An integrated supply chain risk mitigation tool – model, analysis and insights

This study proposes a two-phase approach to create insights into disruptions, allowing managers to mitigate supply chain risks using an integrated risk-mitigation tool. In the first phase, we formulate the problem as a Markov Decision Process (MDP) that maximizes the expected long-run revenue induced by supply chain risks, given the current state of the risk score. We map four decision states to four specific managerial actions: ‘do nothing,’ ‘track the supplier,’ ‘monitor the supplier,’ and ‘change the supplier.’ If the optimal policy is ‘do nothing’, then the supply chain risk scores are continued to be monitored. For ‘track the supplier’, the firm will track the supplier’s performance internally. For ‘monitor the supplier’, the firm will hire an external contractor to monitor the external supply chain risks. If the optimal policy from phase 1 is to ‘change the supplier,’ we identify the best-performing supplier using an integrated best-worst goal programming (BWGP) method. We demonstrate the integrated method on a global automation technologies company. An MDP-based risk mitigation tool yields a promising approach. Results based on metrics show that the BWGP method can be used in the integrated two-phase approach with MDP as a supply chain risk mitigation tool.

本研究提出两阶段方法以剖析供应链中断问题，助力管理者借助集成化风险缓解工具缓解供应链风险。第一阶段中，我们将该问题建模为马尔可夫决策过程（Markov Decision Process, MDP），在给定当前风险评分状态的前提下，最大化供应链风险场景下的预期长期收益。我们将四类决策状态映射至四项具体管理举措：“不采取任何行动”“跟踪供应商”“监督供应商”以及“更换供应商”。若最优策略为“不采取任何行动”，则持续监控供应链风险评分；若选择“跟踪供应商”，企业将在内部跟踪供应商的运营表现；若选择“监督供应商”，企业将聘请外部承包商对外部供应链风险开展监控。若第一阶段得出的最优策略为“更换供应商”，我们将采用集成化最佳-最差目标规划（best-worst goal programming, BWGP）方法识别表现最优的供应商。我们在一家全球自动化技术企业中验证了该集成方法的有效性。基于马尔可夫决策过程的风险缓解工具展现出极具前景的应用路径，基于各项评估指标的实验结果表明，最佳-最差目标规划方法可与马尔可夫决策过程结合，应用于两阶段集成化供应链风险缓解工具中。