1. Viktor Burkovsky, Voronezh State Technical University, Russian Federation
2. Semen Podvalny, Voronezh State Technical University, Russian Federation
3. Mitar Lutovac, FM Herceg Novi , Montenegro
4. Konstantin Gusev, Voronezh State Technical University, Russian Federation
in this paper we consider the problem of optimal scheduling of queueing networks with switching times based on genetic algorithms approach. We propose a special chromosome encoding scheme, crossover and mutation operators. Deterministic mutation is used to prevent premature convergence to a poor solution. Finally, results of computational experiments with distributed components are included to compare deterministic mutation with classical probabilistic one. A system of distributed components is usually configured once the once only during the build-time when known and tested versions of components are used. In practice, large scale applications, especially those from scientific and engineering domain, are highly data centric and the computations are organized around data specific to the problem being solved. To mature software development, engineering methods, such as configuration management architecture and development models related to the component development paradigm, must be introduced.
Ključne rečitestttt :