Quantitative modeling and analysis of supply chain risks using Bayesian theory

Abstract

Purpose – Globally expanding supply chains (SCs) have grown in complexity increasing the nature and magnitude of risks companies are exposed to. Effective methods to identify, model and analyze these risks are needed. Risk events often influence each other and rarely act independently. The SC risk management practices currently used are mostly qualitative in nature and are unable to fully capture this interdependent influence of risks. The purpose of this paper is to present a methodology and tool developed for multi-tier SC risk modeling and analysis. Design/methodology/approach – SC risk taxonomy is developed to identify and document all potential risks in SCs and a risk network map that captures the interdependencies between risks is presented. A Bayesian Theory-based approach, that is capable of analyzing the conditional relationships between events, is used to develop the methodology to assess the influence of risks on SC performance Findings – Application of the methodology to an industry case study for validation reveals the usefulness of the Bayesian Theory-based approach and the tool developed. Back propagation to identify root causes and sensitivity of risk events in multi-tier SCs is discussed. Practical implications – SC risk management has grown in significance over the past decade. However, the methods used to model and analyze these risks by practitioners is still limited to basic qualitative approaches that cannot account for the interdependent effect of risk events. The method presented in this paper and the tool developed demonstrates the potential of using Bayesian Belief Networks to comprehensively model and study the effects or SC risks. The taxonomy presented will also be very useful for managers as a reference guide to begin risk identification. Originality/value – The taxonomy developed presents a comprehensive compilation of SC risks at organizational, industry, and external levels. A generic, customizable software tool developed to apply the Bayesian approach permits capturing risks and the influence of their interdependence to quantitatively model and analyze SC risks, which is lacking.