Laboratorio de Pronóstico Inteligente y Sistemas Estructurales Ciber-físicos

This paper proposes a new mathematical methodology to model expert systems with the ability to sequentially learn from data. To this end, the Plausible Petri nets (PPNs) methodology, first developed in M. Chiachío et al. [Proceedings of the Future Technologies Conference, San Francisco, (2016), pp. 165-172] is used due to their ability to integrate continuous and discrete dynamics in a single net model, which allows us to analyse hybrid systems with interaction of diverse sources of information, like in expert systems. The efficiency of the proposed approach is demonstrated in an expert system model for railway track inspection management taken as case study using published data from a laboratory simulation of train loading on ballast, carried out at the Nottingham Railway Test Facility, University of Nottingham.