Supply network resilience learning: An exploratory data analytics study

Abstract

When a supplier experiences a disruption, it learns how to better prevent and recover from future disruptions. As suppliers learn to become more resilient, the overall supply network also learns to become more resilient. This research draws on the organizational learning literature to introduce the concept of supply network resilience learning, which we define as the improvement of supply network resilience when suppliers learn from their own disruptions. The analysis integrates agent-based modeling, experimental design, data analytics, and analytical modeling to investigate how supplier learning improves supply network learning. We examine how two types of supplier learning, namely, learning-to-prevent and learning-to-recover, affect supply network learning. The results show that suppliers' learning-to-prevent results in a disruption-free supply network when time approaches infinity. However, the results differ across a more realistic finite time horizon. In this setting, learning-to-recover improves network learning when suppliers face a lower chance of disruption. The analysis also shows that centrally located suppliers enhance network learning, except when the risk of a disruption is high and the chance of diffusing a disruption to another supplier is high. In this setting, noncentral suppliers become more critical to supply network learning. This research provides a framework that will help practitioners understand the contingencies that influence the effect of supplier learning on the overall supply network resilience learning.