This project is aimed at simulating the performance of a manufacturing system in a single day (8 hours). This system consists of 3 machines; A, B, and C.
Machine A can only process a-parts, while machine B can process a-parts and b-parts. Machine C further processes the parts received from machine A and B (prioritizing parts received from machine A).
Machine B only processes a-parts when machine A is out of order. When this happens, machine B will also prioritize processing a-parts, meaning that a-parts are processed ahead of b-parts in the queue of machine B, but after the b-part that is being processed by machine B.
When machine B breaks down, the b-parts in its queue will remain in queue (and new incoming parts will also stay in queue) and after being repaired, machine B will continue processing the b-part that it was processing before breaking down.
Each component of the system has been mapped to a specific probability distribution:
- The interarrival time of a-parts follows a normal distribution with mean 100 and standard deviation of 5 minutes.
- The interarrival time of b-parts follows a normal distribution with mean 12 and standard deviation of 5 minutes.
- The time that machine A takes to process a-parts follows a normal distribution with mean 91 and standard deviation of 9 minutes.
- The time that machine B takes to process b-parts follows a normal distribution with mean 15 and standard deviation of 3 minutes.
- The time that machine B takes to process a-parts follows a normal distribution with mean 110 and standard deviation of 10 minutes.
- The time that machine C takes to process either part follows a normal distribution with mean 25 and standard deviation of 4 minutes.
- The time between each breakdown of machine A follows a normal distribution with mean 400 and standard deviation 30 of minutes.
- The time between each breakdown of machine B follows a normal distribution with mean 200 and standard deviation of 15 minutes.
- Machine C rarely breaks down, therefore we can assume that machine C doesn't break down.
- The repair time of machine A follows a normal distribution with mean 15 and standard deviation of 14 minutes.
- The repair time of machine B follows a normal distribution with mean 10 and standard deviation of 8 minutes.
The simulation model is run 10 times to estimate the performance metrics of the system via point estimation and interval estimation. The performance metrics are: average response time (this means the average time a part spends in the system, for both types of parts), Number of parts processed in a day (each type), occupancy rates of each machine, failure rate of each machine.
Each run consists of two phases: a warmup phase, and a main phase which the system is evaluated upon. Since the system is never empty at the beginning of a day, we implement a 2 hour warmup phase in our simulation to reduce the bias of our estimations. The main phase is an 8 hour simulation of the system (continued from the warmup phase), which is used for estimating the performance of the system.
After the simulation, relevant graphs about the performance of each machine is generated to check for bottlenecks.