Use the Extend model for the three station serial system (Three Station Serial.mox) in the Examples/Models and Methods folder. Use a mean interarrival time of 5 minutes and a mean service time of 4 minutes for each station. Measure the average cycle time, average WIP, and throughput rate.

a. Assume onstant service and interarrival times. Use simulation for 250 minutes.

b. Assume exponential service time and interarrival time distributions. Use simulation for 250 minutes. Comment on the effect of statistical variability on the system measures.

c. Assume exponential service time and interarrival time distributions. Use simulation for 10,000 minutes.

d. Assume exponential service time and interarrival time distributions. Use 10 runs of a simulation for 1,000 minutes.

e. Assume exponential service time and interarrival time distributions. Use analytical formulas to compute the system measures. Compare the analytical results with the simulation results of parts c and d.

Use a mean arrival time of 5 minutes and a mean service time of 8 minutes for each station. Compare average number in the queue, average wait in the queue and number produced in a 250 minute period.

a. Use constant service and interarrival times.

b. Use constant service time and exponential interarrival time.

c. Use exponential service time and constant interarrival time.

d. Use exponential service time and exponential interarrival time.

e. What observation can you make about the effects of statistical variability on the observed quantities?

f. Use queuing formulas to compute mean number in the queue and mean wait in the queue for exponential service time and exponential interarrival time. Compare with part d. What observation can you make about the use of analytic formulas versus simulation to estimate these quantities?

g. Do part f, but use an average of the results of 10 runs for simulation estimates.

h. Compare average number in the queue and average wait in the queue with the single channel case with a mean arrival time of 5 minutes and a mean service time of 4 minutes. We note that the efficiencies of both the parallel and single channel systems are the same. Is it better to have two slow machines are one fast machine?