The reliability algorithm based on binary neural networks

Abstract

According to the limitation in the research of reliability, this paper, starting from the generality of complex systems, firstly converts the system components into the linear combination by training the binary neural network, secondly we solve the distribution function of the linear combination. With this method, this paper finally gets a general method to analyse the reliability of arbitrary systems. Finally, this paper gives its theoretical guarantee and this method is also validated through examples.