Submitted by Group 9 Section F Phanindra Kumar J Maitreyee Korgaonkar S M Prasanna Kumar Neha Hajela Nitin Nagori Subhanan Sahoo Vijay Khuman C
Q1. For Indianapolis, the maximum number of expected calls in the peak hour (10AM to 10:59 AM) is
108 calls. Now let us assume we need to serve all the customers (including peak hours) hence for a stable system λ≤s
(Arrival rate ≤ No. of Servers*Mean service rate)
As we know that each call takes 5 minutes average service time, hence µ=12 calls/hr. Thus, s ≥ 9 108/12
(For stable system)
Also there are 2 conditions that the average “hold” ( Waiting) time is only 1 minute.
Now let us take s=10, 11, 12 & so on till this condition is satisfied. The values for different variables for the iterations are as follows No. of servers 10 11 12
Po( Idle time Prob.) 0.00006959687 0.00009955626 0.00011279897
Lq (No. of calls waiting) 6.018584 1.937116 0.798104
Wq (seconds) 200.6194572 64.57053701 26.60346845
Hence the minimum number of servers for Indianapolis BTC is 12, for maintaining the desired service levels at each hour.
No. of servers s = 12
Hence the management has to hire 12 Travel counsellors.
Q2 a) If the average service time is halved to 2.5 minutes
Now we have µ=24calls/hr. Hence for a stable system
λ≤s
Hence, Now
s ≥ 4.5 108 24
Thus we now calculate the various values for s=5 with the condition of 1 minute ( 60 seconds) waiting time
No. of servers Po( Idle time Prob.) Lq (No. of calls waiting) 5 0.00495854962 6.862439 6 0.00914011511 1.264956 We can find that 6 servers are now sufficient for handling the same BTC
Wq (seconds) 228.7479662 42.16520758
No. of servers s = 6
Q2 b) If the on-hold time is increased to 120 seconds (2 minutes)
Now we assume µ=12 calls/hr. & same arrival rate, but only the waiting time can be increased to 2 minutes or 120 seconds. Everything in Q1 will remain same, but now we can take the minimum number of servers as 11, rather than 12. No. of servers 10 11
Hence
Po( Idle time Prob.) 0.00006959687 0.00009955626
Lq (No. of calls waiting) 6.018584 1.937116
Wq (seconds) 200.6194572 64.57053701
No. of servers s = 11
Q3) Economic impact when the 3 centres (Indianapolis, Kansas & Dallas) have a centralized BTC
Carrying out a similar analysis on Kansas City & Dallas we find that the total number of servers required for each of them is: BTC
No. of servers
Indianapolis Kansas City Dallas
12 7 8 27
Total
Thus, total no. of servers required is 27 When we centralize then all the traffic is centralized; hence the maximum arrival rate is now 201 calls in the peak hour. Again for a stable system, the number of servers should be more than s ≥ 16.75 201/12
Hence once again we try to find the appropriate number of servers for which the waiting time is less than 60 seconds No. of servers 17 18 19 20
Po( Idle time Prob.) 0.00000000756 0.00000002834 0.00000003957 0.00000004572
Lq (No. of calls waiting) 62.30933 9.199211 3.689692 1.801165
Hence the number of servers required is 20 for the centralized BTC. Hence in a centralized system 7 less travel counsellors are required. Thus the economic impact of having a centralized BTC is Savings = 7*30,000 =$ 210,000 (As we are saving the salaries of 7 travel counsellors, which is $30,000 per annum) Savin s= 210 000
But there are certain set-up & maintenance costs which will be incurred for this centralized BTC.
Ex enses= 160 000 Hence the overall benefit for year 1992, will be Benefit= 50 000
Thank you for interesting in our services. We are a non-profit group that run this website to share documents. We need your help to maintenance this website.