Professional Documents
Culture Documents
Case Summary
Pacific National Bank is considered a medium- sized bank with its twentyone branches. The bank has always relied on an outside vendor to operate its
own automatic teller machines (ATMs). Since ninety percent of the ATM
customers prefer other banks to perform their transaction, Operations Vice
President Maria Perez wishes to improve the bank by offering a broader mix of
banking services with its own machines tied into its own data- processing
network. Moreover, by operating its own ATM machines, costs are expected to
lessen because fewer human tellers would be required for the same volume of
transactions as before.
Ms. Perez selected the Greenhills office as the test branch for a captive
ATM. Customers from that branch were enlisted to sign up for a Pacific ATM
card. Once a number of ATM card holders was established, statistics were kept
by recording the average number of arrivals of customers per period or schedule
of the bank. They are as follows:
Previous data show that customers arrive randomly at a stable mean rate, so the
assumption of a Poisson process is valid. Moreover, the mean time required to
complete customer transactions is two minutes, and the individual service times have a
frequency distribution with a pronounced positive skew, so an exponential distribution
may be used to approximate data.
Once Pacific National Bank start operating its own ATM machines, it is expected
that there will be a 30% diversion of regular business away from human tellers to the
ATM, which produced a further 2% expansion beyond the previous level of overall client
transactions. In addition to this, it is also expected that the mean service time at an ATM
is one-half minute, with an exponential distribution serving as an adequate
approximation to the unknown positively skewed unimodal distribution that actually
applies. Ms. Perez believes that, once the ATM is installed, the Greenhills human tellers
will be left with a greater proportion of the more involved and lengthy transactions,
raising their mean service time to 2.5 minutes.
II. Body
1. Assume that Pacific National Bank remains with human tellers only.
a. For each time period given in the table, determine the minimum number of tellers
needed on station to service the customer stream.
In order to know the minimum number of tellers needed on station
to service the customer stream, we first need to know how often do
customers arrive and the time allotted in serving these customers. These
can be determined by solving for the mean arrival rate and mean service
time. After solving for these, the mean arrival rate will then be divided by
the mean service time to get the minimum number of tellers needed. The
solution is as follows:
Period (1)
= no. of calling units/ unit of time
= 155 customers / 2 hours
= 77.5 customers / hour
=/
= 77.5 / 30 = 2.583333 3 tellers
Period (2)
= no. of calling units/ unit of time
= 242 customers / 1 hour
= 242 customers / hour
=/
= 242 / 30 = 8.066666667 9 tellers
Period (3)
= no. of calling units/ unit of time
= 290 customers / 2 hours
= 145 customers / hour
=/
= 145 / 30 = 4.833333 5 tellers
Period (4)
= no. of calling units/ unit of time
= 554 customers / 3 hours
= 184.666667 customers / hour
=/
= 184.666667 / 30 = 6.1555556 7
tellers
To summarize, for periods one, two, three and four, three, nine, five and
seven tellers are needed, respectively.
b. Assume that the number of tellers found in part (a) is used. For each time period,
determine the mean customer waiting time.
In order to determine the mean customer waiting time, it is essential
to know the type of model the situation has in order to know the equations
to be used in solving for the missing data. In this particular case, the
problem follows a multi-server model with poisson arrivals and exponential
service times since, as the problem states, it is exponentially distributed
and may opt to follow a poisson process. Also, its servers are greater than
one and all have the same rate of service.
Once the type of model is determined, solution may begin. First, the
probability of having zero calling units in the system must be solved. Then,
the expected number of calling units must then be found. After solving for
these, the data determined will be used to get the mean customer waiting
time. The solution is as follows:
Period 1:
Probability of having 0 calling units:
Period 2:
Period 2:
Period 3:
Period 4:
c.
Period 2:
Period 3:
Period 4:
2.
Past
= 174.77 pesos
Period 3
AHQSC = 11.0134277 X 0.50 X 145
= 798.47 pesos
Period 4
AHQSC = 1.58159678 X 0.50 X 184.666667
= 146.03 pesos
b. One teller more than was found in part (a) of Question 1
Period 1
AHQSC = 0.49205654 X 0.50 X 77.5
= 19.07 pesos
Period 2
AHQSC
Period (3)
AHQSC
Period (4)
AHQSC
3. Suppose that the ATM is installed and that customers themselves decide whether to use
human tellers or to use the ATM, and that two queues form independently for each.
Finally, assume that a 10% traffic increase is generated by the ATM within each open
time period and that all of it is for the ATM.
a. Determine for each period the mean arrival rate at the human teller windows.
In order to solve for the mean arrival rate, we simply divide
the no. of calling units by its unit of time. As for the number of
calling units, we need to take note of the thirty percent diversion
from the human tellers towards the ATMs, hence:
Period 1
No. of calling units
Period 2
No. of calling units
Period 3
No. of calling units
Period 4
No. of calling units
b. Do the same with regard to the mean arrival rate at the ATM.
Period (1)
155 x 10%* = 15.5
*Assumed traffic increase by the ATM w/in each open time period
155 x 1%* = 1.55
*Half of the percentage of expansion beyond the previous level of overall
client transactions (absorbed by the ATM)
155 x 30%* = 46.5
*Percentage of diversion from human tellers to ATM
Total = 63.55 64 customers
Period (2)
242 x 10%* = 24.2
*Assumed traffic increase by the ATM w/in each open time period
242 x 1%* = 2.42
*Half of the percentage of expansion beyond the previous level of overall
client transactions (absorbed by the ATM)
242 x 30%* = 72.6
*Percentage of diversion from human tellers to ATM
Period (1)
= 54 / hour
S = 54 / 24 = 2.25 3 tellers
Period (2)
= 169 / hour
S = 169 / 24 = 7.041666667
8 tellers
ALTERNATIVE SOLUTION:
Mean Arrival at Human Teller Window
Period (3)
= 102 / hour
S = 102 / 24 = 4.25 5 tellers
Period (4)
= 129 / hour
S =129 / 24 = 5.375 6 tellers
Period (1)
155 x 10% = 15.5 16
155 x 2% = 3.1 4 / 2 = 2
Total = 18
Period (2)
242 x 10% = 24.2 25
242 x 2% = 4.84 5 / 2 2.5 = 3
Total = 28
Period (3)
290 x 10% = 29
290 x 2% = 5.8 6 / 2 = 3
Total = 32
Period (4)
554 x 10% = 55.4 56
554 x 2% = 5.8 6 / 2 = 3
Total = 62
Mean Arrival Rate at ATM
Period (1)
155 x 0.30 = 46.5 47
Human Tellers 155 - 47 = 108
ATM
47 + 18 = 65
Period (2)
242 x 0.30 = 72.6 73
Human Tellers 242 - 73 = 169
ATM
73 + 28 = 101
Period (3)
290 x 0.30 = 87
Human Tellers 290 - 87 = 203
ATM
87 + 32 = 119
Period (4)
544 x 0.30 = 166.2 167
Human Tellers 554 - 167 = 387
ATM
167 + 62 = 229
4. Assume that the number of human tellers is one more than that found in
part (c) of Question 3. Determine for Ms. Perez the mean customer waiting time
during each open period for those customers who seek the following:
a. Human tellers
Period 1
P(0)
1
(7.6796875)+1.067871094(2.285714286)
P(0)
P(0)
= =
54
s 4x24
1
10.12053571
0.098808998 or 9.88%
=
0.5625
Period 2:
P(0) =
1
827.3348168+117.3048945(4.595744681)
P(0) =
1
1366.438162
P(0) = 0.000731829 or 0.07%
=
= 169 = 0.782407407
Expected number of calling units in the
s
9x24
queue
Lq
= 0.00731829 (169/24)9 (0.782407407)
9!(1-0.782407407)2
= 1.418634399
5.098839243
Period 4:
= = 129 = 0.767857142
s 7x24
Expected number of calling units in the
queue
Lq
= 0.003800289 (129/24)7(0.767857142)
7!(1-0.767857142)2
=
378.2216046
271.6071449
=
1.392531867
Expected number of calling units in the
system
Ls
=
1.392531869 + (129 / 24)
=
6.767531867
Period 3:
= = 102
s 6x24
= 0.70833333
Ls
= 30 seconds/customer
= 120 customers/hour
Period 2:
= 100 customers/hour
= 100 customers/hour
Expected Time Waiting in Line:
Wq = 100 / [120 (120 - 100)]
= 100 / [120 (20)]
= 100 / [2,400]
= 0.041666667 hours x 60
minutes/hour
= 2.5 mins or 2 mins. 30
secs.
Period 3:
= 59.5 customers/hour
= 59.5 customers/hour
Expected Time Waiting in Line:
Wq = 59.5 / [120 (120 - 59.5)]
= 59.5 / [120 (60.5)]
= 59.5 / [7,260]
= 0.008195592287 hours x
3,600 seconds/hour
= 29.50 seconds
Period 4:
= 76 customers/hour
= 76 customers/hour
Expected Time Waiting in Line:
Wq = 76 / [120 (120 - 76)]
= 76 / [120 (44.)]
= 76 / [5,280]
=0.0143939 hours x 3,600
seconds/hour
= 51.82 secs
= 30 seconds/customer
= 120 customers/hour
Expected Time in the System:
Ws
= 1 / (120 - 100)
= 1 / 20
(or add 30 seconds service time)
= 0.05 hours x 60 minutes/hour
= 3 minutes
= 30 seconds/customer
= 120 customers/hour
5. Consider the complete 24-hour, 7-day picture. Incorporate whatever information you
need from Questions 1 through 4 and your solutions, plus any additional information in
the case and any necessary assumptions, to compare the net cost of operation with and
without the ATM. Then give your overall recommendation to Ms. Perez.
COMPARISON NO. 1: Without ATM min VS Without ATM min+1
WITHOUT ATM MIN
Total Tellers Wage
Peri
(# of tellers)(# of
od
hours)(wage rate)
Total AHQHC
(hourly AHQHC)(# of
hours)
(1)
(3)(2)(50)
= 300
(139.39)(2)
= 278.78
578.78
2,893.90
(2)
(9)(1)(50)
= 450
(174.77)(1)
= 174.77
624.77
3,123.85
(3)
(5)(2)(50)
= 500
(798.47)(2)
= 1,596.94
2,096.94
10,484.70
(4)
(7)(3)(50)
= 1050
(146.03)(3)
= 438.09
1,488.09
1,488.09
Total AHQHC
(hourly AHQHC)(# of
hours)
(1)
(4)(2)(50)
= 400
(19.07)(2)
= 38.14
438.14
2,190.70
(2)
(10)(1)(50) = 500
(53.10)(1)
= 53.1
553.1
2,765.50
(3)
(6)(2)(50) = 600
(65.76)(2)
= 131.52
731.52
3657.60
(4)
(8)(3)(50) = 1200
(39.49)(3)
= 118.47
1,318.47
1,318.47
Total AHQHC
(hourly AHQHC)(# of
hours)(wage rate)
hours)
(1)
(4)(2)(50)
= 400
(19.07)(2)
= 38.14
438.14
2,190.70
(2)
(10)(1)(50) = 500
(53.10)(1)
= 53.1
553.1
2,765.50
(3)
(6)(2)(50) = 600
(65.76)(2)
= 131.52
731.52
3657.60
(4)
(8)(3)(50) = 1200
(39.49)(3)
= 118.47
1,318.47
1,318.47
Total Tellers
Wage
Total AHQHC
Total Cost per
(# of tellers)(# (hourly AHQHC)
Day
of hours)(wage (# of hours)
rate)
Period
(1)
(4)(2)(50)
= 400
(9.30)(2)
= 18.60
418.60
2,093.00
(2)
(9)(1)(50)
= 450
(42.56)(1)
= 42.56
542.56
2,462.80
(3)
(6)(2)(50)
= 600
(25.47)(2)
= 50.94
650.94
3,254.70
(4)
(7)(3)(50)
= 1050
(41.78)(3)
= 125.34
1,175.34
1,175.34
Net
Advantage
(1)
2,190.70
2,093.00
97.70
(2)
2,765.50
2,462.80
302.70
(3)
3657.60
3,254.70
402.90
(4)
1,318.47
1,175.34
143.13