Contrôle de la production et de la maintenance des systèmes industriels non fiables pour les produits périssables

Abstract

The work presented in this thesis deals with problems related to optimal control. It studies an unreliable and complex manufacturing system. On one hand, the studied production system is subject to random failure and repair times. On the other hand, the family of products manufactured is of a perishable nature with a limited and random shelf life. This project focuses on the design and the optimization of production and maintenance control policies for perishable products. It aims to develop control policies that integrate the different aspects of the system while optimizing its parameters. First taking into account the important balance that must take place between the costs of backlog on the one hand and the costs of disposal on the other hand, the objective is to conduct an appropriate production control policy that minimizes the total cost while respecting the constraint of meeting demand. In addition, the shelf life of perishable products considered in the developed models is both fixed and random. This is considered consistent because it will allow to examine several cases of industrial systems and to see the effect of shelf life variability on the behavior of the system. Moreover, manufacturing systems face several factors that can lead to machine deterioration and increased failure rates. As a result, they aim to maximize their availability and minimize the occurrence of breakdowns by performing preventive maintenance interventions. This task becomes more critical if the deterioration of the machine has an impact on the shelf life of the manufactured products. The objective is then to develop a maintenance control policy that takes into account all the interactions that may exist in the system. On the one hand, the first part of this thesis deals with determining an optimal production control policy specific to unreliable manufacturing systems producing perishable products with random shelf lives. On the other hand, preventive maintenance control policies are introduced in the second part to study how to integrate production activities and maintenance interventions in order to improve machine availability while considering the nature of perishable products. In this second part, we are interested in studying the effect of machine degradation on the product’s shelf life and how this relationship can influence the joint maintenance and production control policy. In order to solve the considered problems, we first adopt an analytical solution in order to develop a parameterized production control policy for an unreliable manufacturing system prone to random failure and repair times. The optimality conditions in the form of Hamilton - Jacobi - Bellman equations (HJB) are developed. Then, by numerically solving the HJB equations, the production policy is established. Subsequently, due to the complexity of the system under consideration and the limitations of the analytical models, we adopt an experimental approach based on simulation as a tool for optimizing decision parameters. The latter integrates the response surface methodology, experimental design and sensitivity analysis. Numerical examples of the models developed were established to assess the performance of the system and confirm the robustness of the used approach. This experimental approach is also adapted in the second part of the thesis to optimize the combined maintenance and production policy.

Date	1 Sept 2021
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Ali Gharbi (Supervisor) & Jean-Pierre Kenné (Co-supervisor)