Özet:
High population and limited amount of resources increase competition, and in order to survive in this competition we must use these resources wisely. Today’s manufacturers are conscious of this fact and they find efficient production methods which minimize the costs and reduce the scraps as well as provide the quality level that satisfy customer needs. In production plants, machines are not one-hundred percent reliable and they may sometimes produce defective parts. Rework, i.e. the transformation of products that do not meet the desired specifications into products that do, is one of the efficient methods that reduces the amount of scraps. In the production systems with rework loop, after the bad parts are detected and repaired by the rework line, they are sent back to the main transfer line. In order to reduce the number of defective parts, some inspection stations are located in the production systems. When a defective part is detected at inspection stations, the machine producing the bad parts is stopped so that it will not produce more defective parts. This is called Quality Information Feedback. In this thesis, we present Markov models for the approximate solution of the production systems with both single rework line and multiple rework lines. We use overlapping decomposition approach that offered in Li (2004) to approximate the throughput rate of the production systems with rework loops. The idea of overlapping decomposition technique is to decompose the system into serial transfer lines. Our model is different from Li in that, we use decomposition technique instead of aggregation procedure while evaluating the serial transfer lines and we formulate the rework rate with respect to yields of the machines that have quality failures. The accuracy of the method is validated by simulation experiments. In this work, we also seek the answer to a fundamental question: Although the bad parts are repaired in a separate line (rework line), should we stop the machines producing bad parts? In other words, we analyze the effect of quality information feedback on the rework production systems in this thesis by using the models that are developed by Kim and Gershwin, (2005).
