Production scheduling using discrete event simulation

Abstract

Production scheduling and job sequencing can be a complex problem especially with multiple machines and a large variety of products. The several different possible combinations of the production plan possess a lot of potential for improvement through better load balancing and minimizing the number of setups. Main objective for this thesis was to develop a simulation model that can be used in production planning and to increase the visibility and control on the production planning process in the selected company. This thesis started looking for better ways to schedule the production list through prioritization rules and a genetic algorithm. Discrete event simulation model was built to compare the different rules and the genetic algorithm. The different simulation experiments compared the throughput of the system and the balance of loads within manufacturing operations. The results showed clear improvement with the tested methods. At best the throughput was improved by 15% while also balancing the load between stations more evenly. The simulation model is able to produce a job sequence for each machine which increases visibility and control in processes.