Designing distributed real time systems is an extremely complex process and manually hardly optimizable. To automize the cost-related optimization by a clever selection of the necessary components, while fulfilling all real-time constraints, different stochastic optimization algorithms were adapted to the problem and tested on example systems by using a prototypic implementation. Good results were achieved by using the algorithms "Great Deluge" and "Threshold Accepting", while the "Genetic Algorithms" seem to be less suitable. Additionally, a deterministic heuristics is proposed which achieves good results with only few effort.
«
Designing distributed real time systems is an extremely complex process and manually hardly optimizable. To automize the cost-related optimization by a clever selection of the necessary components, while fulfilling all real-time constraints, different stochastic optimization algorithms were adapted to the problem and tested on example systems by using a prototypic implementation. Good results were achieved by using the algorithms "Great Deluge" and "Threshold Accepting", while the "Genetic Algor...
»