Please

An e commerce company employs two parallel processors to serve incoming re- quests to their web site. The company wants to replace the two systems by a single newer system that processes requests twice as fast as either of the old systems. They reason that the response time to a request should decrease, since with the new system work is always being performed at the same rate unless there is a single job in the system, in which case the new system would process it twice as fast. Data were collected on the current parallel system. Based on a sample of size 100, the interarrival times had sample mean ia and sample variance (all times in seconds). The service times had sample mean is and sample variance oz. Fur- thermore, 90% of the interarrival times were less that M, and 90% of the service times were less than Ms. The distributions of the interarrival times and service times are unknown; all we have to work with are the sample statistics. We can assume that the new machine processes a given job in exactly half the time taken by either of the old machines, and that jobs arrive at a constant rate. The parameters were ha = 10, = 105, js = 16,63 = 250, Ma = 23, Mg = 37. In order to estimate the average response time for the new system, a simula- tion was carried out (using the program GG1Range.py on Canvas). The interar- rival times were generated from the uniform distribution on 0, Maand the service times were generated from the uniform distribution on [O, M./2). The simulation produced an average time in system of 20.7 with a 95% confidence interval of [20.089, 21.332). The simulation was based on a time horizon of 500,000. For next week, answer the following questions. (1) Describe any weaknesses you see in the approach outlined above, and the problems that might arise because of the shortcomings. (2) How would you go about estimating the average response time, and what is your estimate? Justify any assumptions you make. Submit a text file with your answers by 11:30 am on April 20. If you work with a fellow student, put both names at the top of each of your submissions. You should not require more than a page. An e commerce company employs two parallel processors to serve incoming re- quests to their web site. The company wants to replace the two systems by a single newer system that processes requests twice as fast as either of the old systems. They reason that the response time to a request should decrease, since with the new system work is always being performed at the same rate unless there is a single job in the system, in which case the new system would process it twice as fast. Data were collected on the current parallel system. Based on a sample of size 100, the interarrival times had sample mean ia and sample variance (all times in seconds). The service times had sample mean is and sample variance oz. Fur- thermore, 90% of the interarrival times were less that M, and 90% of the service times were less than Ms. The distributions of the interarrival times and service times are unknown; all we have to work with are the sample statistics. We can assume that the new machine processes a given job in exactly half the time taken by either of the old machines, and that jobs arrive at a constant rate. The parameters were ha = 10, = 105, js = 16,63 = 250, Ma = 23, Mg = 37. In order to estimate the average response time for the new system, a simula- tion was carried out (using the program GG1Range.py on Canvas). The interar- rival times were generated from the uniform distribution on 0, Maand the service times were generated from the uniform distribution on [O, M./2). The simulation produced an average time in system of 20.7 with a 95% confidence interval of [20.089, 21.332). The simulation was based on a time horizon of 500,000. For next week, answer the following questions. (1) Describe any weaknesses you see in the approach outlined above, and the problems that might arise because of the shortcomings. (2) How would you go about estimating the average response time, and what is your estimate? Justify any assumptions you make. Submit a text file with your answers by 11:30 am on April 20. If you work with a fellow student, put both names at the top of each of your submissions. You should not require more than a page