2. We consider a coffee stand with 2 servers. The servers work for 8 hours per day. The customers arrive with EXPO (3) minutes. The service takes TRIA (3,4,7) minutes, and customers are served in a FIFO basis. We are interested in the average time a customer spends in the system. a. We run some replications for this system. Arena results show the following. (All values are in minutes.) Given this information, build a 95\% confidence interval for the true mean of average time spent in the system. b. Consider the output statistics in part b above. What are the values of accuracy level and precision level for the estimation made for the true average time spent in the system? c. Same system is tested with a different number of replications. Output Analyzer shows the following: According to this output, answer the following True/False question. You have to explain your reasoning. - 5% of the customers in this system will either wait longer than 8.71 minutes or shorter than 8.27 minutes. d. We want to test the performance of alternative designs for this coffee stand. Scenario 1: Existing system that serves the customers in FIFO basis. Scenario 2: Separate lines for servers where the customers are directed to the shortest queue upon arrival. 30 replications are run for each scenario, and the average time a customer spends in the system is compared for each pair of scenarios using Output Analyzer. The interval is computed for ii+1, where i represents the output of scenario i. According to this result which scenario performs better? What would you propose to the coffee stand management? Explain your reasoning