Inclusive risk modeling for manufacturing firms: a Bayesian network approach

This paper focuses on modelling the enterprise level risks from the perspective of an original equipment manufacturer. We intend to converge on an overall risk measure that is representative of the cumulative effect of risks emanating from considerations pertaining to respective functional divisions within the enterprise. Further, due to multitude of interplays between the core objectives of various functional divisions, modeling the cumulative risk pertaining to any project within a firm presents significant challenges. This paper proposes a systematic risk assessment methodology considering various enterprise specific risk characteristics (primarily technical, commercial, and operational in nature) related to multiple functional divisions of an enterprise. Specifically, we consider six different functional divisions i.e. planning, sourcing, operations, marketing, logistics and service. A Bayesian network model is then evolved by mapping the risk parameters related to various functional divisions and their interdependencies. Further, each of these risk parameters are represented in terms of parent and root nodes. In order to determine the probabilities of existing nodes in a Bayesian network, a methodical approach is developed that focuses on obtaining the conditional probabilities of the nodes with multiple parents. Thereafter, an enterprise level value chain risk measure is proposed that evaluates the feasible risk states in terms of an aggregate risk number. Employing an example of a typical automotive company, the methodology is illustrated.