Determining the exact reliability of a network by computing its associated reliability polynomial is known to be a difficult task (#P-complete in general). That is why plenty of approximate solutions have been proliferating since quite some time, and artificial neural network approaches have been suggested. Still, learning the exact coefficients of the associated reliability polynomials is an approach that has not yet been pursued. One would need to use a database for training, while, unfortunately, none seem to be available. That is why in this work, we introduce an algorithm that counts the number of shortest paths from the input (start) to the output (terminus) for a particular class of two-terminal networks known as hammocks. These numbers are the first non-zero coefficients of the reliability polynomials of hammock networks, hence a first step towards building a training database. We shall, first of all, prove the optimality of the algorithm and we shall use it to generate a first training dataset, while targeting the second coefficient as the next step.
Short Bio:
Carmen Țerei is currently enrolled in the first year of PhD at the West University Timișoara. Carmen holds a Bachelor’s Degree in Computer Science and a Master’s Degree in Economic Informatics. Her professional background combines academic teaching and research with industrial experience in automotive software engineering, where she has been involved in software development, testing and requirements engineering. Her current research focuses on the application of artificial intelligence methods to the reliability analysis of complex networked systems.
