Year of Publication


Document Type





Manufacturing Systems Engineering

First Advisor

Lawrence E. Holloway

Second Advisor

Jon C.Yingling


This study deals with a very common problem encountered in many automotive industries. Automotive companies try to level the production of different models over time based on the demands for these models in the market. In order to achieve this, they introduce a leveled stream of cars in the beginning of the production line. But because of many reasons this leveled stream gets disturbed in its course. In order to re-level the stream, buffers are used between the shops. One such buffer is called as selectivity bank and it sits between paint shop and assembly shop. This buffer receives a disturbed sequence from the paint shop. The thesis tries to develop different algorithms that can be used to discharge cars from this buffer in order to achieve better leveling in the presence of rework and assembly constraints. These algorithms continuously try to steer the system from an undesirable state to a more desirable state by keeping track of current conditions in the plant. A simulation model is developed, which gives a platform for comparing relative performance of these logics under different conditions. The simulation tool is also helpful in designing optimum size of this buffer that will result in desired leveling performance.