"""
Copyright 2013 Steven Diamond, 2022 - the CVXPY Authors
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
"""
from __future__ import annotations
import warnings
from typing import List, Tuple, TypeVar
import numpy as np
from cvxpy.constraints.cones import Cone
from cvxpy.expressions import cvxtypes
from cvxpy.utilities import scopes
Expression = TypeVar('Expression')
[docs]
class ExpCone(Cone):
"""A reformulated exponential cone constraint.
Operates elementwise on :math:`x, y, z`.
Original cone:
.. math::
K = \\{(x,y,z) \\mid y > 0, ye^{x/y} <= z\\}
\\cup \\{(x,y,z) \\mid x \\leq 0, y = 0, z \\geq 0\\}
Reformulated cone:
.. math::
K = \\{(x,y,z) \\mid y, z > 0, y\\log(y) + x \\leq y\\log(z)\\}
\\cup \\{(x,y,z) \\mid x \\leq 0, y = 0, z \\geq 0\\}
Parameters
----------
x : Expression
x in the exponential cone.
y : Expression
y in the exponential cone.
z : Expression
z in the exponential cone.
"""
def __init__(self, x: Expression, y: Expression, z: Expression, constr_id=None) -> None:
Expression = cvxtypes.expression()
self.x = Expression.cast_to_const(x)
self.y = Expression.cast_to_const(y)
self.z = Expression.cast_to_const(z)
args = [self.x, self.y, self.z]
for val in args:
if not (val.is_affine() and val.is_real()):
raise ValueError('All arguments must be affine and real.')
xs, ys, zs = self.x.shape, self.y.shape, self.z.shape
if xs != ys or xs != zs:
msg = ("All arguments must have the same shapes. Provided arguments have"
"shapes %s" % str((xs, ys, zs)))
raise ValueError(msg)
super(ExpCone, self).__init__(args, constr_id)
def __str__(self) -> str:
return "ExpCone(%s, %s, %s)" % (self.x, self.y, self.z)
def __repr__(self) -> str:
return "ExpCone(%s, %s, %s)" % (self.x, self.y, self.z)
@property
def residual(self):
# TODO(akshayka): The projection should be implemented directly.
from cvxpy import Minimize, Problem, Variable, hstack, norm2
if self.x.value is None or self.y.value is None or self.z.value is None:
return None
x = Variable(self.x.shape)
y = Variable(self.y.shape)
z = Variable(self.z.shape)
constr = [ExpCone(x, y, z)]
obj = Minimize(norm2(hstack([x, y, z]) -
hstack([self.x.value, self.y.value, self.z.value])))
problem = Problem(obj, constr)
return problem.solve()
@property
def size(self) -> int:
"""The number of entries in the combined cones.
"""
return 3 * self.num_cones()
def num_cones(self):
"""The number of elementwise cones.
"""
return self.x.size
def as_quad_approx(self, m: int, k: int) -> RelEntrConeQuad:
return RelEntrConeQuad(self.y, self.z, -self.x, m, k)
def cone_sizes(self) -> List[int]:
"""The dimensions of the exponential cones.
Returns
-------
list
A list of the sizes of the elementwise cones.
"""
return [3]*self.num_cones()
[docs]
def is_dcp(self, dpp: bool = False) -> bool:
"""An exponential constraint is DCP if each argument is affine.
"""
if dpp:
with scopes.dpp_scope():
return all(arg.is_affine() for arg in self.args)
return all(arg.is_affine() for arg in self.args)
def is_dgp(self, dpp: bool = False) -> bool:
return False
def is_dqcp(self) -> bool:
return self.is_dcp()
@property
def shape(self) -> Tuple[int, ...]:
s = (3,) + self.x.shape
return s
def save_dual_value(self, value) -> None:
# TODO(akshaya,SteveDiamond): verify that reshaping below works correctly
value = np.reshape(value, newshape=(-1, 3))
dv0 = np.reshape(value[:, 0], newshape=self.x.shape)
dv1 = np.reshape(value[:, 1], newshape=self.y.shape)
dv2 = np.reshape(value[:, 2], newshape=self.z.shape)
self.dual_variables[0].save_value(dv0)
self.dual_variables[1].save_value(dv1)
self.dual_variables[2].save_value(dv2)
def _dual_cone(self, *args):
"""Implements the dual cone of the exponential cone
See Pg 85 of the MOSEK modelling cookbook for more information"""
if args == ():
return ExpCone(-self.dual_variables[1], -self.dual_variables[0],
np.exp(1) * self.dual_variables[2])
else:
# some assertions for verifying `args`
args_shapes = [arg.shape for arg in args]
instance_args_shapes = [arg.shape for arg in self.args]
assert len(args) == len(self.args)
assert args_shapes == instance_args_shapes
return ExpCone(-args[1], -args[0],
np.exp(1)*args[2])
[docs]
class RelEntrConeQuad(Cone):
"""An approximate construction of the scalar relative entropy cone
Definition:
.. math::
K_{re}=\\text{cl}\\{(x,y,z)\\in\\mathbb{R}_{++}\\times
\\mathbb{R}_{++}\\times\\mathbb{R}_{++}\\:x\\log(x/y)\\leq z\\}
Since the above definition is very similar to the ExpCone, we provide a conversion method.
More details on the approximation can be found in Theorem-3 on page-10 in the paper:
Semidefinite Approximations of the Matrix Logarithm.
Parameters
----------
x : Expression
x in the (approximate) scalar relative entropy cone
y : Expression
y in the (approximate) scalar relative entropy cone
z : Expression
z in the (approximate) scalar relative entropy cone
m: Parameter directly related to the number of generated nodes for the quadrature
approximation used in the algorithm
k: Another parameter controlling the approximation
"""
def __init__(self, x: Expression, y: Expression, z: Expression,
m: int, k: int, constr_id=None) -> None:
Expression = cvxtypes.expression()
self.x = Expression.cast_to_const(x)
self.y = Expression.cast_to_const(y)
self.z = Expression.cast_to_const(z)
args = [self.x, self.y, self.z]
for val in args:
if not (val.is_affine() and val.is_real()):
raise ValueError('All Expression arguments must be affine and real.')
self.m = m
self.k = k
xs, ys, zs = self.x.shape, self.y.shape, self.z.shape
if xs != ys or xs != zs:
msg = ("All arguments must have the same shapes. Provided arguments have"
"shapes %s" % str((xs, ys, zs)))
raise ValueError(msg)
super(RelEntrConeQuad, self).__init__([self.x, self.y, self.z], constr_id)
def get_data(self):
return [self.m, self.k, self.id]
def __str__(self) -> str:
tup = (self.x, self.y, self.z, self.m, self.k)
return "RelEntrConeQuad(%s, %s, %s, %s, %s)" % tup
def __repr__(self) -> str:
return self.__str__()
@property
def residual(self):
# TODO(akshayka): The projection should be implemented directly.
from cvxpy import Minimize, Problem, Variable, hstack, norm2
if self.x.value is None or self.y.value is None or self.z.value is None:
return None
cvxtypes.expression()
x = Variable(self.x.shape)
y = Variable(self.y.shape)
z = Variable(self.z.shape)
constr = [RelEntrConeQuad(x, y, z, self.m, self.k)]
obj = Minimize(norm2(hstack([x, y, z]) -
hstack([self.x.value, self.y.value, self.z.value])))
problem = Problem(obj, constr)
return problem.solve()
@property
def size(self) -> int:
"""The number of entries in the combined cones.
"""
return 3 * self.num_cones()
def num_cones(self):
"""The number of elementwise cones.
"""
return self.x.size
def cone_sizes(self) -> List[int]:
"""The dimensions of the exponential cones.
Returns
-------
list
A list of the sizes of the elementwise cones.
"""
return [3]*self.num_cones()
[docs]
def is_dcp(self, dpp: bool = False) -> bool:
"""An exponential constraint is DCP if each argument is affine.
"""
if dpp:
with scopes.dpp_scope():
return all(arg.is_affine() for arg in self.args)
return all(arg.is_affine() for arg in self.args)
def is_dgp(self, dpp: bool = False) -> bool:
return False
def is_dqcp(self) -> bool:
return self.is_dcp()
@property
def shape(self) -> Tuple[int, ...]:
s = (3,) + self.x.shape
return s
def save_dual_value(self, value) -> None:
# TODO: implement me.
pass
[docs]
class OpRelEntrConeQuad(Cone):
"""An approximate construction of the operator relative entropy cone
Definition:
.. math::
K_{re}^n=\\text{cl}\\{(X,Y,T)\\in\\mathbb{H}^n_{++}\\times
\\mathbb{H}^n_{++}\\times\\mathbb{H}^n_{++}\\:D_{\\text{op}}\\succeq T\\}
More details on the approximation can be found in Theorem-3 on page-10 in the paper:
Semidefinite Approximations of the Matrix Logarithm.
Parameters
----------
X : Expression
x in the (approximate) operator relative entropy cone
Y : Expression
y in the (approximate) operator relative entropy cone
Z : Expression
Z in the (approximate) operator relative entropy cone
m: int
Must be positive. Controls the number of quadrature nodes used in a local
approximation of the matrix logarithm. Increasing this value results in
better local approximations, but does not significantly expand the region
of inputs for which the approximation is effective.
k: int
Must be positive. Sets the number of scaling points about which the
quadrature approximation is performed. Increasing this value will
expand the region of inputs over which the approximation is effective.
This approximation uses :math:`m + k` semidefinite constraints.
"""
def __init__(self, X: Expression, Y: Expression, Z: Expression,
m: int, k: int, constr_id=None) -> None:
Expression = cvxtypes.expression()
self.X = Expression.cast_to_const(X)
self.Y = Expression.cast_to_const(Y)
self.Z = Expression.cast_to_const(Z)
if (not X.is_hermitian()) or (not Y.is_hermitian()) or (not Z.is_hermitian()):
msg = ("One of the input matrices has not explicitly been declared as symmetric or"
"Hermitian. If the inputs are Variable objects, try declaring them with the"
"symmetric=True or Hermitian=True properties. If the inputs are general "
"Expression objects that are known to be symmetric or Hermitian, then you"
"can wrap them with the symmetric_wrap and hermitian_wrap atoms. Failure to"
"do one of these things will cause this function to impose a symmetry or"
"conjugate-symmetry constraint internally, in a way that is very"
"inefficient.")
warnings.warn(msg)
self.m = m
self.k = k
Xs, Ys, Zs = self.X.shape, self.Y.shape, self.Z.shape
if Xs != Ys or Xs != Zs:
msg = ("All arguments must have the same shapes. Provided arguments have"
"shapes %s" % str((Xs, Ys, Zs)))
raise ValueError(msg)
super(OpRelEntrConeQuad, self).__init__([self.X, self.Y, self.Z], constr_id)
def get_data(self):
return [self.m, self.k, self.id]
def __str__(self) -> str:
tup = (self.X, self.Y, self.Z, self.m, self.k)
return "OpRelEntrConeQuad(%s, %s, %s, %s, %s)" % tup
def __repr__(self) -> str:
return self.__str__()
@property
def residual(self):
# TODO: implement me
raise NotImplementedError()
@property
def size(self) -> int:
"""The number of entries in the combined cones.
"""
return 3 * self.num_cones()
def num_cones(self):
"""The number of elementwise cones.
"""
return self.X.size
def cone_sizes(self) -> List[int]:
"""The dimensions of the exponential cones.
Returns
-------
list
A list of the sizes of the elementwise cones.
"""
return [3]*self.num_cones()
[docs]
def is_dcp(self, dpp: bool = False) -> bool:
"""An operator relative conic constraint is DCP when (A, b, C) is affine
"""
if dpp:
with scopes.dpp_scope():
return all(arg.is_affine() for arg in self.args)
return all(arg.is_affine() for arg in self.args)
def is_dgp(self, dpp: bool = False) -> bool:
return False
def is_dqcp(self) -> bool:
return self.is_dcp()
@property
def shape(self) -> Tuple[int, ...]:
s = (3,) + self.X.shape
return s
def save_dual_value(self, value) -> None:
# TODO: implement me.
pass