Testguy
- 5
- 0
I have a question regarding functional differentiablility.
I understand that Frechet differntiability of a functional T with respect to a norm \rho_1 implies Hadamard differentiability of the functional T with respect to the same norm.
However, it is no surprise that there would be cases where a functional T is not Hadamard differentiable with respect to a norm \rho, but that the same functional is Frechet differentiable with respect to a different norm \rho_2. Especially, this turn out to be the case for some functionals when \rho_1(\Delta)=sup_x |\Delta(x)| is the infinity norm, and
\rho_2(\Delta)=\int |\Delta(x)|dx is the L_1-norm.
According to a number of sources this should be the case for some functionals on the quite simple form T(H)=s(\int x dH(x) )=s(\int x h(x)dx),
where h(x) is the usual derivative of the function H(x), and s is some differentiable function.
I do however not find any s where this is the case.
Can anyone help me out with such a function s?
Any help is appreciated.
I understand that Frechet differntiability of a functional T with respect to a norm \rho_1 implies Hadamard differentiability of the functional T with respect to the same norm.
However, it is no surprise that there would be cases where a functional T is not Hadamard differentiable with respect to a norm \rho, but that the same functional is Frechet differentiable with respect to a different norm \rho_2. Especially, this turn out to be the case for some functionals when \rho_1(\Delta)=sup_x |\Delta(x)| is the infinity norm, and
\rho_2(\Delta)=\int |\Delta(x)|dx is the L_1-norm.
According to a number of sources this should be the case for some functionals on the quite simple form T(H)=s(\int x dH(x) )=s(\int x h(x)dx),
where h(x) is the usual derivative of the function H(x), and s is some differentiable function.
I do however not find any s where this is the case.
Can anyone help me out with such a function s?
Any help is appreciated.