An application of Bayesian networks for the optimization of a bridge layout

Abstract

Advantages in the adoption of integrated systems on a ship's bridge are quantified with respect to a more traditional bridge layout. The study is performed by modelling the causation factor for a ship-to-ship collision scenario. The assessment is provided in terms of the probability distribution of the time required to the officer on the watch (OOW) to complete a foreseen procedure aimed at avoiding a collision with another ship. The time needed to identify and agree the evasive manoeuvre with the second ship is evaluated for the two layouts, considering also a probabilistic occurrence of disturbances distracting and delaying the OOW during the procedure. Bayesian networks are employed to model the scenario. A general increment of the reactivity of the OOW in the integrated bridge is seen, represented by a shorter time to complete the procedure and by a reduced probability of being interrupted. The overall effect is an increase in the time available to put into practice the manoeuvre itself and, eventually, to avoid the collision.