supported through study contributions %SBBANNER%
- 11 April added solution hints
- 21. May schedule updated again
- 30. May schedule updated
DescriptionIn this tutorial, you will learn to apply the theory from the lectures Fundamentals of Convex Optimization (MA2504) und Discrete Optimization (MA 3502) to practical examples. Using a professional MILP-solver (FICO Xpress) you will practice how to transform a given problem into a Mixed-Integer Linear Programming model, to input the model to the solver and to tailor the solution process to certain properties of the problem. The tutorial is reccomended as a preparation for advanced optimization classes (especially Case Studies Discrete Optimization) as well as Bachelor's and Master's theses. Agenda You will get to know the Mosel programming language that will be used throughout the course. Some linear optimization problems will serve to practice and deepen your knowledge. We then move on to problems of discrete optimization and explore the different possibilities to control the optimization process using cutting planes and Branch & Bound. Throughout the course, the focus lies on practical problems with some brief theory revisions in between. You will work in small groups of 2-3 students to solve the problems and implement the solutions.
- Fundamentals of Convex Optimization (MA2504) (or the previously offered Lineare Optimierung (MA3501))
- for the second part: Discrete Optimization (MA 3502) and/or Combinatorial Optimization (MA4502)
- basic programming skills
RegistrationRegistration for the course is still possible on the holiday course website.
|Wednesday, 6 April 2016||10:00 - 12:00 and 13:00 - ca. 17:00||02.04.011||Introduction to Mosel, programming exercises Linear Optimization|
|Thursday, 7 April 2016||9:00 - 12:00 and 13:00 - ca. 17:00||02.04.011||Introduction to Mosel (part II), Discrete Optimization: Cutting Planes|
|Friday, 8 April 2016||9:00 - 12:00 and 13:00 - ca. 16:00||02.04.011||Discrete Optimization: Branch & Bound|
PreparationsDuring the tutorial, you will work on programming problems in small groups of 2-3 students. As we cannot provide enough laptops, we ask all participants to bring their own laptops. You should try and install the software FICO Xpress (see next section) ahead of the first day of the tutorial as we will start with the first excercises right away. Should you encounter any problems, please contact us as soon as possible. Furthermore, you should make sure that you can access at least one of the Wi-Fi networks available in the building as we will provide you with documents and data files during the course.
|Mosel Reference Sheet:|
|Mosel Cutting Planes:|
|Mosel Branch and Bound:|
|Exercise||Problem Sheet||Data||Templates||Solution Hints|
|0||Crude Oil Refinement|
Installation of course softwareIn the tutorial, we use the commercial software FICO Xpress.
WindowsFor Windows systems a restricted version for educational purposes is available free of charge. This version will suffice for the purposes of the tutorial and can be downloaded from https://community.fico.com/download.jspa .
Linux/MacUnfortunately, no free-of-charge version is available for these systems. Users can choose from these options:
- share a laptop with a fellow student (recommended)
- use Wine (Linux: http://www.winehq.org/download , Mac: http://wiki.winehq.org/MacOSX/ , Version 1.4) to run the Windows version of FICO Xpress
- set up a virtual machine (e.g. from Maniac) running Windows and use the Windows version of FICO Xpress.
- A restricted version of FICO Xpress for educational purposes can be downloaded from https://community.fico.com/download.jspa (unfortunately Windows-only).
- An overview of Xpress Mosel is available here , some examples can be found here .
- Other commercial optimization software:
- Free optimization software:
- An overview of optimization tools (commercial, free-of-charge und open source) for different problems is provided by Decision Tree for Optimization Software .