OPTIMIZATION Course

Quadratic problems

Go back

A quadratic problem is something like this:

@ \min f(x) = \frac{1}{2} X^t * A * X + b^T X @

with

  • X the vector for your variables like x,y,...
  • A the matrix in Ax=b
  • b the vector of results in Ax=b

After some calculations [hidden], you got these formulas

  • $\nabla f(X) = AX-b$
  • $H_{f}(X) = A$

I have seen it use quite a lot in constrained optimization (on the web) but we didn't work on this enough, so I don't know much.