Master's Theses

Permanent URI for this communityhttps://hdl.handle.net/10679/9875

Browse

Browsing by Author "Akdeniz, Hekimhan"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Placeholder
    Master ThesisPublicationMetadata only
    Mathematical modeling and math heuristic approach for shift selection, lot sizing and worker assignment problem of Vestel
    Akdeniz, Hekimhan; Yanıkoğlu, İhsan; Yanıkoğlu, İhsan; Önal, Mehmet; Yılmaz, Görkem; Atan, T.; Çavdaroğlu, B.; Department of Industrial Engineering; Akdeniz, Hekimhan
    The simultaneous lot sizing and scheduling studies do not implement a worker assignment problem in the literature. This study provides a solution for worker assignment problems with overtime constraints of the Government. The study also provides fairness for overtime decision in a model which try to manage the unfairness between workers. The shift and overtime types are decided initially. Then, the production lot sizes are de ned and the demanded models are scheduled according to the due dates of the orders. This study examines the simultaneous lot sizing and scheduling problem of the serial production which has specific constraints that affect the lot size of the production set and schedule of the orders. The study presents two different models which are called combined Model and two-phase model. The comparison parameters between combined model and two-phase model are in terms of solution time, real cost difference, and achieving the purpose. The heuristic methods are combined with mixed-integer programming to solve the worker assignment problem. A fairness restriction has been proposed in the last section of the study. The models try to solve the problems in a reasonable time. The developed models are solved by Gurobi Optimizer and Python. The results show that the two-phase model provides very close output with a 0.27% difference in comparison to the combined model in a reasonable time. The results show also that the two-phase model can solve the problem faster time than the combined model in most of the real instances.