Matrix multiplication to addition

Click For Summary

Discussion Overview

The discussion revolves around the possibility of relating matrix multiplication to matrix addition through a transformation, specifically exploring whether an equation of the form AB = PA + QB can be established. The scope includes theoretical considerations and mathematical reasoning regarding matrix dimensions and transformations.

Discussion Character

  • Exploratory
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant questions if a transformation exists that relates matrix multiplication to addition, providing an initial example of AB = PA + QB.
  • Another participant argues that such a transformation is generally not possible due to dimensional mismatches, explaining the conditions under which matrix multiplication is defined.
  • A third participant acknowledges the dimensional concerns but suggests that there may be "clever" ways to achieve a transformation, referencing the logarithm of a matrix as a potential avenue for exploration.
  • There is a proposal to consider a transformation of the form AB*u = f(A)*u + f(B)*u, where u is a vector and f is a function, indicating a shift in focus from matrices to their action on vectors.
  • One participant reiterates the idea of treating numbers as "one by one" matrices and questions whether a function f can exist to satisfy the transformation for all numbers u.
  • There is a request for clarification on whether the transformation holds, indicating uncertainty about the validity of the proposed relationship.

Areas of Agreement / Disagreement

Participants express differing views on the feasibility of the transformation relating matrix multiplication to addition. While some acknowledge the dimensional issues, others remain open to the possibility of alternative approaches, leading to an unresolved discussion.

Contextual Notes

The discussion highlights limitations related to dimensionality and the specific forms of matrices and vectors involved in the proposed transformations. The exploration of functions f and their applicability remains open-ended.

yiorgos
Messages
18
Reaction score
0
I am looking for a transformation that relates a matrix product with a matrix addition, e.g.
AB = PA + QB

Is there any such transformation?
Thnx
 
Physics news on Phys.org
In general, the answer would be no, simply because of a dimensional mismatch.
Suppose A is an n x m matrix and B is m x r (the first dimension has to be m otherwise AB does not make sense). Then AB is n x r. However, P will be p x n and Q will be q x m, for some numbers p and s (the second coordinate is fixed because the products will have to make sense). All this only works out if p = q = m = r which restricts the validity of the theorem, if it were true, quite a lot.

Oh, and if you do get the dimensions to work out, there is of course the trivial solution P = 0n x m, Q = A.
 
Thanks for the answer.
You are right about the particular transformation yet that was just a naive example I gave.

I am sure there exist "clever" ways to achieve this.

For example
http://en.wikipedia.org/wiki/Logarithm_of_a_matrix

I don't know if that helps more but what I actually need is
the below transformation.
AB*u -> f(A)*u + f(B)*u

where u is a vector and f,g() some functions (like the one in the link I gave)
 
Well, a single number can be thought of as a "one by one" matrix so the first thing you should think about is "if A and B are numbers, do there necessarily exist a function f such that ABu= f(A)u+ f(B)u for every number u?"
 
HallsofIvy said:
Well, a single number can be thought of as a "one by one" matrix so the first thing you should think about is "if A and B are numbers, do there necessarily exist a function f such that ABu= f(A)u+ f(B)u for every number u?"

So in short you are saying that this transformation doesn't hold?
 

Similar threads

  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 12 ·
Replies
12
Views
4K
  • · Replies 15 ·
Replies
15
Views
5K
  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 14 ·
Replies
14
Views
4K
  • · Replies 33 ·
2
Replies
33
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 4 ·
Replies
4
Views
4K