Supply chain risk management modelling: Bayes networks approach and an application in the fuel distribution sector

The phenomenon of uncertainty is a prominent factor that prevent supply chains from achieving their goals. When the complexity of supply chains increases, the need to explore risks and develop tailored strategies becomes very crucial for surviving in the competitive environment. In this study, we propose a conceptual model based on Bayes networks and content analysis for managing the downstream risks of supply chains. The proposed model consists of four stages and twelve steps. It was applied at Opet Oil Company in 2019 as a case study. In this study, customer complaints are evaluated as an information source for the risks occurring in the lower echelons of the supply chain and signals for future risks. In the content analysis, two breakdown risks and nine pre-risks associated with these two risks and three risk sources are discovered. In the Bayesian network, these indicators are defined as random variables and the interdependencies between those variables are structured. Thus, the prior probabilities of the risk variables are calculated. Afterward, various evidence is added to the model under six different scenarios and it is observed how the posterior probabilities changed. In addition, the sensitivity of these probabilities with respect to the level of responsiveness is calculated. As a result, five different response strategies are proposed taking into account the magnitude and sensitivity of posterior probabilities of breakdown risks. We believe that the developed model can be used as an early warning system against supply chain breakdowns, as well as supporting operational and tactical decisions in the supply chain.