InfimalConvolution

class odl.functionals.functional.InfimalConvolution(*args, **kwargs)[source]

Bases: Functional

Functional representing h(x) = inf_y f(x-y) + g(y).

__init__(left, right)[source]

Initialize a new instance.

Parameters

leftFunctional: Function corresponding to f.
rightFunctional: Function corresponding to g.

Examples

>>> space = odl.rn(3)
>>> l1 = odl.functionals.L1Norm(space)
>>> l2 = odl.functionals.L2Norm(space)
>>> f = odl.functionals.InfimalConvolution(l1.convex_conj, l2.convex_conj)
>>> x = f.domain.one()
>>> f.convex_conj(x) - (l1(x) + l2(x))
0.0