pyBNBowTie: Python library for Bow-Tie Analysis based on Bayesian Networks

Abstract In addition to more conventional but often less precise methods, the risk assessment as part of the risk management process can be performed with the bow-tie analysis method. A bow-tie analysis describes the effects of causes on a top event and the resulting consequences. Bayesian networks, on the other hand, offer a mathematically concise way of describing dependencies between events under uncertainty. The mapping of bow-tie analysis into Bayesian networks is intended to make their superior calculation options available. While the mapping algorithm of a bow-tie method into a Bayesian network is described in the literature, no computer program carrying out this mapping has been found so far. In this text, a Python library, that is validated using published examples, is presented and made publicly available for mapping bow-tie methods into Bayesian networks.