Probabilistic analysis of supply chains resilience based on their characteristics using dynamic Bayesian networks

Student thesis: Doctoral Thesis

Abstract

There is an increasing interest in the resilience of supply chains given the growing awareness of their vulnerabilities to natural and man-made hazards. Contemporary academic literature considers, for example, so-called resilience enablers and strategies, such as improving the nature of collaboration and flexibility within the supply chain. Efforts to analyse resilience tend to view the supply chain as a complex system. The present research adopts a distinctive approach to the analysis of supply resilience by building formal models from the perspective of the responsible manager. Dynamic Bayesian Networks (DBNs) are selected as the modelling method since they are capable of representing the temporal evolution of uncertainties affecting supply. They also support probabilistic analysis to estimate the impact of potentially hazardous events through time. In this way, the recovery rate of the supply chain under mitigation action scenarios and an understanding of resilience can be obtained. The research is grounded in multiple case studies of manufacturing and retail supply chains, involving focal companies in the UK, Canada and Malaysia, respectively. Each case involves building models to estimate the resilience of the supply chain given uncertainties about, for example, business continuity, lumpy spare parts demand and operations of critical infrastructure. DBNs have been developed by using relevant data from historical empirical records and subjective judgement. Through the modelling practice, It has been found that some SC characteristics (i.e. level of integration, structure, SC operating system) play a vital role in shaping and quantifying DBNs and reduce their elicitation burden. Similarly, It has been found that the static and dynamic discretization methods of continuous variables affect the DBNs building process. I also studied the effect of level of integration, visibility, structure and SC operating system on the resilience level of SCs through the analysis of DBNs outputs. I found that the influence of the integration intensity on supply chain resilience can be revealed through understanding the dependency level of the focal firm on SC members resources. I have also noticed the relationship between the span of integration and the level of visibility to SC members. This visibility affects the capability of SC managers in the focal firm to identify the SC hazards and their consequences and, therefore, improve the planning for adverse events. I also explained how some decision rules related to SC operating system such as the inventory strategy could influence the intermediate ability of SC to react to adverse events. By interpreting my case data in the light of the existing academic literature, I can formulate some specific propositions.
Date of Award26 Oct 2016
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SupervisorRobert van der Meer (Supervisor) & Lesley Walls (Supervisor)

Cite this

'