Řešení optimalizačních inženýrských úloh pomocí efektivních simulačních metod

Abstract

This master thesis focuses on using Cross-entropy method as a tool for solving optimalization problems. The aim of the work is the application of Cross-entropy method on the Credit Risk Economic Capital model, continuous optimization problems and combinatorial optimization problems, specifically the traveling salesman problem, max-cut problem and the sequence alignment problem. An important part of this thesis is the parallel implementation of the aforementioned tasks on the GPU. The main tools for implementing all mentioned problems are Matlab 2014b and CUDA 6.0 toolkit.