Dear students,
let me inform you that unfortunately I had some typos in the final BFGS formula (the one for the update of the inverse of the Hessian). Slides with a corrected formula have been uploaded.
By the way, the origin of the typo is the otherwise celebrated and recommendable monograph on Linear and Nonlinear Programming by Luenberger. Only after hitting my head against the wall while writing the code for the solver myself did I check the errata (chapter 10, page 294). I recommend doing the same on some other occasions if you run into troubles.
Z.H.