Product scheduling problem with learning effect in a flexible flow shop: an application in defense industry

If the works to be done in the workshop follow the same machines in the same order, this is called the flow type production line. In recent years, work on flow type scheduling problems has been increasing. However, when the literature is examined, it is seen that there is a lack of studies which can be applicable to the real life problems. In this thesis, the scheduling problem which aims to find the minimum completion time is solved by examining the flexible workflow type assembly line belonging to a defense company which produces electronic components. Thus, it is aimed to contribute to the application of theoretical knowledge in practice. In this problem, the chief technician responsible for the production line has to manually assign jobs to the stations. However, while some stations on the line are heavily cluttered, waiting times are quite high. The goal is to automatically generate a production schedule that minimizes makespan, taking into account product priority relationships within the given weekly production schedule. In this real life problem, the durations of operation are considered deterministic. An important distinction that separates the problem from classical assembly line problems is that products can be produced more than one in the some stations according to their size, which causes that the products must be produced in subgroups. A mixed integer mathematical model has been developed which allows the production of subgroups of products with mixing and without mixing. The problem is NP-Hard, and the mathematical model gives optimal results for small-sized problems, while large-scale problems are solved by adapted NEH Algorithm which shows high performance in terms of time and produces near-optimal results. The algorithm is coded in Excel VBA program. Additionally, as there is a significant involvement of humans in the scheduling environment, the number of activities subject to learning is high. Therefore, learning effects is examined in the production line. For the cases where subgroups of products are produced by mixing and without mixing, results were obtained and analyzed. After analyzing the learning effect on the product scheduling, the decision support system for the company was developed. Through the decision support system, the weekly production schedule creation process has been automated, which has resulted in an improvement of approximately 21% in the weekly makespan of the products.