Evaluation of control strategies for managing supply chain risks using Bayesian Belief Networks

Abstract

Ekici, Şule Önsel (Dogus Author) -- Conference full title: 2015 International Conference on Industrial Engineering and Systems Management (IESM) : October 21-23, 2015, Seville, Spain.Supply chains have become complex and vulnerable and therefore, researchers are developing effective techniques in order to capture the complex structure of the supply network and interdependency between supply chain risks. Researchers have recently started using Bayesian Belief Networks for modelling supply chain risks. However, these models are still focused on limited domains of supply chain risk management like supplier selection, supplier performance evaluation and ranking. We have developed a comprehensive risk management process using Bayesian networks that captures all three stages of risk management including risk identification, risk assessment and risk evaluation. Our proposed new risk measures and evaluation scheme of different combinations of control strategies are considered as an important contribution to the literature. We have modelled supply network as a Bayesian Belief Network incorporating the supply network configuration, probabilistic interdependency between risks, resulting losses, risk mitigation control strategies and associated costs. An illustrative example is presented and three different models are solved corresponding to different risk attitudes of the decision maker. Based on our results, it is not always viable to implement control strategy at the most important risk factor because of the consideration of mitigation cost, relative loss and probabilistic interdependency between connected risk factors

Enclosing rail capacity/punctuality in accessibility measures: a two-step approach by using Data Envelopment Analysis and Analytic Hierarchy Process

“…Nevertheless, for urban railway systems, which are more independent and closed systems, the B method has been accepted and has been applied in such critical systems and some key components. The early success stories are [12]: KVB, an Automatic Train Protection system for SNCF since 1993; SAET-METEOR [13], a driverless Train Automation and Operation system in metro Line 14 in Paris in 1998; Roissy VAL, a Section Automatic Pilot system for light driverless shuttle for Paris-Roissy airport in 2006. A recent application is the COPPILOT system [14], [15], which is a metro platform screen door controlling system of the ClearSy company.…”

“…Accessibility is a complex concept with various facets; it has been widely treated in the scientific literature of the last years [11][12][13][14][15][16][17][18][19][20][21] and also in several European and international research projects or studies [22][23][24]. Inter alia accessibility could be defined as 'the amount of effort for a person to reach a destination' or 'the number of activities which can be reached from a certain location'; indicators of accessibility measure the benefits households and firms in an area enjoy from the existence and use of the transport infrastructure relevant for their area.…”

“…Among this works, we proposed a mixed-integer linear programming (MILP) formulation for solving to optimality the problem of routing and scheduling trains in case of railway traffic perturbation [7], [10], [11], [12], [13], [14]. Although very often this MILP formulation quickly finds the optimal solution to realistic instances, it fails sometime in delivering it within a computation time in line with real-time purposes.…”

“…We compared the performance of the three methods against the simple reactive rule based actions typically employed in real manufacturing settings. The statistical analysis of the results evidenced the outperformance of the Iterated Greedy algorithm (IG) of [13] over the rest of the methods. The results depend neither on the type of disruption affecting the system nor on the relative importance of makespan or instability in the objective function.…”

“…After the publication of [3], several authors proposed various heuristics to solve this problem ([4]- [13]), but the blocking constraint has been considered in none of them. The blocking flow shop scheduling problem allows many productive systems to be modeled when there are no buffers between consecutive machines.…”