Forschungsthemen
[MA] Parameter Tuning in Combinatorial Search Spaces
The problem of software variant selection and hardware resource allocation is currently being tackled by two types of the approaches. The first type uses model transformation techniques and creates an integer linear program to solve it, while the second type uses metaheuristic-based solvers, which are optimized at design time by a parameter tuning software product line. The goal of this thesis is to increase the application area of parameter tuning by applying it directly to the optimization problem. The research objective is to investigate whether a parameter tuning approach can go beyond parameter optimization and be fully or partially applicable for tasks such as combinatorial optimization based on the problem of software variant selection and hardware resource allocation.
For this thesis, the following tasks have to be fulfilled:
- Literature analysis covering closely related work.
- Development of a strategy for embedding the flow of solving the optimization problem into the parameter tuning software product line.
- Design of a search space for a parameter tuning and its navigational rules.
- Implementation of the aforementioned concepts.
- Evaluation of the developed approach with a problem of software variant selection and hardware resource allocation.
Betreuer: Dmytro Pukhkaiev