Understanding the Relationship Between A and B: A Variation Analysis

  • Context: Undergrad 
  • Thread starter Thread starter ChrisVer
  • Start date Start date
  • Tags Tags
    Variation
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
4 replies · 2K views
Messages
3,372
Reaction score
465
Suppose I have such an equation:
[itex]A= f(B)[/itex]
so [itex]A[/itex] is a function of [itex]B[/itex].

Can I really use the fact that a variation of [itex]A[/itex] is like taking the differential of it?
[itex]\delta A= dA[/itex]

so that:

[itex]\delta A = \frac{d f(B)}{dB} \delta B[/itex]
?
 
Physics news on Phys.org
the reason I ask is because I have a confusion with one more way to see variation.
Suppose I vary [itex]B[/itex], so that [itex]A[/itex] is going to be varied too:

[itex]A+ \delta A = f(B+\delta B)[/itex]
And suppose [itex]f(x) = \frac{1}{\sqrt{x}}[/itex]

Then:

[itex]A( 1+ \frac{\delta A}{A}) = \frac{1}{\sqrt{B+\delta B}}= \frac{1}{\sqrt{B}} \frac{1}{\sqrt{1+\frac{\delta B}{B}}}≈ A -A \frac{1}{2} \frac{\delta B}{B}[/itex]

[itex]\frac{\delta A}{A}= - \frac{1}{2} \frac{\delta B}{B}[/itex]This result would imply that:
[itex]\frac{d f(B)}{dB} = -\frac{1}{2}\frac{A}{B}[/itex]

in my OP...

with a simple check, we have:

[itex]\frac{d (1/ \sqrt{B})}{dB} = -\frac{1}{2} \frac{1}{B^{3/2}}[/itex]

This method works fine for such an easy function as [itex]f(x)=\frac{1}{\sqrt{x}}[/itex]
Now suppose I have instead:
[itex]f(x)= \frac{1}{\sqrt{x}} + c[/itex]

with [itex]c[/itex] a constant. The same method won't work because I will end up with something like:

[itex]\frac{\delta A}{A} = - \frac{1}{A} \frac{1}{2} \frac{1}{\sqrt{B}} \frac{\delta B}{B}[/itex]

unfortunately now [itex]\frac{1}{A}[/itex] can't be canceled with the [itex]\frac{1}{\sqrt{x}}[/itex] ... Only maybe, up to some error, if I consider [itex]c[/itex] very small... The derivative however will kill [itex]c[/itex] and give some other result (the same as for without [itex]c[/itex] which I think is more physical- isn't it? a constant shouldn't destroy the variations)..
 
Last edited:
I think I found my misconception... the variation is still independent of [itex]c[/itex]... it's 1/A which gives me that dependence. So there's no way to escape from c rather than neglecting it :) that's fine...i guess
 
You can basically use Taylor expansion to show your equality.

If ##A=f(B)## and we add a small quantity ##\delta B## to ##B## then by Taylor's theorem we will have:

$$A'\equiv A+\delta A=f(B+\delta B)=f(B)+\frac{df}{dB}\delta B+\mathcal{O}(\delta B^2)$$

This gives as the limit ##\delta B\rightarrow 0##:

$$\delta A=\frac{df}{dB}\delta B$$

Of course, this is usually much more interesting when we consider functionals rather than simple functions.