The High Performance Optimization Lab focuses on the practical study of optimization problems and on their solution using state-of-the-art methods and modern computing devices.
A typical example is the examination of a practical combinatorial or continuous/numerical optimization problem that is to be solved on real-world instances (e.g., stemming from industry), and either using or developing algorithms or solvers that are tailored to exploit e.g. parallel hardware.
The participants are expected to be able to work under a Unix system environment (e.g. Linux, it is okay to use a virtual machine) and to have some first experience in programming (ideally, with C/C++).
There are two topic options (also depending on the number of participants), the first of which is currently preferred:
1st Topic Option: We will study the Capacitated Vehicle Routing Problem: Given a network with travel costs, a depot, a set of customers with demands, and a vehicle capacity, calculate optimal routes to serve all customers. Here, we look at (extensions of) exact and heuristic algorithms and an automated evaluation of the solution quality. The problem and potential approaches will be introduced in the beginning of the lab. If we are successful, we may even look at some instances from the DIMACS Implementation Challenge.
2nd Topic Option: We will study a Car Wiring Optimization problem that occurs in practice in the fabrics of large-scale manufacturers. Here, a special focus will be on dynamic programming and min-cost flow algorithms. So it is a plus if the applicants are interested in and already comfortable with these techniques.
For planning purposes and since the number of participants is limited, please let us know if you are interested in joining this lab by writing an email to smallach [at] cs.uni-bonn.de until April 6th the latest. Please also briefly highlight your previously gained experience (e.g., lectures, exercises, PGs and other LABs, theses, work in industry) in the mentioned required and/or desired fields.
You will then receive information to join the first meeting which is tentative, i.e. comes with no obligation to you. You may still decide whether to join the Lab afterwards.
The lab will be held in digital manner only, i.e., especially meetings will take place online via video conferences.
Dozent: Dr. Sven Mallach