A Markov Decision Process model for maximizing overall equipment effectiveness

Purpose: The purpose of this paper is to present a novel high-level decision-making framework for optimizing Overall Equipment Effectiveness (OEE) in production systems. The model can be used to schedule the daily production and maintenance tasks to maximize the OEE—a topic that is seldom addressed in the existing literature within dynamic settings. Design/methodology/approach: This study introduces a novel Markov Decision Process (MDP) model to optimize OEE in production systems. The model uses daily availability and quality rate as the system state vector to streamline decision-making regarding maintenance and production scheduling. The MDP model was solved by the policy iteration algorithm, with the objective of maximizing the OEE. The model is demonstrated through an industrial case study demonstrating the practical effectiveness of the MDP framework in improving OEE. Findings: The study finds that using an MDP-based approach significantly enhances OEE. Through a case study, we show that this approach provides better outcomes compared to traditional methods. The model effectively balances availability, performance, and quality components, leading to improved production efficiency and reduced downtime. Research limitations/implications: This research primarily focuses on using daily availability and quality as the system state, which, while effective for high-level decision-making—such as allocating time for maintenance rather than specifying exact maintenance tasks—may not capture the detailed states of production systems. Additionally, the model’s reliance on real-time data may require high levels of data accuracy and consistency, which could be a limitation in systems with less comprehensive monitoring capabilities. Future studies could expand the model to include additional system states and quality metrics. Practical implications: The proposed MDP model provides a practical decision-support tool for managers in production environments. By optimizing maintenance schedules and the production rate, the model helps reduce downtime, improve equipment utilization, and enhance product quality. This model can be useful for industries looking for the optimal control of OEE. Social implications: Improving OEE has wide social implications, particularly in industries where enhanced operational efficiency leads to reduced resource consumption and waste. By optimizing the utilization of machinery and minimizing downtime, the MDP model contributes to more sustainable production practices. Originality/value: This paper offers a unique contribution to the field by modeling OEE using the MDP, with daily availability as the system state. Unlike most studies in the literature that focus on measuring OEE, this research aims to facilitate optimal control of OEE, and with the increasing availability of sensory data in modern industries, this direction holds significant potential for practical application.