Total potential energy in matrix

Click For Summary
SUMMARY

The discussion focuses on the properties of the symmetric matrix A = [3 -1; -1 3] in relation to quadratic forms and linear equations. It concludes that the quadratic form x'*A*x cannot evaluate to zero for any non-zero vector x, confirming that A is invertible due to a non-zero determinant. Furthermore, it establishes that the total potential energy function f(x) = 1/2*x'*A*x - b'*x achieves a minimum at the unique solution of the equation A*x = b, which can be derived using the Jacobian and Hessian of f.

PREREQUISITES
  • Understanding of quadratic forms and their properties
  • Knowledge of matrix invertibility and determinants
  • Familiarity with optimization concepts, including Jacobians and Hessians
  • Proficiency in MATLAB notation for matrix operations
NEXT STEPS
  • Study the properties of symmetric matrices and their quadratic forms
  • Learn about matrix determinants and conditions for invertibility
  • Explore optimization techniques involving Jacobians and Hessians
  • Practice solving linear equations using matrix inverses in MATLAB
USEFUL FOR

Mathematicians, engineers, and students in applied mathematics or optimization who are working with quadratic forms and linear algebra concepts.

omgitsroy326
Messages
29
Reaction score
0
Consider the matrix A=[3 -1; -1 3]. (a) Can the quadratic form x'*A*x evaluate to zero for some non-zero vector x? (x is a 2x1 column vector, and x' means x transposed: usual Matlab notation). (b) Does the equation A*x=b, b arbitrary, have a unique solution? If it does, prove it. (c) Show that the total potential energy f(x)=1/2*x'*A*x-b'*x has a minimum for x that solves the linear equations A*x=b.

This is the question which I'm approached with
my answer is as following
a) No, because it's a symetric matrix only x'Ax = 0 only if vector x = 0.
b) Yes , it is not underdetermined and det does not equal 0
c) iono

i was wondering what you guys thought
 
Last edited:
Physics news on Phys.org
a) Compute [itex]x^tAx[/itex] and you will obtain an expression in terms of the entries of x, and this expression, as is given, will be a quadratic form. With this expression, you should easily be able to see that unless x is the zero vector, the expression will not be equal to zero (if x is to have real entries).

b) Compute det(A). Note that it is non-zero. Therefore, A is invertible, so there exists an inverse of A, [itex]A^{-1}[/itex]. If [itex]Ax = b[/itex], then [itex]A^{-1}Ax = A^{-1}b[/itex], so [itex]x = A^{-1}b[/itex], so clearly x exists and is unique.

c) The best way I can think to do this is to compute f(x) to obtain an expression in terms of the entries of x. Essentially, f is a function from [itex]\mathbb{R}^2[/itex] to [itex]\mathbb{R}[/itex]. Compute the Jacobian of f at x, and find the x such that the Jacobian is zero. For all those points, compute the Hessian, and show that there is only one that has a positive definite Hessian. Calculate the value of f at this point. Also, calculate the value of [itex]f(A^{-1}b)[/itex], and show that the two values are the same, thereby showing that the vector where f has it's minimum is indeed the unique solution to [itex]Ax = b[/itex].

If you are unfamiliar with any of the terminology above, you can easily look it up at wikipedia.com or mathworld.com, and then, if you're still stuck, ask for further clarification.
 
wow.. thanks.. that was very detailed
 

Similar threads

  • · Replies 33 ·
2
Replies
33
Views
4K
  • · Replies 24 ·
Replies
24
Views
3K
  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
5K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 7 ·
Replies
7
Views
5K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 34 ·
2
Replies
34
Views
3K