IFAC PapersOnLine 52-13 (2019) 1022–1027 ScienceDirect Available online at www.sciencedirect.com 2405-8963 © 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Peer review under responsibility of International Federation of Automatic Control. 10.1016/j.ifacol.2019.11.329 © 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. 1. INTRODUCTION Mathematical models and experimental procedures are employed to estimate the value of the performance metrics of manufacturing production systems, which are charac- terized by their complex, dynamic and stochastic natures. Even when it is possible through direct observation to estimate the value of the performance metrics, and to verify how improving measures impact the system, the use of mathematical models of the production system, for predicting its behaviour, constitutes a good analysis approach. Also, taking into account that decision sup- port tools based on digital models and simulation are increasingly used to improve manufacturing competitive- ness, mathematical models of systems performance facil- itate a straightforward computer implementation of con- tinuous monitoring and improvement of the performance of production systems. Control (and therefore the measurement of performance) of manufacturing systems is a daunting task, it needs information applications as a part of the modern Man- ufacturing Execution Systems (MES) / Manufacturing Universidad Nacional de Colombia sede Bogot´a. Universit´ e de Technologie de Troyes Operations Management (MOM) as those belonging to Industry 4.0. Through MES, the process control systems is supervised, decisions about the routes that products follow through the production system and starting point of operations on products are made, and disruptions of the manufacturing operations such as machine failures are handled (Valckenaers and Van Brussel, 2005). MES are used for acquiring accurate and reliable data from ma- chines and their components and for inferring meaningful information from this data, as the system performance. Overall Equipment Effectiveness (OEE) is an opera- tional measure tool used for evaluating the manufacturing systems performance that can be monitored using MES in both, batch and continuous production. OEE is also an indicator of process improvement activities for manufac- turing systems (Dal et al., 2000). OEE is also applied to identify different types of produc- tion losses and provides directions about improvement ar- eas of processes related to manufacturing production sys- tems performance metrics. OEE is considered to combine the operation, maintenance and management of manufac- turing equipment and resources through determining oper- ational losses which are mainly functions of the availability, performance rate and quality rate of the machine, produc- Keywords: Multi-state system, universal generating function, overall equipment effectiveness, cost-efficient corrective maintenance Abstract: The monitoring and control of production system performance is a continuing con- cern within prioritization and optimization decisions regarding manufacturing systems. There is interest on mathematical models of systems performance that facilitate a straightforward computer implementation of continuous monitoring and improvement of the performance of production systems. In this work, it is considered a multi-state system approach for modeling production manufacturing systems and for measuring its long-term performance. One opera- tional measure used for evaluating the manufacturing systems performance is overall equipment effectiveness. Here, the focus on overall equipment effectiveness is the total stoppage time due to corrective maintenance. This study therefore set out to assess a mathematical modeling approach for evaluating decisions about cost-efficient corrective maintenance. This approach is used for computing the manufacturing system states probability distribution, since calculating the total production cost per unit (total unit cost of production plus total unit cost of maintenance). In order to evaluate the goodness of the mathematical approach, a continuous time Markov chain model and a discrete event simulation model are used, by comparing the values obtained using the three techniques. The results demonstrate the effectiveness of the universal generating function for assessing system performance metrics. Copyright c 2019 IFAC * MindLab Research Group. Systems and Industrial Engineering Department.Universidad Nacional de Colombia, Bogot´a, Colombia (e-mail: gabula@ unal.edu.co). ** Laboratoire de Mod´ elisation et Sˆ uret´ e des Syst` emes (LM2S-STMR). Universit´ e de Technologie de Troyes, Troyes, 10000 France (e-mail: nacef.tazi@utt.fr ; eric.chatelet@utt.fr) BULA, Gustavo * TAZI, Nacef ** CHATELET, Eric ** Determining Production Systems Performance Metrics Considering Machine Downtime