Chain rule (multivariable calculus)

Poetria
Messages
267
Reaction score
42
Homework Statement
Let ##f(x,y)=x^3+y^2+x*y##
Suppose that a point is moving through the plane. At time t , the point is at ## (x(t), y(t))=(t^2, e^{t-1})##. Use linear approximation to estimate the change in f as t goes from 1 to 1.1 . In other words, approximate
Relevant Equations
Multivariable chain rule
##f_x=3*x^2+y##
##f_y=2*y+x##

##(3*(t^2)^2+e^{t-1})*2*t+(2*e^{t-1}+t^2)*e^{t-1}##
Well, I am not sure how to evaluate it.
I got a wrong result by multiplying by 0.1, i.e.
##((3*(t^2)^2+e^{t-1})*2*t+(2*e^{t-1}+t^2)*e^{t-1})*0.1##

I guess it is trivial but I am lost. :(
 
Physics news on Phys.org
We are interested in the change along ##t##. So why don't you write ##f(x,y)=f(x(t),y(t))=f(t)## in the first place? Then we have two points at ##t=1## and ##t=1.1##. The linear approximation is the secant through these points:
$$
y=\underbrace{\dfrac{f(1)-f(1.1)}{1-1.1}}_{=:m}\cdot x + b\;\wedge\;f(1)=m\cdot 1+b
$$
Finally, you can check the quality of this approximation by calculating the tangent at ##t=1##:
$$
y=\left. \dfrac{d}{dt}\right|_{t=1}f(t) \cdot x + c =f'(1)x+c\;\wedge\;f(1)=f'(1)\cdot 1 +c
$$
which is the linear approximation if only one point is given.
 
  • Like
Likes Poetria
Many thanks. But I can't read this:
"The linear approximation is the secant through these points:
You can't use 'macro parameter character #' in math mode ?

A nice quote from Leibniz. :) Émilie du Châtelet (my avatar) was Leibnizian. :)
 
Poetria said:
Many thanks. But I can't read this
You need to give me a second to correct my format. Please reload.

"He who hasn't tasted bitter things hasn't earned sweet things." (G. Leibniz)

"Let us choose for ourselves our path in life, and let us try to strew that path with flowers." (E. du Châtelet)

##t\longmapsto f(t)## is the path, and you should plant the flowers at ##t=1## and ##t=1.1.##
 
  • Love
Likes Poetria
fresh_42 said:
You need to give me a second to correct my format. Please reload.

"He who hasn't tasted bitter things hasn't earned sweet things." (G. Leibniz)

"Let us choose for ourselves our path in life, and let us try to strew that path with flowers." (E. du Châtelet)

##t\longmapsto f(t)## is the path, and you should plant the flowers at ##t=1## and ##t=1.1.##
I got:
-10.3022 + 13.3022*x = y
x=1.1
Approximation of the change:
4.33022-3=1.33022

A beautiful metaphor: ##t\longmapsto f(t)## is the path, and you should plant the flowers at ##t=1## and ##t=1.1.##
 
Perhaps a gradient would be also a good idea:
Gradient ##\vec (0.21, 0.105171)##
A tangent plane: 0.21*x + 0.105171*y-0.32

Slope of the gradient: 0.500814
 
Last edited:
Poetria said:
A beautiful metaphor: ##t\longmapsto f(t)## is the path, and you should plant the flowers at ##t=1## and ##t=1.1.##
Yes, but a little bit too fast posted. The path is actually ##t\longmapsto (x(t),y(t))## and ##f(t)## the function values along the path.
 
  • Haha
Likes Poetria
  • Like
Likes fresh_42
Back
Top