Linearizing f(x,y,z) at (1,1,1) for Quick Calculations

  • Thread starter Thread starter Mindscrape
  • Start date Start date
  • Tags Tags
    Linearization
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
1 reply · 2K views
Mindscrape
Messages
1,854
Reaction score
1
I am linearizing the function [tex]f(x,y,z) = tan^{-1}(xyz)[/tex] at the point (1,1,1).

Since [tex]f(x_0,y_0,z_0)= \frac{\pi}{4} + \pi*n[/tex] should I just take the first value or do I have to carry all the solutions through the linearization process?

Um, anybody remember this? I can put up some work if it helps.

[tex]L(x,y,z) = \Delta f = f_x \Delta x + f_y \Delta y + f_z \Delta z[/tex]

So

[tex]L(x,y,z) = f(x_0,y_0,z_0) + f_x(x-x_0) + f_y (y-y_0) + f_z (z-z_0)[/tex]

where [tex]f_x = \frac{\partial f}{\partial x}[/tex] (and f_y and f_z)

So does [tex]L(x,y,z) = \frac{\pi}{4} +\pi n + \frac{1}{2}[(x-1)+(y-1)+(z-1)][/tex]
?
 
Last edited:
Physics news on Phys.org
Surely somebody must know. I have doubts that I should include the general solutions, because that would be a strange linearization. I'm not really sure though, since the general solutions would be just as valid, unless I am overlooking something.