brydustin
Jul28-11, 01:08 PM
If one wanted to construct a Hessian matrix the obvious solution would be to do the following:
-------
for i= 1:n
for j=i:n
H(i,j)= DEFINITION
end
end
H = H + H' - diag(diag(H));
-------
in this was way one only has to compute the upper half of the matrix and can then transpose it to itself to make the full Hessian.
What would be the similar solution for the third derivative tensor?
i.e. T_(ijk) = all permutations of {i,j,k}.
-------
for i= 1:n
for j=i:n
H(i,j)= DEFINITION
end
end
H = H + H' - diag(diag(H));
-------
in this was way one only has to compute the upper half of the matrix and can then transpose it to itself to make the full Hessian.
What would be the similar solution for the third derivative tensor?
i.e. T_(ijk) = all permutations of {i,j,k}.