Q#1: Let unit time be 1 minnute. Then, lambda1 = 1, lambda2 = 1.5, mu = 2. (a) 1 Ws1 = ------------ = 1 (minute). mu - lambda1 (b) lambda2 rho2 = ------- = 0.75, mu rho2 Wq2 = ------------ = 1.5 (minutes). mu - lambda2 Q#2: Let unit time be 1 hour. Then, lambda = 60, mu = 30. The cost factors are Cb = $18 per booth per hour, and Cd = $15 per booth per hour. Let the # of booths in operation be n. (a) No, because if n = 2, then lambda rho = ------- = 1, n*mu which means the average length of queue will be infinity. (b) lambda rho = -------. n*mu Given "n" and "rho", find "Wqmu" by using the "Wqmu - n" chart. Then, Wqmu Wq = -----. mu The total system cost is TC(n) = n*Cb + Ws*lambda*Cd = n*Cb + (Wq + 1/mu)*lambda*Cd = 18*n + 30*Wqmu + 30 --------------------------------- n rho Wqmu TC(n) 3 2/3 0.4 96 4 2/4 0.08 104.4 (c) In terms of total system cost, the optimal number of booths in operation should be n=3. We assume that the total system cost is a convex function of "n", and actually, it really is. Thus, from (b), we know TC(n) is increasing in "n" when "n" is greater than 3. Hence, the optimal number is less than or equal to 3. When "n" is less than 3, the average length of queue is infinity, so is TC(n). Therefore, the optimal number is exactly 3.