Using Predictive Modeling to Predict Failures in Supply Chain

Target Objective

• Implementing manufacturing traceability solutions would allow firms to improve data accuracy, reduce human error, increase customer service levels, and save money in lost revenues, recall costs, litigation and fines.
• It would also allow firms to prevent quality issues before they occur.
• Our focus in this study will be on process control, specifically on predicting failures within processes.

Project Detail

• We would like to provide a Markov Decision Process based framework for increasing process reliability and identifying replacement times for components and products before failures.
• To illustrate, we aim to implement our model on the We Energies data set related to identifying failing meters in an electrical distribution system.
• To achieve our goals, we will utilize prescriptive, predictive, and descriptive analytics.

Target Outcomes

• Increase equipment and process reliability with prescriptive maintenance;
• Minimize the total operational cost for the company given the set of constraints to be satisfied;
• Provide dynamic and optimal recommendations and develop a decision support system that can achieve “what-if” analysis.
• Build a system that allows real time remote monitoring, asset location and status.
• Publish research papers in top operations management journals and develop whitepapers.