Optimization of unpaced synchronous production lines

• Applicant: Professorin Dr. Sigrid Knust
• Subject Area: Accounting and Finance
• Term: from 2013 to 2016
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 240376541

Following the paradigm of mass customization, the production process shifts from mass to variant production. For this, in almost all industrial settings automatic production lines are used. Raw material is transferred between different stations by a conveyor belt until the finished product can be taken from the last station. At the stations the products are processed in different operations by specialized machines or employees. Depending on the products the production times may vary a lot at different stations. While in mass production no sequencing problem has to be solved, in variant production a good sequence is important to use the optimization potential of the production line.In this project we consider unpaced production lines with synchronous transportation. In these systems the conveyor belt is moved after all operations of the current workpieces are finished, i.e. all stations wait for the workpiece finished last. Hence, production times depend on all products which are simultaneously placed on the production line. Especially, if we have a large variety in the durations of the operations, for an efficient production it is important to find a sequence of the products such that waiting times are minimized. Here, we have indirect setup times occurring between non-consecutive products with a certain distance in the sequence. Additional complex sequence dependencies may occur due to changeover times for resources.Despite their high relevance in practice, these problems have been treated only rarely in the literature. Especially, there are no theoretical considerations regarding the sequencing problems depending on the configuration of the production lines. The main goals of this project are a well-founded theoretical analysis and suitable models of unpaced synchronous production lines as well as the development of efficient optimization algorithms. In the project we especially consider production lines with additional resources and setup times leading to two kinds of sequence dependencies: sequence dependent setup times when waiting for the last operation to be finished as well as changeover times for resources.All algorithms will be tested and evaluated on real-world data from practice. In discussions with companies we realized that production planning problems are often solved manually in an unsatisfactory way due to their high complexity. Therefore, besides new theoretical insights we also expect a profit for industrial applications from our project.

• DFG Programme: Research Grants