|Topic:||Evoluce heuristických optimalizačních algoritmů|
|Department:||Analýza a interpretace biomedicínských dat|
|Supervisor:||Ing. Petr Pošík, Ph.D.|
|Announce as:||Diplomová práce, Bakalářská práce, Semestrální projekt|
|Description:||Heuristic search algorithms (e.g. Nelder-Mead Simplex Search) often have some hard-coded internal strategy parameters, or strategy components, which may not be set up optimally with respect to the target class of problems. The goal of this project and subsequent bachelor project is to explore the possibilities of evolving instances of algorithms specialized to a particular class of problems.
* Choose a particular heuristic algorithm and design its representation/parameterization that is suitable for evolution.
* Choose several sets of functions/problems representing different target problem classes.
* Design/choose a suitable measure describing the performance of an algorithm instance on a set of problems.
* Design an evolutionary algorithm to evolve heuristics specialized to the chosen problem class.
* Explore the differences of the evolved algorithms from the original generic version.
|Bibliography:||Will be provided by the project supervisor.|