Using a simulation of the manufacturing process Mr. Gürkan found a method to find the optimal location to place buffers to help machines run smoother and longer without breaking down. The simulation ran a maximum of 50 machines and took data on cycle times and downtime to set up the production process. Then using the data collected and forming a directional derivative could be used to infer the approximate location where the process became too loud or uneven. These points were isolated and selected for the optimal location for the buffers. Once added into the simulation the productivity of the process was found to increase as was the time between break down on the production line.
To further test the simulation additional production line simulations were produced. These additional simulations were made with fewer machines used in the process. These simulations were tested utilizing the directional derivative and more traditional stochastic testing methods and the level of error found between the two on the location of buffers was found to be negligible.
Gürkan, G. (2000). Simulation optimization of buffer allocations in production lines with unreliable machines. Annals of Operations Research, 93(1–4), 117–216. https://libproxy.ecpi.edu:2111/10.1023/a:1018900729338