Change of variables and discrete derivatives

Click For Summary

Discussion Overview

The discussion revolves around the evaluation of partial derivatives of a wavefunction defined on a grid, specifically transitioning from derivatives with respect to new variables (a, b, c) to the original variables (x, y, z). The context includes numerical methods and finite differences.

Discussion Character

  • Technical explanation
  • Homework-related
  • Mathematical reasoning

Main Points Raised

  • One participant seeks to express the derivatives d/dx, d/dy, and d/dz using the chain rule, involving derivatives d/da, d/db, and d/dc, but is uncertain about how to express da/dx, db/dx, and dc/dx in matrix form.
  • Another participant suggests looking into "stencils" in numerical analysis as a method for approximating partial derivatives from grid values.
  • A later reply clarifies that while they have obtained partial derivatives with respect to a, b, and c, they specifically want to convert these to derivatives with respect to x, y, and z.

Areas of Agreement / Disagreement

Participants express different approaches to the problem, with some focusing on the chain rule and others suggesting numerical methods like stencils. There is no consensus on the best method to transition between the variable derivatives.

Contextual Notes

The discussion highlights the challenge of expressing derivatives in terms of different variable sets, indicating potential limitations in the mathematical expressions and assumptions involved in the transformation process.

pericles
Messages
3
Reaction score
0
Hey

I am trying to evaluate d/dx, d/dy and d/dz of a wavefunction defined on a grid. I have the wavefunction defined on equally spaced points along three axes a=x+y-z, b=x-y+z and c=-x+y+z. I can therefore construct the derivative matrices d/da, d/db and d/dc using finite differences but I don't know how to convert these to d/dx, d/dy and d/dz.

I have been trying to use some sort of chain rule

d/dx=d/da*da/dx+d/db*db/dx+d/dc*dc/dx ... but I do not know how to express da/dx, db/dx and dc/dx using matrices.

Any help would be greatly appreciated
Thanks
 
Physics news on Phys.org

Homework Statement



Hey

I am trying to evaluate d/dx, d/dy and d/dz of a wavefunction defined on a grid. I have the wavefunction defined on equally spaced points along three axes a=x+y-z, b=x-y+z and c=-x+y+z. I can therefore construct the derivative matrices d/da, d/db and d/dc using finite differences but I don't know how to convert these to d/dx, d/dy and d/dz.

I have been trying to use some sort of chain rule

d/dx=d/da*da/dx+d/db*db/dx+d/dc*dc/dx ... but I do not know how to express da/dx, db/dx and dc/dx using matrices.

Any help would be greatly appreciated
Thanks

Homework Equations


a=x+y-z,
b=x-y+z
and c=-x+y+z
x=½(a+b)
y=½(a+c)
z=½(c+b)

The Attempt at a Solution



I have been trying to use some sort of chain rule

d/dx=d/da*da/dx+d/db*db/dx+d/dc*dc/dx ... but I do not know how to express da/dx, db/dx and dc/dx using matrices.

I tried using [(d/da*x)]^(-1) but it does not work

any help would be greatly appreciated.
Thanks
 
If the problem is to approximate the partial derivatives of a smooth function from the values of the function on a grid, you should look up the topic of "stencils" in numerical analysis. Search on keywords like "stencil, derivatives, 3D".
 
Stephen Tashi said:
If the problem is to approximate the partial derivatives of a smooth function from the values of the function on a grid, you should look up the topic of "stencils" in numerical analysis. Search on keywords like "stencil, derivatives, 3D".

My problem is somewhat different...

I have the partial derivatives with respect to a, b and c but would like partial derivatives with respect to x, y and z.

I have used a stencil (I think) in order to get the partials with respect to a, b and c.
 

Similar threads

  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 6 ·
Replies
6
Views
4K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 10 ·
Replies
10
Views
3K
Replies
1
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K