If Iam going to derive the DD of a function without using a unit vector i.e.$u=<a,b>$ (not unit vector). I get the same result $Du=(Fx).a+(Fy).b$.So why are we considering unit vector always?

  • $\begingroup$ write $a$ and $b$ using polar coordinates. Now do you see the problem? $\endgroup$ – The Great Duck Dec 29 '17 at 7:35
  • $\begingroup$ How is this “the same result?” $\nabla f\cdot v \ne \nabla f\cdot {v\over\|v\|}$ if $v$ isn’t a unit vector. $\endgroup$ – amd Dec 29 '17 at 8:44
  • $\begingroup$ Yes true, but i get the same formula , but different results when i change the magnitude of the vector but in the same direction.By same formula i mean Du=(fx)a+(fy)b, whether it may or many not be a unit vector. $\endgroup$ – CHAND Dec 29 '17 at 16:23
  • $\begingroup$ @CHAND Please, if you are ok, you can accept the answer and set it as solved. Thanks! $\endgroup$ – user Feb 1 '18 at 22:48

If V = < a, b> is not a unit vector then afx + bfy is not the directional derivative of f at the direction of V. It is the directional derivative multiplied by the norm of V. Therefore you do not get the same result unless you normalize V.

  • $\begingroup$ That depends on your definition of directional derivative. Not all authors require a unit vector. Indeed, if you’re not working in a normed space, then what’s a unit vector in the first place? $\endgroup$ – amd Dec 29 '17 at 8:43
  • $\begingroup$ Solution of direction derivative changes when we use a vector other than a unit vector at a point but in the same direction, but I'm asking why is it changing even though it is in the same direction? $\endgroup$ – CHAND Dec 29 '17 at 12:23
  • $\begingroup$ That is exactly why you need to consider a unit vector so you get the same answer for all vectors in the same direction. $\endgroup$ – Mohammad Riazi-Kermani Dec 29 '17 at 13:07
  • $\begingroup$ The directional derivative is a dot product of the gradient vector and the unit vector in the desired direction. As you know the dot product depends on the norms of vectors so you get a different answer if you do not normalize your vector. $\endgroup$ – Mohammad Riazi-Kermani Dec 29 '17 at 13:13
  • $\begingroup$ The dot product of the gradient and a unit vector happens to equal the directional derivative for differentiable functions (and some others), but that’s not the fundamental definition of a directional derivative. $\endgroup$ – amd Dec 29 '17 at 23:28

For any $\vec v=(a,b)$ we can define the directional derivative as:

$$\frac{\partial f}{\partial \vec v}=\lim_{h\to 0}\frac{f(x_0+ah,y_0+bh)-f(x_0,y_0)}{h}$$

Often the directional derivative are defined assuming $\vec v$ as a unit vector, indeed it can be shown that:

$$\frac{\partial f}{\partial \lambda \vec v}=\lim_{h\to 0}\lambda\frac{f(x_0+\lambda ah,y_0+\lambda bh)-f(x_0,y_0)}{\lambda h}=\lambda \frac{\partial f}{\partial \vec v}$$

Notably the partial derivatives are the directional derivatives corresponding to the unit vectors $(1,0)$ and $(0,1)$.

Wheter $f$ is differentiable the following holds:

$$\frac{\partial f}{\partial \vec v}=\nabla f\cdot\vec v=v_1\frac{\partial f}{\partial x}+v_2\frac{\partial f}{\partial y}$$


The formula stated here gives the derivative of f, a real valued function of two variables, with respect to a vector u whose components are 'a' and 'b' at a point where the partial derivatives(denoted in the question by fx and fy) of f are evaluated, in case the function f is differentiable there, as a function of two variables. If we put the additional condition that u is a unit vector we get what is usually referred to as the directional derivative of f,at the point, along u and which is equal to the component of the gradient vector of f along u. Ref: Tom Apostol: Mathematical Analysis.

  • $\begingroup$ The formula stated where? $\endgroup$ – amd Dec 29 '17 at 7:00
  • $\begingroup$ @amd in the rest of that reference somewhere. beats me. $\endgroup$ – The Great Duck Dec 29 '17 at 7:26
  • $\begingroup$ That’s not very useful. $\endgroup$ – amd Dec 29 '17 at 8:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.