(477f) A Model-Based Risk Management Strategy for Distributed Supply Chains

Supply chain management is a key consideration for today's enterprises to increase their operational efficiencies and minimize costs while maintaining quality. Globalization due to the lowered trade barriers has increased competitiveness and put higher pressure on companies' bottom lines. To reduce costs and increase efficiency, strategies such as outsourcing, single sourcing, multi-site production, centralized distribution, etc are employed by companies. This leads to lengthy and complex supply chains, which suffer from the lack of visibility among supply chain entities and vulnerability to disruptions. The lack of visibility hinders supply chain entities from making reliable forecasts, which results in inefficient supply chain operations and increases risks. Risk is the probability of a disruption occurrence in the supply chain operations. The material, information, and finance flows among entities of a complex supply chain are disruption-prone as blockage in any of these would lead to undesirable events like process shutdown, financial loss, under- or over-supply, etc. Hence, supply chain risk and disruption management is crucial.

The literature in the area of supply chain risk and disruption management is limited and no general structured methodology has been proposed to date. Sheffi et al. (2003) describe mechanisms that companies follow to assess terrorism-related risks, protect the supply chain from those risks, and attain resilience. They provide classifications of disruptions, security measures, and ideas to achieve resilience. They report various case studies and interviews with executives of companies. Wilson (2003) focuses only on transportation disruptions. Gaonkar and Viswanadham (2004) propose a conceptual framework to approach supply chain risk problems. They classify supply chain risks in three forms ? deviation, disruption and disaster ? and noted that supply chains need to be robust in three levels ? strategic, tactical and operational. Julka et al. (2002a, b) propose an agent-based framework for modeling a supply chain. In their framework, the behavior of every entity in the supply chain is emulated using an agent that imitates the behaviors of various departments in a refinery. Mishra et al. (2003, 2004) extend this approach to manage disruptions in a refinery supply chain. In the event of a disruption, agents collaborate to identify a holistic rectification strategy using heuristic rules. Adhitya et al. (2005) propose a heuristic-based rescheduling strategy to manage refinery supply chain disruptions. Heuristic approaches lack generality and flexibility; this paper, therefore, proposes a general model-based framework for risk and disruption management in a supply chain.

The key element of the proposed framework is a causal model of the supply chain operations. The model consists of modules which represent the supply chain entities. Each module consists of nodes which represent the supply chain variables related to that particular entity. Nodes are interconnected by arcs, which represent the cause-and-effect relationships between the variables they connect. Disruptions to the supply chain can be imposed on the causal model. They may lead to undesirable consequences such as infeasible supply chain operations, insufficient inventory, production shortfall, unfulfilled demand, etc., which are modeled as violations of limits in the causal model. These violations are identified by using the causal model to propagate the effects of disruptions, by tracing the cause-and-effect relationships starting from the disrupted nodes to other nodes as dictated by the connections (arcs).

Once the violations are identified, rectification options are searched using the rectifications-graph, which captures all possible options to overcome the disruption effects, based on a utility function (for example: minimum costs, minimum changes in the operations schedule, earliest rectifications, etc). The rectifications-graph is generated based on the causal relationships captured in the causal model. In contrast to heuristic approaches, the main advantages of the proposed model-based method are generality, completeness of solution search and flexibility of the utility function. Capabilities of the proposed framework are illustrated using an industrial supply chain example.

References: 1) Adhitya, A., Srinivasan, R. and Karimi, I. A. (2005). Managing Abnormal Events in Refinery Supply Chains by Heuristic Rescheduling. Submitted to AIChE J. 2) Gaonkar, R. and Viswanadham, N. (2004). A conceptual and analytical framework for the management of risk in supply chains. Proceedings of the 2004 IEEE International Conference on Robotics and Automation, 2699-2704. 3) Julka, N., Srinivasan, R. and Karimi, I. A. (2002a). Agent-based Supply Chain Management-1: Framework, Comput. Chem. Eng. 26(12), 1755-1769. 4) Julka, N., Srinivasan, R. and Karimi, I. A. (2002b). Agent-based Supply Chain Management-2: A Refinery Application, Comput. Chem. Eng. 26(12), 1771-1781. 5) Mishra, M., Srinivasan, R. and Karimi, I. A. (2003). Managing disruptions in refinery supply chain using an agent-based decision support system. Presented in the AIChE annual meeting, San Francisco, Nov 16-21, 2003. 6) Mishra, M., Srinivasan, R. and Karimi, I.A. (2004). A model-based framework for detecting, diagnosing, and rectifying supply chain disruptions. Presented in the AIChE annual meeting, Austin, TX, Nov 7-12, 2004. 7) Sheffi, Y., Rice, Jr., J. B., Fleck, J. M. and Caniato, F. (2003). Supply Chain Response to Global Terrorism: A Situation Scan. Eur OMA-POMS Conference, 2003. 8) Wilson, M. C. (2002). Transportation Disruptions in the Supply Chain: Simulator as a Decision Support Tool. Proceedings of the 31st Annual Logistics Educators Conference.