What is the definition of log(P) in the von neumann entropy formula?

Click For Summary

Discussion Overview

The discussion revolves around the definition of log(P) in the context of the von Neumann entropy formula, specifically addressing how this logarithmic operation is interpreted when applied to a density matrix P. The scope includes theoretical aspects of quantum mechanics and entropy calculations.

Discussion Character

  • Technical explanation
  • Conceptual clarification

Main Points Raised

  • One participant defines the von Neumann entropy as S(P) = -tr(P*log(P)) and questions the definition of log(P).
  • Another participant provides a series expansion for ln(P) as ln(P) = -∑(1/n)(I-A)^n, suggesting a mathematical approach to the logarithm of the density matrix.
  • A later reply reiterates the independence of the trace from representation, stating that if P is diagonalized with eigenvalues {k}, then S(P) can be expressed as H({k}), where H is the Shannon entropy.
  • Another participant presents an alternative definition of entropy as S:=-k⟨ln(ρ)⟩ρ, drawing a parallel to Gibbs' entropy.

Areas of Agreement / Disagreement

Participants express differing views on the definition of log(P) and its implications for calculating von Neumann entropy. There is no consensus on a singular definition, and multiple interpretations are presented.

Contextual Notes

Some definitions depend on the representation of the density matrix, and the discussion includes various mathematical formulations that may not be universally applicable without further context.

trosten
Messages
47
Reaction score
0
Let P be a density matrix. Then the von neumann entropy is defined as
S(P) = -tr(P*log(P))

But how is log(P) defined ?

--edit--
found the answer. the trace is independent of representation so that if P is diagonalized with eigenvalues {k} then S(P) = H({k}) where H is the shannon entropy.
 
Last edited:
Physics news on Phys.org
[tex]ln(P) = -\sum_{n=1}^\infty \frac{1}{n}(I-A)^n[/tex]
 
trosten said:
found the answer. the trace is independent of representation so that if P is diagonalized with eigenvalues {k} then S(P) = H({k}) where H is the shannon entropy.
right. Usually easier by the way !
 
The definition is really

[tex]S:=-k\langle \ln\hat{\rho}\rangle_{\hat{\rho}}[/tex]

,quite similar to Gibbs' entropy.

Daniel.
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 6 ·
Replies
6
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 3 ·
Replies
3
Views
3K