Object Oriented Bayesian Network for complex system risk assessment

In this communication, we present a new way of modelling risk management process in a large scale, complex and dynamic system using an extended object oriented Bayesian network (EOOBN) as the underlying mathematical tool. The communication begins by presenting the proposed extension that permits to take into account the necessity of varying parameters in object oriented Bayesian network (OOBN) and dynamics for practical purposes. The second part is devoted to adaption of existing information propagation algorithms in classical Bayesian network (BN) and dynamic Bayesian network (DBN) to this extended model. A small practical case study is used along the communication to illustrate each modelling step.