UCL School of Management

Research seminar

Professor Alp Akcay, Eindhoven University of Technology


Wednesday, 28 April 2021
15:00 – 16:15
Research Group
Operations and Technology

UCL School of Management is delighted to welcome, Professor Alp Akcay, Eindhoven University of Technology to host a research seminar discussing ‘Data-Driven Inspection Planning for Production Processes with Hidden Defects’.

Abstract: In this talk, I will present two research papers based on a collaboration with Philips’ shaver factory at Drachten, the Netherlands. We consider high-precision machine tools in a discrete manufacturing setting. Before a tool fails, it goes through a defective phase where it can continue processing new products. However, the products processed by a defective tool do not necessarily generate the same reward obtained from the ones processed by a healthy tool. The defective phase of the tool is not visible and can only be detected by an inspection with the objective of avoiding the costly failure. However, the inspections are also costly and when to perform an inspection is an important decision.

The first part of the talk is about a problem where the tool can be retired from production to avoid a tool failure and save its salvage value, while doing so too early causes not fully using the production potential of the tool. We build a Markov decision process model and study when it is the right moment to inspect or retire a tool with the objective of maximizing the total expected reward obtained from an individual tool. The structure of the optimal policy is characterized. The implementation of our model by using the real-world maintenance logs at Philips shows that the value of the optimal policy can be substantial compared to the policy currently used in practice.

The second part focuses on the inspection optimization problem when there are imperfect signals coming from a data-driven prediction model (i.e., a binary classifier with a known false-positive and false-negative probability). A new hybrid inspection policy, which generalizes the classical age-based inspection scheduling, is introduced. To optimize the proposed inspection policy, a stochastic dynamic programming model is formulated with the objective of minimizing the long-run expected cost rate. The performance improvement achieved by the proposed policy is quantified by comparing it to practically relevant benchmark policies. Numerical experiments with a set of realistic problem instances show that adding alert-triggered inspections to age-based inspection scheduling brings substantial reduction in the expected cost rate when the predictive model is sufficiently accurate.

I will conclude my talk with some reflection on future research challenges in smart industry.

Open to
Last updated Monday, 24 May 2021