PDA

View Full Version : Newton's method


shwin
Feb5-08, 04:35 PM
lets say i have the functions f(x,y) = cos(x-y) - y and g(x,y) = sin(x+y) - x. I want to use newton's method to approximate these functions. Do I just take the partial derivatives with respect to x and y of each function and plug in a given point (a,b)?

EnumaElish
Feb5-08, 05:27 PM
Do you mean to approximate the roots? Or do you mean to approximate an extremum (min or max)?

HallsofIvy
Feb6-08, 07:24 AM
As EnumaElish said, you don't use Newton's method to "approximate" functions- you use Newton's method to find approximate (numerical) solutions to equations. What really is the problem?

shwin
Feb6-08, 04:39 PM
Both, I am not sure how to use it for one and for the other. And yes I meant approximate solutions, I assumed it was just semantics but obviously it isnt.

HallsofIvy
Feb6-08, 05:35 PM
The difference is that to solve and equation, you need an equation! What exactly do you want to do? Perhaps you are just referring to "tangent plane approximations" to the functions (not normally called Newton's method).

The tangent plane to z= f(x,y) at (x_0, y_0), is
z= f_x(x_0,y_0)(x- x_0)+ f_y(x_0,y_0)(y- y_0)