Practical Distributed Optimization in Julia
Credits: 5 ECTS
Term: Summer 23
Lecture & Exercise (block course)
Lecturers: Prof. Dr.-Ing. Timm Faulwasser, Dr.-Ing. Alexander Engelmann,
Schedule: 12.04. | 26.04. | 10.05. | 24.05. | 26.07.-28.07.
Lecture hall: April & May: CT Geschossbau III - G3 425 (Seminarraum); July: Elektrotechnik - 3.21
Examination: Take-home project and oral exam
Enrollment: Please sign up for the course on LSF. You will then be enrolled in Moodle automatically.
Office hours: On demand
For further details, please check out our moodle course.
Majors and Module Numbers:
M.Sc. Electrical Engineering and Information Technology: Module ETIT-405
M.Sc. Automation and Robotics: Module
Numerical optimization methods form the basis of many of today's technical systems. Many of these systems are composed of subsystems, which make classical optimization hard to apply simply because of the problem size or because of additional requirements such as a limited information exchange. This course covers distributed and decentralized optimization algorithms overcoming limitations of classical centralized optimization. The central idea in these algorithms is to shift the computational burden to the subsystems. The course will cover
- basics on nonlinear programming and convex optimization;
- dual decomposition;
- the alternating direction method of multipliers (ADMM);
- decomposition strategies for sequential quadratic programming and interior point methods;
- additional topics such as primal decomposition methods for machine learning (if time permits).