[Gesetzt]

Hello,

I have applied the second order backward difference now. It works, the gradient is zero now. But I'm still not sure about one thing with the backward difference at 0 °.

When I would stick strictly to the formula :
for the gradient dc(10)/d_theta=0=(3*c10-4*c1-1+c1-2)/2*h I would have two ghostpoints. So I thought the gradient is in the wrong direction here. I have used instead : (3*c10-4*c11+c12)/2*h=0 . Is this right or not correct ? For clarity of the location of the points :

Best regards

Dan

[FMDenaro]

The backward second order formula is

[3*f(i)-4*f(i-1)+f(i-2)]/(2*h)

the forward is

-[3*f(i)-4*f(i+1)+f(i+2)]/(2*h)