Jedno rješenje primjene genetskog algoritma za analizu protokola usklađivanja vremena u TTEthernet mrežama
Safety-critical systems like airplanes and cars demand high-reliable communication between components within the system, which is achieved by using deterministic networks. Proper establishing and maintenance of synchronization of device clocks in the network components represents one of crucial aspects in deterministic networks where belong TTEthernet as well. If device clocks are not synchronized, deterministic communication is not feasible. Keeping in mind that most critical information has been exchanged between the network components using deterministic traffic class, it is obvious that such services will not be available until the clocks in the network are synchronized. The thesis deals with estimation of worst-case startup time for observed TTEthernet network, in case that one device in the network is under failure. The estimation is performed by OMNeT++ simulations and using genetic algorithm. The simulations show that startup time of the network is extended significantly under impact of faulty component. Also, unavailability of most critical services in the network is extended for the same time. For the network simulated in this thesis, estimated median value equals 489579 μs for worst-case startup time.
https://www.cris.uns.ac.rs/DownloadFileServlet/Disertacija163412279787293.pdf?controlNumber=(BISIS)