def setFlowMaterialCoeff(self, flow, material, coefficient):
if self._eqConstBuilt:
if material not in self._materialIdxLookup:
raise Exception("Invalid material")
if flow not in self._flows:
raise Exception("Invalid flow")
materialIdx = self._materialIdxLookup[material]
flowIdx = self._flows[flow]
coeffs, flowIdxs = zip(*self._materialCoeffs[material])
coeffs = list(coeffs)
flowLoc = flowIdxs.index(flowIdx)
coeffs[flowLoc] = coefficient
self._materialCoeffs[material] = zip(coeffs, flowIdxs)
rowIdx = int(materialIdx + 1)
length = len(flowIdxs)
if length != len(coeffs):
raise ValueError("Array sizes must match")
colIdxs = _toIndexArray(flowIdxs)
data = _toDoubleArray(coeffs)
glp.glp_set_mat_row(self._lp, rowIdx, length, colIdxs, data)
else:
idx = self._getVar(flow)
self._materialCoeffs[material].append((coefficient, idx))
self._solved = False
def set_linear_coefficients(self, coefficients):
if self.problem is not None:
problem = self.problem.problem
self.problem.update()
num_cols = glp_get_num_cols(problem)
ia = intArray(num_cols + 1)
va = doubleArray(num_cols + 1)
num_rows = glp_get_mat_row(self.problem.problem, self._index, ia, va)
variables_and_coefficients = {var.name: coeff for var, coeff in six.iteritems(coefficients)}
final_variables_and_coefficients = {
glp_get_col_name(problem, ia[i]): va[i] for i in range(1, num_rows + 1)
}
final_variables_and_coefficients.update(variables_and_coefficients)
ia = intArray(num_cols + 1)
va = doubleArray(num_cols + 1)
for i, (name, coeff) in enumerate(six.iteritems(final_variables_and_coefficients)):
ia[i + 1] = self.problem._variables[name]._index
va[i + 1] = float(coeff)
glp_set_mat_row(problem, self._index, len(final_variables_and_coefficients), ia, va)
else:
raise Exception("Can't change coefficients if constraint is not associated with a model.")
def set_constraints_csr(self, data, glpk_indices, indptr, shape):
'''
set the constrains row by row to be equal to the passed-in csr matrix attribues
glpk_indices is already offset by one
'''
Timers.tic('set_constraints_csr')
assert shape[0] <= self.get_num_rows()
assert shape[1] <= self.get_num_cols()
# actually set the constraints row by row
assert isinstance(data, list), "data was not a list"
for row in range(shape[0]):
# we must copy the indices since glpk is offset by 1 :(
count = int(indptr[row + 1] - indptr[row])
#indices_list = glpk_indices[indptr[row]:indptr[row+1]]
#indices_vec = SwigArray.as_int_array(indices_list)
indices_vec = SwigArray.as_int_array(
glpk_indices[indptr[row]:indptr[row + 1]], count)
#data_row_list = [float(d) for d in data[indptr[row]:indptr[row+1]]]
#data_vec = SwigArray.as_double_array(data_row_list)
data_vec = SwigArray.as_double_array(
data[indptr[row]:indptr[row + 1]], count)
glpk.glp_set_mat_row(self.lp, 1 + row, count, indices_vec,
data_vec)
Timers.toc('set_constraints_csr')
def set_linear_coefficients(self, coefficients):
if self.problem is not None:
problem = self.problem.problem
num_cols = glp_get_num_cols(problem)
ia = intArray(num_cols + 1)
va = doubleArray(num_cols + 1)
num_rows = glp_get_mat_row(self.problem.problem, self._index, ia, va)
variables_and_coefficients = {var.name: coeff for var, coeff in six.iteritems(coefficients)}
final_variables_and_coefficients = {
glp_get_col_name(problem, ia[i]): va[i] for i in range(1, num_rows + 1)
}
final_variables_and_coefficients.update(variables_and_coefficients)
ia = intArray(num_cols + 1)
va = doubleArray(num_cols + 1)
for i, (name, coeff) in enumerate(six.iteritems(final_variables_and_coefficients)):
ia[i + 1] = self.problem._variables[name]._index
va[i + 1] = coeff
glp_set_mat_row(problem, self._index, len(final_variables_and_coefficients), ia, va)
else:
raise Exception("Can't change coefficients if constraint is not associated with a model.")
def set_constraints_swigvec_rows(self, data_vec_list, indices_vec_list, count_list, row_offset):
'''An optimized / lower level way to set row constraints compared with set_constraints_csr
The passed in fields are a list of swig data vector and indices vectors (one for each row), as
well as a row offset.
'''
Timers.tic('set_constraints_swigvec_rows')
for row, (data, indices, count) in enumerate(zip(data_vec_list, indices_vec_list, count_list)):
glpk.glp_set_mat_row(self.lp, 1 + row_offset + row, count, indices, data)
Timers.toc('set_constraints_swigvec_rows')
def _add_constraints(self, constraints, sloppy=False):
super(Model, self)._add_constraints(constraints, sloppy=sloppy)
for constraint in constraints:
constraint._problem = None # This needs to be dones in order to not trigger constraint._get_expression()
glp_add_rows(self.problem, 1)
index = glp_get_num_rows(self.problem)
glp_set_row_name(self.problem, index, str(constraint.name))
num_cols = glp_get_num_cols(self.problem)
index_array = intArray(num_cols + 1)
value_array = doubleArray(num_cols + 1)
num_vars = 0 # constraint.variables is too expensive for large problems
if constraint.expression.is_Atom and constraint.expression.is_Symbol:
var = constraint.expression
index_array[1] = var.index
value_array[1] = 1
num_vars += 1
elif constraint.expression.is_Mul:
args = constraint.expression.args
if len(args) > 2:
raise Exception(
"Term(s) %s from constraint %s is not a proper linear term." % (args, constraint))
coeff = float(args[0])
var = args[1]
index_array[1] = var.index
value_array[1] = coeff
num_vars += 1
else:
for i, term in enumerate(constraint.expression.args):
args = term.args
if args == ():
assert term.is_Symbol
coeff = 1
var = term
elif len(args) == 2:
assert args[0].is_Number
assert args[1].is_Symbol
var = args[1]
coeff = float(args[0])
elif len(args) > 2:
raise Exception(
"Term %s from constraint %s is not a proper linear term." % (term, constraint))
index_array[i + 1] = var.index
value_array[i + 1] = coeff
num_vars += 1
glp_set_mat_row(self.problem, index, num_vars,
index_array, value_array)
constraint._problem = self
self._glpk_set_row_bounds(constraint)
def set_constraints_csr(self, csr_mat, offset=None):
'''set the constrains row by row to be equal to the passed-in csr matrix
offset is an optional tuple (num_rows, num_cols) which tells you the top-left offset for the assignment
'''
assert isinstance(csr_mat, csr_matrix)
assert csr_mat.dtype == float
if offset is None:
offset = (0, 0)
assert len(
offset) == 2, "offset should be a 2-tuple (num_rows, num_cols)"
# check that the matrix is in bounds
lp_rows = self.get_num_rows()
lp_cols = self.get_num_cols()
if offset[0] < 0 or offset[1] < 0 or \
offset[0] + csr_mat.shape[0] > lp_rows or offset[1] + csr_mat.shape[1] > lp_cols:
raise RuntimeError("Error: set constraints matrix out of bounds (offset was " + \
"{}, matrix size was {}), but lp size was ({}, {})".format(
offset, csr_mat.shape, lp_rows, lp_cols))
# actually set the constraints row by row
indptr = csr_mat.indptr
indices = csr_mat.indices
data_list = csr_mat.data.tolist()
for row in range(csr_mat.shape[0]):
# we must copy the indices since glpk is offset by 1 :(
count = int(indptr[row + 1] - indptr[row])
indices_list = [
1 + offset[1] + int(indices[index])
for index in range(indptr[row], indptr[row + 1])
]
indices_vec = SwigArray.as_int_array(indices_list)
data_row_list = data_list[indptr[row]:indptr[row + 1]]
data_vec = SwigArray.as_double_array(data_row_list)
glpk.glp_set_mat_row(self.lp, offset[0] + row + 1, count,
indices_vec, data_vec)
def _set_coefficients_low_level(self, variables_coefficients_dict):
if self.problem is not None:
problem = self.problem.problem
indices_coefficients_dict = dict(
[(variable.index, coefficient) for variable, coefficient in six.iteritems(variables_coefficients_dict)])
num_cols = glp_get_num_cols(problem)
ia = intArray(num_cols + 1)
da = doubleArray(num_cols + 1)
index = self.index
num = glp_get_mat_row(self.problem.problem, index, ia, da)
for i in range(1, num + 1):
try:
da[i] = indices_coefficients_dict[ia[i]]
except KeyError:
pass
glp_set_mat_row(self.problem.problem, index, num, ia, da)
else:
raise Exception(
'_set_coefficients_low_level works only if a constraint is associated with a solver instance.')
def _add_constraints(self, relation):
"""Add the given relation as one or more constraints.
Return a list of the names of the constraints added.
"""
expression = relation.expression
constr_count = sum(True for _ in expression.value_sets())
if constr_count == 0:
return []
row_indices = count(swiglpk.glp_add_rows(self._p, constr_count))
names = []
for i, value_set in zip(row_indices, expression.value_sets()):
value_set = list(value_set)
var_indices = swiglpk.intArray(1 + len(value_set))
var_values = swiglpk.doubleArray(1 + len(value_set))
for j, (variable, coeff) in enumerate(value_set):
var_indices[1 + j] = self._variables[variable]
var_values[1 + j] = float(coeff)
swiglpk.glp_set_mat_row(self._p, i, len(value_set), var_indices,
var_values)
if relation.sense == RelationSense.Greater:
swiglpk.glp_set_row_bnds(self._p, i, swiglpk.GLP_LO,
-float(expression.offset), 0)
elif relation.sense == RelationSense.Less:
swiglpk.glp_set_row_bnds(self._p, i, swiglpk.GLP_UP, 0,
-float(expression.offset))
else:
swiglpk.glp_set_row_bnds(self._p, i, swiglpk.GLP_FX,
-float(expression.offset), 0)
names.append(i)
self._do_presolve = True
return names
def _add_constraints(self, relation):
"""Add the given relation as one or more constraints.
Return a list of the names of the constraints added.
"""
expression = relation.expression
constr_count = sum(True for _ in expression.value_sets())
if constr_count == 0:
return []
row_indices = count(swiglpk.glp_add_rows(self._p, constr_count))
names = []
for i, value_set in zip(row_indices, expression.value_sets()):
value_set = list(value_set)
var_indices = swiglpk.intArray(1 + len(value_set))
var_values = swiglpk.doubleArray(1 + len(value_set))
for j, (variable, coeff) in enumerate(value_set):
var_indices[1 + j] = self._variables[variable]
var_values[1 + j] = float(coeff)
swiglpk.glp_set_mat_row(
self._p, i, len(value_set), var_indices, var_values)
if relation.sense == RelationSense.Greater:
swiglpk.glp_set_row_bnds(
self._p, i, swiglpk.GLP_LO, -float(expression.offset), 0)
elif relation.sense == RelationSense.Less:
swiglpk.glp_set_row_bnds(
self._p, i, swiglpk.GLP_UP, 0, -float(expression.offset))
else:
swiglpk.glp_set_row_bnds(
self._p, i, swiglpk.GLP_FX, -float(expression.offset), 0)
names.append(i)
self._do_presolve = True
return names
def _add_constraints(self, constraints, sloppy=False):
super(Model, self)._add_constraints(constraints, sloppy=sloppy)
for constraint in constraints:
constraint._problem = None # This needs to be done in order to not trigger constraint._get_expression()
glp_add_rows(self.problem, 1)
index = glp_get_num_rows(self.problem)
glp_set_row_name(self.problem, index, str(constraint.name))
num_cols = glp_get_num_cols(self.problem)
index_array = intArray(num_cols + 1)
value_array = doubleArray(num_cols + 1)
num_vars = 0 # constraint.variables is too expensive for large problems
offset, coef_dict, _ = parse_optimization_expression(constraint, linear=True)
num_vars = len(coef_dict)
for i, (var, coef) in enumerate(coef_dict.items()):
index_array[i + 1] = var._index
value_array[i + 1] = float(coef)
glp_set_mat_row(self.problem, index, num_vars,
index_array, value_array)
constraint._problem = self
self._glpk_set_row_bounds(constraint)
def add_dense_row(self, vec, rhs):
'''
add a row from a dense nd.array, row <= rhs
'''
Timers.tic('add_dense_row')
assert isinstance(vec, np.ndarray)
assert len(vec.shape) == 1 or vec.shape[0] == 1
assert len(vec) == self.get_num_cols(
), f"vec had {len(vec)} values, but lpi has {self.get_num_cols()} cols"
rows_before = self.get_num_rows()
self.add_rows_less_equal([rhs])
data_vec = SwigArray.as_double_array(vec, vec.size)
indices_vec = SwigArray.get_sequential_int_array(vec.size)
glpk.glp_set_mat_row(self.lp, rows_before + 1, vec.size, indices_vec,
data_vec)
Timers.toc('add_dense_row')
def _add_constraints(self, constraints, sloppy=False):
super(Model, self)._add_constraints(constraints, sloppy=sloppy)
for constraint in constraints:
constraint._problem = None # This needs to be done in order to not trigger constraint._get_expression()
glp_add_rows(self.problem, 1)
index = glp_get_num_rows(self.problem)
glp_set_row_name(self.problem, index, str(constraint.name))
num_cols = glp_get_num_cols(self.problem)
index_array = intArray(num_cols + 1)
value_array = doubleArray(num_cols + 1)
num_vars = 0 # constraint.variables is too expensive for large problems
coef_dict, _ = parse_optimization_expression(constraint, linear=True)
num_vars = len(coef_dict)
for i, (var, coef) in enumerate(coef_dict.items()):
index_array[i + 1] = var._index
value_array[i + 1] = float(coef)
glp_set_mat_row(self.problem, index, num_vars,
index_array, value_array)
constraint._problem = self
self._glpk_set_row_bounds(constraint)
def _import_problem(self):
import swiglpk as glpk
if self.verbosity() >= 1:
glpk.glp_term_out(glpk.GLP_ON)
else:
glpk.glp_term_out(glpk.GLP_OFF)
# Create a problem instance.
p = self.int = glpk.glp_create_prob();
# Set the objective.
if self.ext.objective[0] in ("find", "min"):
glpk.glp_set_obj_dir(p, glpk.GLP_MIN)
elif self.ext.objective[0] is "max":
glpk.glp_set_obj_dir(p, glpk.GLP_MAX)
else:
raise NotImplementedError("Objective '{0}' not supported by GLPK."
.format(self.ext.objective[0]))
# Set objective function shift
if self.ext.objective[1] is not None \
and self.ext.objective[1].constant is not None:
if not isinstance(self.ext.objective[1], AffinExp):
raise NotImplementedError("Non-linear objective function not "
"supported by GLPK.")
if self.ext.objective[1].constant.size != (1,1):
raise NotImplementedError("Non-scalar objective function not "
"supported by GLPK.")
glpk.glp_set_obj_coef(p, 0, self.ext.objective[1].constant[0])
# Add variables.
# Multideminsional variables are split into multiple scalar variables
# represented as matrix columns within GLPK.
for varName in self.ext.varNames:
var = self.ext.variables[varName]
# Add a column for every scalar variable.
numCols = var.size[0] * var.size[1]
glpk.glp_add_cols(p, numCols)
for localIndex, picosIndex \
in enumerate(range(var.startIndex, var.endIndex)):
glpkIndex = self._picos2glpk_variable_index(picosIndex)
# Assign a name to the scalar variable.
scalarName = varName
if numCols > 1:
x = localIndex // var.size[0]
y = localIndex % var.size[0]
scalarName += "_{:d}_{:d}".format(x + 1, y + 1)
glpk.glp_set_col_name(p, glpkIndex, scalarName)
# Assign bounds to the scalar variable.
lower, upper = var.bnd.get(localIndex, (None, None))
if lower is not None and upper is not None:
if lower == upper:
glpk.glp_set_col_bnds(
p, glpkIndex, glpk.GLP_FX, lower, upper)
else:
glpk.glp_set_col_bnds(
p, glpkIndex, glpk.GLP_DB, lower, upper)
elif lower is not None and upper is None:
glpk.glp_set_col_bnds(p, glpkIndex, glpk.GLP_LO, lower, 0)
elif lower is None and upper is not None:
glpk.glp_set_col_bnds(p, glpkIndex, glpk.GLP_UP, 0, upper)
else:
glpk.glp_set_col_bnds(p, glpkIndex, glpk.GLP_FR, 0, 0)
# Assign a type to the scalar variable.
if var.vtype in ("continuous", "symmetric"):
glpk.glp_set_col_kind(p, glpkIndex, glpk.GLP_CV)
elif var.vtype == "integer":
glpk.glp_set_col_kind(p, glpkIndex, glpk.GLP_IV)
elif var.vtype == "binary":
glpk.glp_set_col_kind(p, glpkIndex, glpk.GLP_BV)
else:
raise NotImplementedError("Variable type '{0}' not "
"supported by GLPK.".format(var.vtype()))
# Set objective function coefficient of the scalar variable.
if self.ext.objective[1] is not None \
and var in self.ext.objective[1].factors:
glpk.glp_set_obj_coef(p, glpkIndex,
self.ext.objective[1].factors[var][localIndex])
# Add constraints.
# Multideminsional constraints are split into multiple scalar
# constraints represented as matrix rows within GLPK.
rowOffset = 1
for constraintNum, constraint in enumerate(self.ext.constraints):
if not isinstance(constraint, AffineConstraint):
raise NotImplementedError(
"Non-linear constraints not supported by GLPK.")
# Add a row for every scalar constraint.
# Internally, GLPK uses an auxiliary variable for every such row,
# bounded by the right hand side of the scalar constraint in a
# canonical form.
numRows = len(constraint)
glpk.glp_add_rows(p, numRows)
self._debug("Handling PICOS Constraint: " + str(constraint))
# Split multidimensional constraints into multiple scalar ones.
for localConIndex, (glpkVarIndices, coefficients, rhs) in \
enumerate(constraint.sparse_Ab_rows(
None, indexFunction = lambda picosVar, i:
self._picos2glpk_variable_index(picosVar.startIndex + i))):
# Determine GLPK's row index of the scalar constraint.
glpkConIndex = rowOffset + localConIndex
numColumns = len(glpkVarIndices)
# Name the auxiliary variable associated with the current row.
if constraint.name:
name = constraint.name
else:
name = "rhs_{:d}".format(constraintNum)
if numRows > 1:
x = localConIndex // constraint.size[0]
y = localConIndex % constraint.size[0]
name += "_{:d}_{:d}".format(x + 1, y + 1)
glpk.glp_set_row_name(p, glpkConIndex, name)
# Assign bounds to the auxiliary variable.
if constraint.is_equality():
glpk.glp_set_row_bnds(p, glpkConIndex, glpk.GLP_FX, rhs,rhs)
elif constraint.is_increasing():
glpk.glp_set_row_bnds(p, glpkConIndex, glpk.GLP_UP, 0, rhs)
elif constraint.is_decreasing():
glpk.glp_set_row_bnds(p, glpkConIndex, glpk.GLP_LO, rhs, 0)
else:
assert False, "Unexpected constraint relation."
# Set coefficients for current row.
# Note that GLPK requires a glpk.intArray containing column
# indices and a glpk.doubleArray of same size containing the
# coefficients for the listed column index. The first element
# of both arrays (with index 0) is skipped by GLPK.
glpkVarIndicesArray = glpk.intArray(numColumns + 1)
for i in range(numColumns):
glpkVarIndicesArray[i + 1] = glpkVarIndices[i]
coefficientsArray = glpk.doubleArray(numColumns + 1)
for i in range(numColumns):
coefficientsArray[i + 1] = coefficients[i]
glpk.glp_set_mat_row(p, glpkConIndex, numColumns,
glpkVarIndicesArray, coefficientsArray)
rowOffset += numRows
