Instructions: In-class project 1.

  • Work in a team of 2 or 3 people.
  • Choose one application of either semidefinite programming or conic programming. You may use one of the references below or any other reference.
  • Tell me which application your team is considering, so that each team works on a different one.
  • Prepare a few slides to explain the application to your classmates. (Use powerpoint, latex, handwriting, or whatever you prefer.)
  • Email the slides to me (oleary@cs.umd.edu) or hand them in. Make sure the team members' names are on the first slide.
  • Then do the same for a second application. (Don't choose two from the Vandenberghe & Boyd paper.)
  • Then we will talk about some of these applications.

    • Some resources:

  • Semidefinite Programming website
  • Applications of semidefinite programming
  • Lattice packing ** Slides
  • Vandenberghe & Boyd paper:
  • Structural optimization
  • "Pattern separation by ellipsoids" (Support Vector Machines) ** Slides and Slides
  • Control and System Theory
  • "Unit commitment" in power scheduling ** Slides
  • Layout of high-dimensional data ** Slides
  • The max-cut problem ** Slides and Slides and supplement
  • l_p norm optimization
  • linear precoding for MIMO receivers ** Slides
  • Applications of conic programming
  • Generalization of traveling salesperson problem ** Slides
  • Financial applications ** Slides

    • SDP test problem:

  • torusg3-8.mat test problem
    obtained from from this webpage
  • lib.pdf solution: see pp. 4-5

    • Software Reviews

  • MAXDET
  • SDPA-M
  • SDPT3
  • SeDuMi
  • DSDP
  • YALMIP: installs
  • SPARSEPOP: missing start file