly defined operator on a
    field, this class gives the necessary functionality. Inheriting
    functionality from this class provides the derived class with the inverse
    to the given implicitly definied application of the operator on a field.
    (e.g. .inverse_times vs. .times and
    .adjoint_times vs. .adjoint_inverse_times)
33
34
35

    Parameters
    ----------
Theo Steininger's avatar
Theo Steininger committed
36
37
    inverter : Inverter
        An instance of an Inverter class.
38
39
40
        (default: ConjugateGradient)

    preconditioner : LinearOperator
Martin Reinecke's avatar
Martin Reinecke committed
41
        Preconditioner that is used by ConjugateGradient if no minimizer was
42
        given.
43
44
45
46
47
48
49
50
51
52
53
54

    Attributes
    ----------

    Raises
    ------

    Notes
    -----

    Examples
    --------
Theo Steininger's avatar
Theo Steininger committed
55
    The PropagatorOperator inherits from InvertibleOperatorMixin.
56
57
58

    See Also
    --------
Theo Steininger's avatar
Theo Steininger committed
59
    PropagatorOperator
60
61
62

    """

63
64
    def __init__(self, inverter=None, preconditioner=None,
                 forward_x0=None, backward_x0=None, *args, **kwargs):
65
66
67
68
69
        self.__preconditioner = preconditioner
        if inverter is not None:
            self.__inverter = inverter
        else:
            self.__inverter = ConjugateGradient(
70
                                        preconditioner=self.__preconditioner)
71
72
73

        self.__forward_x0 = forward_x0
        self.__backward_x0 = backward_x0
74
        super(InvertibleOperatorMixin, self).__init__(*args, **kwargs)
75

76
77
78
79
    def _times(self, x, spaces):
        if self.__forward_x0 is not None:
            x0 = self.__forward_x0
        else:
80
81
            x0 = Field(self.target, val=0., dtype=x.dtype,
                       distribution_strategy=x.distribution_strategy)
82

83
84
85
        (result, convergence) = self.__inverter(A=self.inverse_times,
                                                b=x,
                                                x0=x0)
86
87
        return result

88
89
90
91
    def _adjoint_times(self, x, spaces):
        if self.__backward_x0 is not None:
            x0 = self.__backward_x0
        else:
92
93
            x0 = Field(self.domain, val=0., dtype=x.dtype,
                       distribution_strategy=x.distribution_strategy)
94

95
96
97
        (result, convergence) = self.__inverter(A=self.adjoint_inverse_times,
                                                b=x,
                                                x0=x0)
98
99
        return result

100
101
102
103
    def _inverse_times(self, x, spaces):
        if self.__backward_x0 is not None:
            x0 = self.__backward_x0
        else:
104
105
            x0 = Field(self.domain, val=0., dtype=x.dtype,
                       distribution_strategy=x.distribution_strategy)
106

107
108
109
        (result, convergence) = self.__inverter(A=self.times,
                                                b=x,
                                                x0=x0)
110
111
        return result

112
113
114
115
    def _adjoint_inverse_times(self, x, spaces):
        if self.__forward_x0 is not None:
            x0 = self.__forward_x0
        else:
116
117
            x0 = Field(self.target, val=0., dtype=x.dtype,
                       distribution_strategy=x.distribution_strategy)
118

119
120
121
        (result, convergence) = self.__inverter(A=self.adjoint_times,
                                                b=x,
                                                x0=x0)
122
        return result