№1 (28), 2014. INFORMATICS AND MODELING

I.S. Kabak

Neuronetwork model for forecasting and the assessment of reliability of the software

The  model  developed  by  the  author  is  presented  in  article  for  forecasting  and  an  assessment  of  reliability  of  thesoftware on the example   of automated  control systems. The  model  is realized on the basis of methods  of artificial intelligence, with use of the device of artificial neural networks. On an entrance of model debugging time moves, at the exit the forecast for value of failure rate is formed. The neural network of a special look is developed for realization of model — a  vertically  layered  neural  network.  Increase  of  accuracy  of  model  leads  to  building  of  layers  in  a  neural network. Realization of an artificial neural network of this type — is considered to a modular way. The main results of article  are  protected  by  patents  of  the  Russian  Federation  for  inventions  and  useful  models. 

 

Keywords: simulating, reliability of the software, artificial neural networks..

[ Back to contents ]
[ Get the article ]