Integrated decision-making framework for preventive maintenance scheduling and spare part logistic planning
MetadataShow full item record
Maintenance scheduling for geographically distributed assets intricately and closely depends on the locations and availability of spare parts, which motivates the joint decision-making on the maintenance scheduling and spare part logistics, including optimization of the system operations, as well as the design of the underlying spare part logistic network. These close interactions between the maintenance and spare part logistic activities have been ignored or oversimplified in the existing research and practice, leading to the inappropriate maintenance resource allocations and excessive maintenance waiting times. Unfortunately, such kind of joint decision-making problems are challenging due to the exceptionally large size of the decision space, as well as the strong inter-dependencies in the system operations, especially for large-scale systems with complex maintenance/logistic structures. Challenges become even more pronounced if one acknowledges that those system operations and degradation processes of the assets are greatly influenced by numerous uncertain factors, yielding highly stochastic system behaviors. To address the aforementioned problems and challenges, in this doctoral dissertation, an integrated decision-making framework is proposed to effectively schedule preventive maintenance (PM) for geographically distributed assets and properly manage inventories in distributed logistic facilities storing the necessary spare parts. In addition, several factors are discussed within the proposed decision-making framework, including the inventory-sharing structure, imperfect maintenance, transportation options and spare parts logistic network design. To capture the stochastic nature of the system operations and the trade-offs between decisions in the domains of maintenance scheduling, spare part inventory management, transportation selection and logistic network construction, a discrete-event simulation-based optimization paradigm was used to minimize generic and customizable cost functions, that reward functioning of the assets, while penalizing asset downtime and consumptions of maintenance/logistic resources. The benefits of the newly proposed integrated decision-making framework are illustrated in simulations, through comparisons between the integrated policies with several traditional, fragmented decision-making processes. Moreover, a design of experiment (DOE) based sensitivity analysis is introduced to evaluate the effects of a variety of relevant systems parameters on the resulting system operations. Future work should be aimed at incorporating robustness to uncertainties in model structures and system parameters into the newly proposed decision-making and system design methodologies, as well as implementing these methods in a real-life system settings, rather than simulations alone.