CSCI 6961 : Geometric Optimization for Robotics

M-R 2-3:20pm
Location: MRC 334 (conf. room)
Instructor: Ibrahim Volkan Isler
Tentative schedule
Summary:
Robots and sensor-networks are inherently tied to the physical world. Consequently, most optimization problems which arise in these fields can be solved using geometric techniques. This course will cover a variety of techniques that should be useful for robotics and sensor-networks researchers. The list of topics include:

Part I: Basic Geometric Optimization Problems
  • Point Location
  • Convex Hulls
  • Linear Programming
  • Proximity queries (Voronoi Diagrams, Triangulations, ...)
  • Geometric Searching
  • Sampling and shape approximation

Part II: NP-hard Geometric Optimization Problems
  • Covering and partitioning
  • Network design: steiner tree problem and variants
  • Routing: Traveling Salesperson Problem and Variations

Part III: Selected Topics
  • an assortment of  robotics and sensor-networks papers that utilize geometric optimization techniques

Suggested texts:
  • Computational Geometry: Algorithms and Applications by de Berg, van Kreveld, Overmars and Schwarzopf
  • Approximation Algorithms by Vazirani
Prerequisites:

CSCI-4020 Computer Algorithms, or equivalent.
Grading:
  • Project: 60%. The project consists of literature & background review + lecture and solution of  a novel/open problem and a presentation. For the first part, you should prepare a webpage with pointers and references to relevant papers, slides etc. For the second part, a term report is required. We will discuss the project in detail in the first class.
  • Take-home Final: 30%
  • Participation: 10%



For more information, please contact Volkan Isler.