Lecture 3.2: "real" quadratic functions and how they work
Aggregazione dei criteri
Quadratic optimization: the real (nonseparable) case. (Epi)graph, (sub)level sets and tomography of quadratic (homogeneous) functions, the several different cases. Characterising level sets of quadratic (homogeneous) functions out of the spectral decomposition of Q. Consequence: characterising when a quadratic (homogeneous) function has minimum and maximum and where it lies. Extension to the box-constrained case and why it is *not* a good idea. Computing the "center" of the level sets of a quadratic non-homogeneous function: the simple (nonsingular) case.