Differentiating implicitly twice
By Anonymous on Tuesday, March 27, 2001
- 11:52 am :
How would you differentiate this implicitly twice?
dy/dx = exy
Thanks.
By Emma McCaughan (Emma) on Wednesday,
March 28, 2001 - 11:33 am :
You could try taking (natural) logs
first. When you differentiate ln(dy/dx), you can differentiate
with respect to (dy/dx), and then multiply by (dy/dx
differentiated with respect to x), which is d2
y/dx2 .
By Anonymous on Wednesday, March 28,
2001 - 07:08 pm :
Hi Emma,
Can you show me what you describe?
Thanks again.
By Anonymous on Wednesday, March 28,
2001 - 09:36 pm :
How do you differentiate ln(dy/dx) with respect to (dy/dx)?
By Emma McCaughan (Emma) on Thursday,
March 29, 2001 - 10:20 am :
Do you know how to differentiate log t
with respect to t? That's 1/t.
The chain rule says that if you want to differentiate f(t) with
respect to x, you do (df/dt) x (dt/dx), so that's (1/t) x
(dt/dx).
In this case, t=dy/dx, so we get (1/(dy/dx)) x (d2
y/dx2 ).
Do come back to us if you get stuck further on. Multiplying
through by dy/dx will be helpful, and then you'll have to keep
very clear about the product rule.
By The Editor :
Here is the complete working:
|
|
d2 y dx2
|
=y |
dy dx
|
+x |
æ ç
è
|
dy dx
|
ö ÷
ø
|
2
|
|
|
And if you really want it
differentiated twice:
|
|
d3 y dx3
|
= |
dy dx
|
|
dy dx
|
+y |
d2 y dx2
|
+ |
æ ç
è
|
dy dx
|
ö ÷
ø
|
2
|
+x |
æ ç
è
|
2 |
dy dx
|
|
d2 dy2
|
ö ÷
ø
|
|
|
|
=2 |
æ ç
è
|
dy dx
|
ö ÷
ø
|
2
|
+ |
d2 y dx2
|
|
æ ç
è
|
y+2x |
dy dx
|
ö ÷
ø
|
|
|