Derivative of a scalar product with itself using distributive law

46 Views Asked by At

We know that $\frac{\partial (\hat{a}.\hat{b})}{\partial (\hat{a}.\hat{b})} = 1$

But, if we use distributive law

$\frac{\partial (\hat{a}.\hat{b})}{\partial (\hat{a}.\hat{b})} = \frac{\partial \hat{a}}{\partial (\hat{a}.\hat{b})}.\hat{b} + \hat{a}.\frac{\partial \hat{b}}{\partial (\hat{a}.\hat{b})}$

with $\frac{\partial \hat{a}}{\partial (\hat{a}.\hat{b})} = \frac{1}{b_{x}} \hat{i} + \frac{1}{b_{y}} \hat{j} + \frac{1}{b_{z}} \hat{k}$ and $\frac{\partial \hat{b}}{\partial (\hat{a}.\hat{b})} = \frac{1}{a_{x}} \hat{i} + \frac{1}{a_{y}} \hat{j} + \frac{1}{a_{z}} \hat{k}$

If we use that in the previous equation,

$\frac{\partial (\hat{a}.\hat{b})}{\partial (\hat{a}.\hat{b})} = 3 + 3 = 6$

Where did I go wrong here?

1

There are 1 best solutions below

0
On

Lets take one dimensional case where, $\hat{a}.\hat{b} = ab$ ;(can be trivially generalized to higher dimensions). Then, what we need to prove is $\frac{d(\hat{a}.\hat{b})}{d(\hat{a}.\hat{b})} = 1$, NOT $\frac{\partial (\hat{a}.\hat{b})}{\partial (\hat{a}.\hat{b})} = 1$.

By distributive law,

$\frac{d(\hat{a}.\hat{b})}{d(\hat{a}.\hat{b})}$ = $\frac{d(\hat{a})}{d(\hat{a}.\hat{b})}.\hat{b}$ + $\hat{a}.\frac{d(\hat{b})}{d(\hat{a}.\hat{b})}$

= $\frac{d(a)}{d(ab)}b + a\frac{d(b)}{d(ab)}$ ...................1

$a = \frac{ab}{b}$

$da = \frac{d(ab)}{b} - \frac{(ab)db}{b^2}$

divide by $d(ab)$,

$\frac{da}{d(ab)} = \frac{1}{b} - \frac{(ab)db}{b^2d(ab)}$ NOT just $\frac{1}{b}$

Substituiting $\frac{da}{d(ab)}$ in equation 1,

$\frac{d(\hat{a}.\hat{b})}{d(\hat{a}.\hat{b})} = 1 - a\frac{db}{d(ab)} + a\frac{db}{d(ab)} = 1$.

Hope this helps.