"Real-time Solution of Network Optimization Problems”
Event details
| Date | 18.11.2013 |
| Hour | 10:30 › 11:30 |
| Speaker | Mahmoud Jafargholi |
| Location |
PPB 017
|
| Category | Conferences - Seminars |
Solving nonlinear optimisation problems for real-time applications is a challenging task and has been restricted to applications with rather slow cycle times. The main reason for this restriction can be traced to the computational demand that optimisation-based algorithms pose to the computational hardware. Although there exists a variety of both commercial and free solvers for nonlinear optimisation problems, their rather generic nature prevents them from exploiting the problem structure inherent in some specific problem classes to solve a problem instance as efficient as possible. Furthermore, the complexity and/or licenses of standard solvers restrict their deployment on non-standard platforms such as embedded control systems. The effectiveness of structure-exploiting algorithms was recently shown for model predictive control problems with the solver FORCES (http://forces.ethz.ch). A tailored high-speed nonlinear optimisation solver for network problems will increase the range of possible applications which can be tackled with today’s hardware platforms. The aim of the thesis is to develop an efficient numerical solver for small-scale nonlinear network optimisation problems. These problems can be posed as quadratic programming problems with bilinear equality constraints. The problem data is highly sparse due to the underlying network structure. A fast linear solver needs to be developed that exploits the aforementioned structure of the problem data. The linear solver should allow integration with the interior point solver called IPOPT which is an open source interior point optimization solver.
Based on a thorough analysis of the problem data structure and state-of-the-art literature, the most suitable algorithmshould be identified/developed to efficiently solve a given class of problems. To facilitate deployment on exotic hardware platforms, the solver should be implemented in C/C++ using only standard libraries.
Based on a thorough analysis of the problem data structure and state-of-the-art literature, the most suitable algorithmshould be identified/developed to efficiently solve a given class of problems. To facilitate deployment on exotic hardware platforms, the solver should be implemented in C/C++ using only standard libraries.
Practical information
- Informed public
- Free
Organizer
- Prof. P. Ricci
Contact
- Prof. P. Ricci