Modèlisation et prédiction de la fiabilité des logiciels et des réseaux
Modeling and Prediction of Software Reliability and Network Reliability
This dissertation research attempts to explore on the one hand, models for software reliability prediction in term of cumulative failure in the software, on the other hand, models
for networks reliability evaluation or any other system which can be modeled as a network
(directed or not).
Artificial neural networks and the Auto-regression methods have been used to predict the
cumulative software failure. These methods are trained by evolutionary and simulated annealing algorithms. The developed approaches are qualified as non-parametric models. Numerical results show that both the goodness-of-fit and the next-step-predictability of our
proposed approaches have greater accuracy in predicting software cumulative failure compared to other approaches.
For evaluating the reliability of the networks we propose two ecient algorithms. The first
one for enumerating minimal pathsets in directed networks and the second one for enumerating minimal cutsets in non directed networks. These algorithms are coupled with the
inclusion-exclusion principle for computing the reliability of a network. Both algorithms
are tested with a set of networks proposed in the literature and give interesting results in
terms of accuracy and execution time.
File listTips: You can preview the content of files by clicking file names^_^
|THESE_BENADDY.pdf||1.92 MB||21-12-13 16:12|
(Add your comment, get 0.1 Point)
- Page 1
- Total 1