def vstack(seq):
"""
| Stacks objects vertically.
:param seq: tuple or list of numbers,
:ref:`scalar objects<scalar_ref>` or
:ref:`multidimensional objects<multi_ref>`.
:return: :ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
"""
# Check input
if (type(seq) is not tuple and type(seq) is not list):
raise TypeError('Invalid argument')
# Process input
new_list = []
numeric = True
for x in seq:
if np.isscalar(x):
new_list += [matrix(x)]
elif type(x) is cvxpy_obj:
new_list += [matrix(x.value)]
elif type(x).__name__ in SCALAR_OBJS:
numeric = False
new_x = cvxpy_array(1, 1)
new_x[0, 0] = x
new_list += [new_x]
elif type(x).__name__ in ARRAY_OBJS:
numeric = False
new_list += [x]
elif type(x) is cvxpy_matrix:
new_list += [x]
else:
raise TypeError('Invalid Input')
# Input is numeric
if numeric:
return np.vstack(new_list)
# Verify dimensions
n = new_list[0].shape[1]
for x in new_list:
if x.shape[1] != n:
raise ValueError('Invalid Dimensions')
# Allocate new array
m = int(np.sum([x.shape[0] for x in new_list]))
new_ar = cvxpy_array(m, n)
# Fill new array
k = 0
for x in new_list:
for i in range(0, x.shape[0], 1):
for j in range(0, x.shape[1], 1):
new_ar[i + k, j] = x[i, j]
k = k + x.shape[0]
# Return new array
return new_ar
def diag(v):
"""
| Extracts diagonal or constructs a diagonal array.
:param v: number,
:ref:`scalar object<scalar_ref>`,
square or one dimensional
:ref:`multidimensional object<multi_ref>`.
:return: number,
:ref:`scalar object<scalar_ref>` or
:ref:`multidimensional object<multi_ref>`.
"""
# Check input
if (np.isscalar(v) or
type(v).__name__ in SCALAR_OBJS):
return v
elif (type(v) is cvxpy_matrix or
type(v).__name__ in ARRAY_OBJS):
if v.shape[1]==1:
v = v.T
else:
raise TypeError('Invalid argument type')
# Get shape
(m,n) = v.shape
# Input is 2D
if m!=1 and n!=1:
# Must be square
if m!=n:
raise ValueError('Invalid dimensions')
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.diag(v)).T
# Object array
else:
new_ar = cvxpy_array(m,1)
for i in range(0,m,1):
new_ar[i,0] = v[i,i]
return new_ar
# Input is 1D
else:
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.diagflat(v))
# Object array
else:
new_ar = cvxpy_array(n,n)
for i in range(0,n,1):
new_ar[i,i] = v[0,i]
return new_ar
def diag(v):
"""
| Extracts diagonal or constructs a diagonal array.
:param v: number,
:ref:`scalar object<scalar_ref>`,
square or one dimensional
:ref:`multidimensional object<multi_ref>`.
:return: number,
:ref:`scalar object<scalar_ref>` or
:ref:`multidimensional object<multi_ref>`.
"""
# Check input
if (np.isscalar(v) or type(v).__name__ in SCALAR_OBJS):
return v
elif (type(v) is cvxpy_matrix or type(v).__name__ in ARRAY_OBJS):
if v.shape[1] == 1:
v = v.T
else:
raise TypeError('Invalid argument type')
# Get shape
(m, n) = v.shape
# Input is 2D
if m != 1 and n != 1:
# Must be square
if m != n:
raise ValueError('Invalid dimensions')
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.diag(v)).T
# Object array
else:
new_ar = cvxpy_array(m, 1)
for i in range(0, m, 1):
new_ar[i, 0] = v[i, i]
return new_ar
# Input is 1D
else:
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.diagflat(v))
# Object array
else:
new_ar = cvxpy_array(n, n)
for i in range(0, n, 1):
new_ar[i, i] = v[0, i]
return new_ar
def vstack(t):
"""
Vertical stack.
"""
# Verify input type
new_list = []
numeric = True
for x in t:
if(np.isscalar(x)):
new_list += [matrix(x)]
elif(type(x) is cvxpy_obj):
new_list += [matrix(x.data)]
elif(type(x).__name__ in SCALAR_OBJS):
numeric = False
new_x = cvxpy_expression(1,1)
new_x[0,0] = x
new_list += [new_x]
elif(type(x).__name__ in ARRAY_OBJS):
numeric = False
new_list += [x]
elif(type(x) is cvxpy_matrix):
new_list += [x]
else:
raise ValueError('Invalid Input')
# Input is numeric
if(numeric):
return np.vstack(new_list)
# Verify dimensions
n = new_list[0].shape[1]
for x in new_list:
if(x.shape[1] != n):
raise ValueError('Invalid Dimensions')
# Allocate new expression
m = 0
for x in new_list:
m += x.shape[0]
new_exp = cvxpy_expression(m,n)
# Fill new expression
k = 0
for x in new_list:
for i in range(0,x.shape[0],1):
for j in range(0,x.shape[1],1):
new_exp[i+k,j] = x[i,j]
k = k + x.shape[0]
# Return new expression
return new_exp
def vstack(t):
"""
Vertical stack.
"""
# Verify input type
new_list = []
numeric = True
for x in t:
if (np.isscalar(x)):
new_list += [matrix(x)]
elif (type(x) is cvxpy_obj):
new_list += [matrix(x.data)]
elif (type(x).__name__ in SCALAR_OBJS):
numeric = False
new_x = cvxpy_expression(1, 1)
new_x[0, 0] = x
new_list += [new_x]
elif (type(x).__name__ in ARRAY_OBJS):
numeric = False
new_list += [x]
elif (type(x) is cvxpy_matrix):
new_list += [x]
else:
raise ValueError('Invalid Input')
# Input is numeric
if (numeric):
return np.vstack(new_list)
# Verify dimensions
n = new_list[0].shape[1]
for x in new_list:
if (x.shape[1] != n):
raise ValueError('Invalid Dimensions')
# Allocate new expression
m = 0
for x in new_list:
m += x.shape[0]
new_exp = cvxpy_expression(m, n)
# Fill new expression
k = 0
for x in new_list:
for i in range(0, x.shape[0], 1):
for j in range(0, x.shape[1], 1):
new_exp[i + k, j] = x[i, j]
k = k + x.shape[0]
# Return new expression
return new_exp
def diagflat(arg):
"""
List, row or column vector to diagonal matrix.
"""
# Argument is a list
if (type(arg) is list):
arg = hstack(arg)
# Argument is a matrix or object array
elif (type(arg) is cvxpy_matrix or type(arg).__name__ in ARRAY_OBJS):
(m, n) = arg.shape
if (m != 1 and n != 1):
raise ValueError('Argument must be one dimensional')
elif (n == 1):
arg = arg.T
# Invalid argument
else:
raise ValueError('Invalid argument')
# Argument is numeric
numeric = True
for i in range(0, arg.shape[1], 1):
if (not np.isscalar(arg[0, i])):
numeric = False
if (numeric):
return matrix(np.diagflat(arg))
# Not numeric
(m, n) = arg.shape
new_exp = cvxpy_expression(n, n)
for i in range(0, n, 1):
new_exp[i, i] = arg[0, i]
return new_exp
def eye(n):
"""
| See **numpy.eye**.
:param n: rows.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.eye(n))
def eye(n):
"""
| See **numpy.eye**.
:param n: rows.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.eye(n))
def sqrtm(x):
"""
| See **scipy.linalg.sqrtm**.
:param x: positive semidefinite matrix.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.real(sci_sqrtm(x)))
def sqrtm(x):
"""
| See **scipy.linalg.sqrtm**.
:param x: positive semidefinite matrix.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.real(sci_sqrtm(x)))
def ones(shape):
"""
| See **numpy.ones**.
:param shape: integer or
pair of integers.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.ones(shape))
def rand(m, n):
"""
| See **numpy.random.rand**.
:param m: rows.
:param n: columns.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.random.rand(m, n))
def rand(m,n):
"""
| See **numpy.random.rand**.
:param m: rows.
:param n: columns.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.random.rand(m,n))
def ones(shape):
"""
| See **numpy.ones**.
:param shape: integer or
pair of integers.
:return: :ref:`matrix<matrix_obj>`.
"""
return matrix(np.ones(shape))
def reshape(v, newshape):
"""
| Reshapes the array to dimensions newshape (see np.reshape)
| in FORTRAN (column-major) order.
:param v: :ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
:param newshape: tuple with two integers.
:return: the reshaped variable or matrix,
:ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
"""
# Check input
if (type(v) is cvxpy_matrix or
type(v).__name__ in ARRAY_OBJS):
pass
else:
raise TypeError('Invalid argument type')
if not (hasattr(newshape, '__iter__') and
len(newshape) == 2):
raise TypeError('Invalid argument for newshape')
# Get shape
(m,n) = v.shape
# Get new shape
(mn,nn) = newshape
if mn*nn != m*n:
raise ValueError('Output dimension size does not match input dimension')
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.reshape(v, newshape, order='F'))
# Object array
else:
new_ar = cvxpy_array(mn,nn)
for j in range(0,n,1):
for i in range(0,m,1):
k = j*m + i
new_ar[k % mn,k / mn] = v[i,j]
return new_ar
def reshape(v, newshape):
"""
| Reshapes the array to dimensions newshape (see np.reshape)
| in FORTRAN (column-major) order.
:param v: :ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
:param newshape: tuple with two integers.
:return: the reshaped variable or matrix,
:ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
"""
# Check input
if (type(v) is cvxpy_matrix or type(v).__name__ in ARRAY_OBJS):
pass
else:
raise TypeError('Invalid argument type')
if not (hasattr(newshape, '__iter__') and len(newshape) == 2):
raise TypeError('Invalid argument for newshape')
# Get shape
(m, n) = v.shape
# Get new shape
(mn, nn) = newshape
if mn * nn != m * n:
raise ValueError(
'Output dimension size does not match input dimension')
# cvxpy matrix
if type(v) is cvxpy_matrix:
return matrix(np.reshape(v, newshape, order='F'))
# Object array
else:
new_ar = cvxpy_array(mn, nn)
for j in range(0, n, 1):
for i in range(0, m, 1):
k = j * m + i
new_ar[k % mn, k / mn] = v[i, j]
return new_ar
def diag(arg):
"""
Extract the diagonal from square array-like object.
"""
# Check size
(m,n) = arg.shape
if(m!=n):
raise ValueError('Invalid dimensions')
# cvxpy matrix
if(type(arg) is cvxpy_matrix):
return matrix(np.diag(arg)).T
# Object array
elif(type(arg).__name__ in ARRAY_OBJS):
new_exp = cvxpy_expression(m,1)
for i in range(0,m,1):
new_exp[i,0] = arg[i,i]
return new_exp
# Error
else:
raise ValueError('Invalid argument type')
def diag(arg):
"""
Extract the diagonal from square array-like object.
"""
# Check size
(m, n) = arg.shape
if (m != n):
raise ValueError('Invalid dimensions')
# cvxpy matrix
if (type(arg) is cvxpy_matrix):
return matrix(np.diag(arg)).T
# Object array
elif (type(arg).__name__ in ARRAY_OBJS):
new_exp = cvxpy_expression(m, 1)
for i in range(0, m, 1):
new_exp[i, 0] = arg[i, i]
return new_exp
# Error
else:
raise ValueError('Invalid argument type')
def diagflat(arg):
"""
List, row or column vector to diagonal matrix.
"""
# Argument is a list
if(type(arg) is list):
arg = hstack(arg)
# Argument is a matrix or object array
elif(type(arg) is cvxpy_matrix or
type(arg).__name__ in ARRAY_OBJS):
(m,n) = arg.shape
if(m!=1 and n!=1):
raise ValueError('Argument must be one dimensional')
elif(n == 1):
arg = arg.T
# Invalid argument
else:
raise ValueError('Invalid argument')
# Argument is numeric
numeric = True
for i in range(0,arg.shape[1],1):
if(not np.isscalar(arg[0,i])):
numeric = False
if(numeric):
return matrix(np.diagflat(arg))
# Not numeric
(m,n) = arg.shape
new_exp = cvxpy_expression(n,n)
for i in range(0,n,1):
new_exp[i,i] = arg[0,i]
return new_exp
def zeros(shape):
"""
Matrix of zeros.
"""
return matrix(np.zeros(shape))
def ones(shape):
"""
Matrix of ones.
"""
return matrix(np.ones(shape))
def zeros(shape):
"""
Matrix of zeros.
"""
return matrix(np.zeros(shape))
def eye(n):
return matrix(np.eye(n))
def rand(m, n):
return matrix(np.random.rand(m, n))
def solve_prog(p):
"""
Description
-----------
Solve optimization program.
Arguments
---------
p: cvxpy_program
"""
# Reset variable counter
var_reset()
# Check parameters
if (len(p.get_params()) != 0):
if (not p.options['quiet']):
print 'Error: Parameters present'
return np.NaN, np.NaN
# Original
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Original *'
print '****************************'
p.show()
# Expanded
p_expanded = p._get_expanded()
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Expanded *'
print '****************************'
p_expanded.show()
# Equivalent
p_equivalent = p_expanded._get_equivalent()
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Equivalent *'
print '****************************'
p_equivalent.show()
# Nonconvex constraints
bad_constr = p_equivalent.constr._get_nonconvex()
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Nonconvex Constraints *'
print '****************************\n'
print bad_constr
# Convex Relaxation
p_cvx_relaxation = p_equivalent._get_cvx_relaxation()
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Convex Relaxation *'
print '****************************'
p_cvx_relaxation.show()
# Solve convex relaxation
if (not p.options['quiet']):
print '\nSolving convex relaxation ...'
sol = solve_convex(p_cvx_relaxation, 'rel')
lagrange_mul_eq = matrix(sol['y'])
if (not p.options['quiet']):
print 'Relaxation status: ', sol['status']
# Relaxation status: primal infeasible
if (sol['status'] == 'primal infeasible'):
if (not p.options['quiet']):
print 'Original program is primal infeasible'
if (p.action == MINIMIZE):
return np.inf, lagrange_mul_eq
else:
return -np.inf, lagrange_mul_eq
# Relaxation status: uknown
elif (sol['status'] == 'unknown'):
relaxation_obj = None
# Relaxation status: dual infeasible
elif (sol['status'] == 'dual infeasible'):
if (p.action == MINIMIZE):
relaxation_obj = -np.inf
else:
relaxation_obj = np.inf
# Relaxacion status: optimal
else:
relaxation_obj = p_cvx_relaxation.obj.get_value()
# Report max slack
if (len(bad_constr) != 0 and not p.options['quiet']):
print 'Max Slack: ',
print max(
map(lambda x: x.get_value(),
[abs(c.left - c.right) for c in bad_constr]))
# Tighten
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Tightening Stage *'
print '****************************'
elif (not p.options['quiet']):
print '\nAttempting tightening ...'
valid_scp = scp(p_cvx_relaxation, bad_constr, sol)
if (not valid_scp):
return np.NaN, np.NaN
# Prepare results
if (len(bad_constr) == 0):
obj = relaxation_obj
else:
obj = p.obj.get_value()
if (p.action == MINIMIZE):
bound_str = 'Lower Bound'
if (relaxation_obj != None):
bound = np.min([relaxation_obj, obj])
gap = np.max([obj - relaxation_obj, 0.0])
else:
bound_str = 'Upper Bound'
if (relaxation_obj != None):
bound = np.max([relaxation_obj, obj])
gap = np.max([relaxation_obj - obj, 0.0])
# Show results
if (p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Results *'
print '****************************'
elif (not p.options['quiet']):
print '\nResults'
if (not p.options['quiet']):
print 'Objective =\t%.5e' % obj
if (relaxation_obj != None):
print bound_str + ' =\t%.5e' % bound
print 'Gap =\t\t%.5e' % gap
else:
print bound_str + ' =\tUnknown'
print 'Gap =\t\tUnknown'
if (len(bad_constr) != 0):
ms = max(
map(lambda x: x.get_value(),
[abs(c.left - c.right) for c in bad_constr]))
print 'Max Slack =\t%.5e' % ms
# Return objective and dual var
return obj, lagrange_mul_eq
def sqrtm(A):
"""
Matrix square root.
"""
return matrix(sci_sqrtm(A))
def sqrtm(A):
"""
Matrix square root.
"""
return matrix(sci_sqrtm(A))
def ones(shape):
"""
Matrix of ones.
"""
return matrix(np.ones(shape))
def solve_prog(p):
"""
Description
-----------
Solve optimization program.
Arguments
---------
p: cvxpy_program
"""
# Reset variable counter
var_reset()
# Check parameters
if(len(p.get_params()) != 0):
if(not p.options['quiet']):
print 'Error: Parameters present'
return np.NaN,np.NaN
# Original
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Original *'
print '****************************'
p.show()
# Expanded
p_expanded = p._get_expanded()
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Expanded *'
print '****************************'
p_expanded.show()
# Equivalent
p_equivalent = p_expanded._get_equivalent()
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Equivalent *'
print '****************************'
p_equivalent.show()
# Nonconvex constraints
bad_constr = p_equivalent.constr._get_nonconvex()
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Nonconvex Constraints *'
print '****************************\n'
print bad_constr
# Convex Relaxation
p_cvx_relaxation = p_equivalent._get_cvx_relaxation()
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Convex Relaxation *'
print '****************************'
p_cvx_relaxation.show()
# Solve convex relaxation
if(not p.options['quiet']):
print '\nSolving convex relaxation ...'
sol = solve_convex(p_cvx_relaxation,'rel')
lagrange_mul_eq = matrix(sol['y'])
if(not p.options['quiet']):
print 'Relaxation status: ',sol['status']
# Relaxation status: primal infeasible
if(sol['status'] == 'primal infeasible'):
if(not p.options['quiet']):
print 'Original program is primal infeasible'
if(p.action == MINIMIZE):
return np.inf,lagrange_mul_eq
else:
return -np.inf,lagrange_mul_eq
# Relaxation status: uknown
elif(sol['status'] == 'unknown'):
relaxation_obj = None
# Relaxation status: dual infeasible
elif(sol['status'] == 'dual infeasible'):
if(p.action == MINIMIZE):
relaxation_obj = -np.inf
else:
relaxation_obj = np.inf
# Relaxacion status: optimal
else:
relaxation_obj = p_cvx_relaxation.obj.get_value()
# Report max slack
if(len(bad_constr) != 0 and not p.options['quiet']):
print 'Max Slack: ',
print max(map(lambda x:x.get_value(),
[abs(c.left-c.right) for c in bad_constr]))
# Tighten
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Tightening Stage *'
print '****************************'
elif(not p.options['quiet']):
print '\nAttempting tightening ...'
valid_scp = scp(p_cvx_relaxation,bad_constr,sol)
if(not valid_scp):
return np.NaN,np.NaN
# Prepare results
if(len(bad_constr) == 0):
obj = relaxation_obj
else:
obj = p.obj.get_value()
if(p.action == MINIMIZE):
bound_str = 'Lower Bound'
if(relaxation_obj != None):
bound = np.min([relaxation_obj,obj])
gap = np.max([obj-relaxation_obj,0.0])
else:
bound_str = 'Upper Bound'
if(relaxation_obj != None):
bound = np.max([relaxation_obj,obj])
gap = np.max([relaxation_obj-obj,0.0])
# Show results
if(p.options['show steps'] and not p.options['quiet']):
print '\n'
print '****************************'
print '* Results *'
print '****************************'
elif(not p.options['quiet']):
print '\nResults'
if(not p.options['quiet']):
print 'Objective =\t%.5e' %obj
if(relaxation_obj != None):
print bound_str+' =\t%.5e' %bound
print 'Gap =\t\t%.5e' %gap
else:
print bound_str+' =\tUnknown'
print 'Gap =\t\tUnknown'
if(len(bad_constr) != 0):
ms = max(map(lambda x:x.get_value(),
[abs(c.left-c.right) for c in bad_constr]))
print 'Max Slack =\t%.5e' %ms
# Return objective and dual var
return obj,lagrange_mul_eq
def eye(n):
return matrix(np.eye(n))
def rand(m,n):
return matrix(np.random.rand(m,n))
def vstack(seq):
"""
| Stacks objects vertically.
:param seq: tuple or list of numbers,
:ref:`scalar objects<scalar_ref>` or
:ref:`multidimensional objects<multi_ref>`.
:return: :ref:`array<array_obj>` or
:ref:`matrix<matrix_obj>`.
"""
# Check input
if (type(seq) is not tuple and
type(seq) is not list):
raise TypeError('Invalid argument')
# Process input
new_list = []
numeric = True
for x in seq:
if np.isscalar(x):
new_list += [matrix(x)]
elif type(x) is cvxpy_obj:
new_list += [matrix(x.value)]
elif type(x).__name__ in SCALAR_OBJS:
numeric = False
new_x = cvxpy_array(1,1)
new_x[0,0] = x
new_list += [new_x]
elif type(x).__name__ in ARRAY_OBJS:
numeric = False
new_list += [x]
elif type(x) is cvxpy_matrix:
new_list += [x]
else:
raise TypeError('Invalid Input')
# Input is numeric
if numeric:
return np.vstack(new_list)
# Verify dimensions
n = new_list[0].shape[1]
for x in new_list:
if x.shape[1] != n:
raise ValueError('Invalid Dimensions')
# Allocate new array
m = int(np.sum([x.shape[0] for x in new_list]))
new_ar = cvxpy_array(m,n)
# Fill new array
k = 0
for x in new_list:
for i in range(0,x.shape[0],1):
for j in range(0,x.shape[1],1):
new_ar[i+k,j] = x[i,j]
k = k + x.shape[0]
# Return new array
return new_ar
