The Integration of Maintenance Decisions and Flow Shop Scheduling

In the conventional production and service scheduling problems, it is assumed that the machines can continuously process the jobs and the information is complete and certain. However, in practice the machines must stop for preventive or corrective maintenance, and the information available to the planners can be both incomplete and uncertain. In this dissertation, the integration of maintenance decisions and production scheduling is studied in a permutation flow shop setting. Several variations of the problem are modeled as (stochastic) mixed-integer programs. In these models, some technical nuances are considered that increase the practicality of the models: having various types of maintenance, combining maintenance activities, and the impact of maintenance on the processing times of the production jobs. The solution methodologies involve studying the solution space of the problems, genetic algorithms, stochastic optimization, multi-objective optimization, and extensive computational experiments. The application of the problems and managerial implications are demonstrated through a case study in the earthmoving operations in construction projects.