This library provides some functions to make optimization in python easier.
It can use scipy.optimize. Also, it provides
an interface that makes minimizing functions of multiple variables easier,
especially if only a subset of the variables should be considered for the
optimization.
This package used to contain a convenience function minimize_ipopt that
mimicked the scipy.mimize.optimize interface. This function is now part of
https://github.com/matthias-k/cyipopt. If you want to use optpy
with ipopt just set the method argument to the minimize function from
cyipopt.
In practise, often one has to optimize functions that depend on multiple variables. Most
optimizers, including scipy.optimize.minimze and ipopt can handle only functions that
depend on one array-like parameter that should be optimized. It gets even more
complicated when one wants to include certain variables in a flexible way into
the optimimzation or keep them constant. optpy provides an easy way to do so:
Here we have a function of three variables x1, x2, x3, but we want to optimize
only x1 and x2.
def f(x1, x2, x3):
return np.sum(x1**2)+np.sum(x2**2)
def constraint(x1, x2, x3):
return x1-1.0
constraints = [{'type': 'eq', 'fun': constraint}]
res = optimization.minimize(f, {'x1': 1.0,
'x2': np.array([2.0, 2.0]),
'x3': 1.0},
optimize=['x1', 'x2'],
method='SLSQP', constraints = constraints)
np.testing.assert_allclose(res.x1, 1.0)
np.testing.assert_allclose(res.x2, [0.0, 0.0])
np.testing.assert_allclose(res.x3, 1.0)There is also a more explicit interface:
parameter_manager = optpy.ParameterManager(['x1', 'x2', 'x3'], ['x1', 'x2'],
x1=1.0, x2=np.array([2.0, 2.0]), x3=1.0)
def f(x1, x2, x3):
return np.sum(x1**2)+np.sum(x2**2)
def constraint(x1, x2, x3):
return x1-1.0
constraints = [{'type': 'eq', 'fun': constraint}]
res = optpy.minimize(f, parameter_manager, method='SLSQP', constraints = constraints)
np.testing.assert_allclose(res.x1, 1.0)
np.testing.assert_allclose(res.x2, [0.0, 0.0])
np.testing.assert_allclose(res.x3, 1.0)