Seminar 9

Dynamic programming

  • 11-01, the rest
  • 11-04. (It might be a bit more quirky than 11-02, so you might want to look at that too.)

 

One full exam set: 2017

(As suggested in the draft. Exam set: https://www.uio.no/studier/emner/sv/oekonomi/ECON4140/oldexams/M3x2017.pdf Links to an external site. . Solution: put "SOLVED" before .pdf. Beware that my handwritten solution is a bit terse - as stated in the typeset part, I would recommend you to spend more ink.) 

 

Convexity/concavity [copied from 2020, this part unchanged since the draft assignment]

  • Show that the intersection of an arbitrary family of convex sets, is convex.
    • Warning: it may be an arbitrarily large family, in particular larger than what you can put in a sequence, so an induction argument is useless!
    • Notation: if J is an arbitrary index set, and for each j in J we have a (convex set), then we can write the family as LaTeX: \{S_j\}_{j\in J}\(\{S_j\}_{j\in J}\) and the intersection  set as LaTeX: \bigcap_{j\in J}S_jjJSj 
  • Use the epigraph definition to show that the maximum of two convex functions (on the same convex domain D) is a convex function.
    • Trick question for the curious: Let LaTeX: \{f_j\}_{j\in J}{fj}jJ be a family of convex function on the same convex domain D. Does it follow that the smallest f that is LaTeX: \geq f_jfj  for all j, is a convex function?
  • The convex hull of a set A is "the smallest convex set that is LaTeX: \supseteq AA". That means, the intersection of all sets S that have the following two properties: S is convex and LaTeX: S \supseteq ASA .  
    • The phrase "convex hull" surely suggests that this is a convex set - it would be quite bad to give it that name if not. Show that the convex hull is indeed convex. 
  • Let f(t) = t + sin t. 
    • Is f quasiconvex? Is f quasiconcave?
    • For what real constants c will f(t)+ct be quasiconcave? For what c will f(t)+ct be quasiconvex?