def jac(self, *args, **kwargs):
"""Evaluate constraints Jacobian at x."""
vals, rows, cols = super(PySparseNLPModel, self).jac(*args, **kwargs)
J = psp(nrow=self.ncon,
ncol=self.nvar,
sizeHint=vals.size,
symmetric=False)
J.put(vals, rows, cols)
return J
def A(self, *args, **kwargs):
"""Evaluate sparse Jacobian of the linear part of the constraints.
Useful to obtain constraint matrix when problem is a linear programming
problem.
"""
vals, rows, cols = super(PySparseAmplModel,
self).A(*args, **kwargs)
A = psp(nrow=self.ncon,
ncol=self.nvar,
sizeHint=vals.size,
symmetric=False)
A.put(vals, rows, cols)
return A
def hess(self, x, z=None, obj_num=0, *args, **kwargs):
"""Evaluate Lagrangian Hessian at (x, z)."""
model = self.model
if isinstance(model, QuasiNewtonModel):
return self.hop(x, z, *args, **kwargs)
if z is None:
z = np.zeros(self.m)
# Create some shortcuts for convenience
model = self.model
on = self.original_n
H = psp(nrow=self.n,
ncol=self.n,
symmetric=True,
sizeHint=self.model.nnzh)
H[:on, :on] = model.hess(x[:on], z, obj_num, *args, **kwargs)
return H
def _jac(self, x, lp=False):
"""Helper method to assemble the Jacobian matrix.
See the documentation of :meth:`jac` for more information.
The positional argument `lp` should be set to `True` only if the
problem is known to be a linear program. In this case, the evaluation
of the constraint matrix is cheaper and the argument `x` is ignored.
"""
n = self.n
m = self.m
model = self.model
on = self.original_n
lowerC = np.array(model.lowerC)
nlowerC = model.nlowerC
upperC = np.array(model.upperC)
nupperC = model.nupperC
rangeC = np.array(model.rangeC)
nrangeC = model.nrangeC
# Initialize sparse Jacobian
nnzJ = self.model.nnzj + m
J = psp(nrow=m, ncol=n, sizeHint=nnzJ)
# Insert contribution of general constraints
if lp:
J[:on, :on] = self.model.A()
else:
J[:on, :on] = self.model.jac(x[:on])
# Create a few index lists
rlowerC = np.array(range(nlowerC))
rupperC = np.array(range(nupperC))
rrangeC = np.array(range(nrangeC))
# Insert contribution of slacks on general constraints
J.put(-1.0, lowerC, on + rlowerC)
J.put(-1.0, upperC, on + nlowerC + rupperC)
J.put(-1.0, rangeC, on + nlowerC + nupperC + rrangeC)
return J
def hess(self, *args, **kwargs):
"""Evaluate Lagrangian Hessian at (x, z)."""
vals, rows, cols = super(PySparseNLPModel, self).hess(*args, **kwargs)
H = psp(size=self.nvar, sizeHint=vals.size, symmetric=True)
H.put(vals, rows, cols)
return H
from cysparse.sparse.ll_mat import * A = LLSparseMatrix(nrow=3, ncol=3) A.put_triplet([0, 1, 2], [0, 1, 2], [1.0, 2.0, 3.0]) print A from pysparse.sparse import PysparseMatrix as psp J = psp(nrow=3, ncol=3) J[0,0] = 1 J[1,1] = 2 J[2,2] = 3 print J # Scale two first rows:" J[:2,:] *= -1.0 print J D = BandLLSparseMatrix(size=3, diag_coeff=[0], numpy_arrays=[np.array([-1, -1, 1], dtype=np.float64)]) A = A * D print A.to_ll()