Differences between revisions 4 and 6 (spanning 2 versions)
Revision 4 as of 2024-06-04 22:22:43
Size: 1397
Comment: More detail
Revision 6 as of 2025-10-16 20:14:53
Size: 1951
Comment: Rewrite
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
The '''chain rule''' is an approach for differentiating a composition of functions in terms of those functions. The '''chain rule''' is an approach for [[Calculus/Derivative|differentiating]] a composition of functions.
Line 11: Line 11:
== Notation == == Description ==
Line 13: Line 13:
The common notation for the chain rule is, given a function ''h(x) = f(g(x))'', the derivative ''h'(x) = f'(g(x)) g'(x)''. Given a function defined like ''f(x) = g(h(x))'', the derivative of ''f'' is ''g'(h(x)) h'(x)''.
Line 15: Line 15:
An equivalent notations is that ''h = fg'' and ''h' = (f ∘ g)' = (f' ∘ g)g' ''. Sometimes this is notated as ''f = gh'' and ''f' = (g' ∘ h)h' ''.
Line 17: Line 17:
A parallel notation is: Given a multivariate composition like ''f(g,h)'' where ''g = g(x,y)'' and ''h = h(x,y)'', and using a notation that expresses [[Calculus/PartialDerivative|partial derivatives]] as ''a,,b,, = ∂a/∂b'', the [[Calculus/Differential|total differential]] demonstrates that:
Line 19: Line 19:
{{attachment:leibniz.svg}} ''df = f,,g,,dg + f,,h,,dh''

''df = f,,g,,(g,,x,,dx + g,,y,,dy) + f,,h,,(h,,x,,dx + h,,y,,dy)''

''df = f,,g,,g,,x,,dx + f,,g,,g,,y,,dy + f,,h,,h,,x,,dx + f,,h,,h,,y,,dy''

''df = (f,,g,,g,,x,, + f,,h,,h,,x,,)dx + (f,,g,,g,,y,, + f,,h,,h,,y,,)dy''

Clearly then ''f,,g,,g,,x,, + f,,h,,h,,x,, = f,,x,,'' and ''f,,g,,g,,y,, + f,,h,,h,,y,, = f,,y,,'', so:
Line 47: Line 55:
Several of these terms are trivial to derive. The derivative of ''e'' to an exponent is itself, so the derivative of ''P'' with respect to ''v'' is ''P''. The derivative of a natural logarithm of some variable is 1 over that variable, so the derivative of ''u'' with respect to ''Q'' is ''1/Q''. This leaves the equation as: Several of these terms are trivial to derive. The derivative of ''e'' to an exponent is itself, so the derivative of ''P = e^v^'' with respect to ''v'' is ''P''. The derivative of a natural logarithm of some variable is 1 over that variable, so the derivative of ''u = ln Q'' with respect to ''Q'' is ''1/Q''. This leaves the equation as:

Chain Rule

The chain rule is an approach for differentiating a composition of functions.


Description

Given a function defined like f(x) = g(h(x)), the derivative of f is g'(h(x)) h'(x).

Sometimes this is notated as f = g ∘ h and f' = (g' ∘ h)h' .

Given a multivariate composition like f(g,h) where g = g(x,y) and h = h(x,y), and using a notation that expresses partial derivatives as ab = ∂a/∂b, the total differential demonstrates that:

df = fgdg + fhdh

df = fg(gxdx + gydy) + fh(hxdx + hydy)

df = fggxdx + fggydy + fhhxdx + fhhydy

df = (fggx + fhhx)dx + (fggy + fhhy)dy

Clearly then fggx + fhhx = fx and fggy + fhhy = fy, so:


Usage

Point elasticity is defined as:

elasticity1.svg

It can be simplified using the chain rule into:

elasticity2.svg

Start with the final formulation and use u- and v-substitution:

substitution1.svg

Note in particular that the equation for v can be rewritten in terms of P:

substitution2.svg

Now the final formulation can be rewritten as:

substitution3.svg

Several of these terms are trivial to derive. The derivative of e to an exponent is itself, so the derivative of P = ev with respect to v is P. The derivative of a natural logarithm of some variable is 1 over that variable, so the derivative of u = ln Q with respect to Q is 1/Q. This leaves the equation as:

substitution4.svg


CategoryRicottone

Calculus/ChainRule (last edited 2025-10-16 20:27:40 by DominicRicottone)