Professional Documents
Culture Documents
Mayank Kakodkar
2007ME10496
Abhishek Kr Singh
2007ME10469
Supervisor:
Prof. Kiran Seth
Examiner:
Prof. Arun Kanda
Department of Mechanical Engineering
IITD
Abstract
In this project, the emphasis was on to study applied probability models and their
various theoretical and practical applications. Different sections of probability like
random variables, conditional probability, distributions, Markov chains, Queuing etc
were thoroughly studied and their various applications were both mathematically
and intuitively analyzed. Finally, as a part of the project, an optimum queuing model
for bank was developed to reduce waiting time of customer and to improve, overall,
efficiency of banking services.
Contents
Abstract...................................................................................................................... 0
List of Figures............................................................................................................. 3
List of Tables............................................................................................................... 4
Chapter 1 Literature review and scope of the work:...................................................5
Theory of Queuing Systems.................................................................................... 6
Chapter 2 LEARNING METHODOLOGY FOLLOWED:...................................................10
Chapter 3 Data collection and related calculations..................................................12
For counter no 1:................................................................................................... 19
For counter no 2:................................................................................................... 19
Chapter 4 Suggested modification........................................................................... 19
Chapter 5 Results and discussion:............................................................................22
List of Figures
Figure 1-1: elements of a single queue queuing system............................................7
Figure 3-1: Histogram of inter-arrival times of customers........................................13
Figure 3-2: Plot of observed and expected distribution of inter-arrival times...........14
Figure 3-3: Histogram of service times at counter no. 1...........................................15
Figure 3-4: Plot of observed and expected distribution of service times at counter
no. 1......................................................................................................................... 16
Figure 3-5: Histogram of service times at counter no. 2...........................................17
Figure 3-6: Plot of observed and expected distribution of service times at counter
no. 2......................................................................................................................... 18
Figure 3-7: Congestion distribution at Bank..............................................................18
Figure 4-1: Suggested arrangement.........................................................................19
Figure 4-2: Plot of service times for one M/G/2 queue..............................................20
Figure 4-3: Plot of observed and expected distribution of service times for one M/G/2
queue....................................................................................................................... 21
List of Tables
Table 3-1: Descriptive statistics of inter-arrival times of customers.........................12
Table 3-2: Frequency table of inter-arrival times of customers.................................12
Table 3-3: Summary of observed and expected distribution of inter-arrival times....13
Table 3-4: Descriptive statistics of service times at counter no. 1............................14
Table 3-5: Frequency table of service times at counter no. 1...................................14
Table 3-6: Summary of observed and expected distribution of service times at
counter no. 1............................................................................................................ 15
Table 3-7: Descriptive statistics of service times at counter no. 2............................16
Table 3-8: Frequency table of service times at counter no. 2...................................16
Table 3-9: Summary of observed and expected distribution of service times at
counter no. 2............................................................................................................ 17
Table 4-1: Descriptive statistics of service times for one M/G/2 queue.....................20
Table 4-2: Frequency table of service times for one M/G/2 queue............................20
Table 4-3: Summary of observed and expected distribution of service times for one
M/G/2 queue............................................................................................................. 21
The various problems to be solved as exercised were solved and its various intuitive
solutions were stressed upon. Also, applications of each theoretical problem in
similar practical problems prevailing in industry were discuswsed with supervisor.
Finally, as a part of this project, an optimum queueing model for Canara bank
was developed using application of queueing theory and various changes to be
done to improve efficacy of the bank applications were recommended.
8
3) SIRO (Serve In Random Order).
4) Priority Queue that may be viewed as a number of queues for various priorities.
5) Many other more complex queuing methods that typically change the customers
position in the queue according to the time spent already in the queue, expected
service duration, and/or priority. These methods are typical for computer multiaccess systems.
Most quantitative parameters (like average queue length, average time spent in the
system) do not depend on the queuing discipline. Thats why most models either do
not take the queuing discipline into account at all or assume the normal FIFO queue.
In fact the only parameter that depends on the queuing discipline is the variance (or
standard deviation) of the waiting time. There is this important rule (that may be
used for example to verify results of a simulation experiment):
The two extreme values of the waiting time variance are for the FIFO queue
(minimum) and the LIFO queue (maximum).
Theoretical models (without priorities) assume only one queue. This is not
considered as a limiting factor because practical systems with more queues (bank
with several tellers with separate queues) may be viewed as a system with one
queue, because the customers always select the shortest queue. Of course, it is
assumed that the customers leave after being served. Systems with more queues
(and more servers) where the customers may be served more times are called
Queuing Networks.
Service represents some activity that takes time and that the customers are
waiting for. Again take it very generally. It may be a real service carried on persons
or machines, but it may be a CPU time slice, connection created for a telephone call,
being shot down for an enemy plane, etc. Typically a service takes random time.
Theoretical models are based on random distribution of service duration also called
Service Pattern. Another important parameter is the number of servers. Systems
with one server only are called Single Channel Systems, systems with more servers
are called Multi Channel Systems.
Output represents the way customers leave the system. Output is mostly ignored
by theoretical models, but sometimes the customers leaving the server enter the
queue again ("round robin" time-sharing systems).
Queuing Theory is a collection of mathematical models of various queuing
systems that take as inputs parameters of the above elements and that provide
quantitative parameters describing the system performance.
Because of random nature of the processes involved the queuing theory is rather
demanding and all models are based on very strong assumptions (not always
9
satisfied in practice). Many systems (especially queuing networks) are not soluble at
all, so the only technique that may be applied is simulation.
Nevertheless queuing systems are practically very important because of the typical
trade-off between the various costs of providing service and the costs associated
with waiting for the service (or leaving the system without being served). High
quality fast service is expensive, but costs caused by customers waiting in the
queue are minimum. On the other hand long queues may cost a lot because
customers (machines e.g.) do not work while waiting in the queue or customers
leave because of long queues. So a typical problem is to find an optimum system
configuration (e.g. the optimum number of servers). The solution may be found by
applying queuing theory or by simulation.
10
Examples:
D/M/1 = Deterministic (known) input, one exponential server, one unlimited FIFO or
unspecified queue, unlimited customer population.
M/G/3/20 = Poisson input, three servers with any distribution, maximum number of
customers 20, unlimited customer population.
D/M/1/LIFO/10/50 = Deterministic arrivals, one exponential server, queue is a stack
of the maximum size 9, total number of customers 50.
Activities
Study of
Applied
probability
and queuing
models
Data
collection
Data
Analysis and
development
of
appropriate
queuing
models
Evaluation of
model
Report
January
Year 2010
February
March
April
11
writing
Table 3-1: Gantt Chart
12
Variable
inter-arrival time
1
1
=
=0.658 customers per minute
E [t ] 87.5
Maximum
350
Std.Dev.
81.4
Category
-50.0000<x<=0.000000
0.000000<x<=50.00000
50.00000<x<=100.0000
100.0000<x<=150.0000
150.0000<x<=200.0000
200.0000<x<=250.0000
250.0000<x<=300.0000
300.0000<x<=350.0000
Missing
Cumulative %
of All
0
42
69
79
87
95
97
100
100
13
No. of obs.
60
50
40
30
20
10
0
-50
50
100
150
200
250
300
350
Figure 4-2 shows the Histogram of inter-arrival times of customers. The line shows
the expected inter arrival time of the customers.
Variable: inter-arrival time, Distribution: Exponential (int data)
Chi-Square = 12.68854, df = 7 (adjusted) , p = 0.08007
Upper
Observed Cumulative
Percent Cumul. % Expected
Boundary
Frequency
Observed
Observed Observed Frequency
<= 30.00000
57
57
30.2
30
54.8
60.00000
34
91
18.0
48
38.9
90.00000
34
125
18.0
66
27.6
120.00000
16
141
8.5
75
19.6
150.00000
9
150
4.8
79
13.9
180.00000
7
157
3.7
83
9.9
210.00000
14
171
7.4
90
7.0
240.00000
7
178
3.7
94
5.0
270.00000
3
181
1.6
96
3.5
300.00000
3
184
1.6
97
2.5
330.00000
3
187
1.6
99
1.8
< Infinity
2
189
1.1
100
4.4
Cumulative
Expected
55
94
121
141
155
165
172
177
180
183
185
189
Table 4-4 shows the Summary of observed and expected distribution of inter-arrival
times of the customers and Figure 4-3: Plot of observed and expected distribution of
inter-arrival times. The line shows the expected distribution.
14
60
No. of observations
50
40
30
20
10
30
60
90
120
150
180
210
240
270
300
330
360
Std.Dev.
54
Category
-50.0000<x<=0.000000
0.000000<x<=50.00000
50.00000<x<=100.0000
100.0000<x<=150.0000
150.0000<x<=200.0000
200.0000<x<=250.0000
250.0000<x<=300.0000
Missing
% of all
Cases
0.0
8.8
23.3
42.5
17.6
5.2
2.6
0.0
Cumulative %
of All
0
9
32
75
92
97
100
100
Figure 4-4: Histogram of service times at counter no. 1. And Table 4-7: Summary of
observed and expected distribution of service times at counter no. 1
Let E[t1] is the mean service time of a customer and 1 is the mean service rate.
Then from Table 4-5:
15
1=
1
1
=
=0.491 customers per minute
E[t 1 ] 122
Histogram: queue1
K-S d=.06589, p> .20; Lil iefors p<.05
Expected Normal
100
90
80
No. of obs.
70
60
50
40
30
20
10
0
-50
50
100
150
200
250
300
Cumulative
Expected
9
24
53
94
135
166
183
190
192
193
Table 4-7: Summary of observed and expected distribution of service times at counter no. 1
Figure 4-5: Plot of observed and expected distribution of service times at counter no.
1.
16
50
No. of observations
40
30
20
10
30
60
90
120
150
180
210
240
270
300
Figure 4-5: Plot of observed and expected distribution of service times at counter no. 1
Std.Dev.
122.8
Category
-100.000<x<=0.000000
0.000000<x<=100.0000
100.0000<x<=200.0000
200.0000<x<=300.0000
300.0000<x<=400.0000
400.0000<x<=500.0000
500.0000<x<=600.0000
Missing
% of all
Cases
0.0
10.0
32.5
35.0
12.5
3.8
6.3
0.0
Cumulative %
of All
0
10
43
78
90
94
100
100
Let E[t2] is the mean service time of a customer and 2 is the mean service rate.
Then from Table 4-5:
2=
1
1
=
=0.252 customers per minute
E[t 2] 238.4
17
Figure 4-6: Histogram of service times at counter no. 2 and Table 4-10: Summary of
observed and expected distribution of service times at counter no. 2. Figure 4-7:
Plot of observed and expected distribution of service times at counter no. 2.
Histogram: queue2
K-S d=.12347, p<.20 ; Lil iefors p<.01
Expected Normal
35
30
No. of obs.
25
20
15
10
-100
100
200
300
400
500
600
Cumulative
Expected
5.86
13.41
25.39
40.43
55.37
67.12
74.43
78.04
79.44
80.00
Table 4-10: Summary of observed and expected distribution of service times at counter no.
2
18
No. of observations
20
15
10
60
120
180
240
300
360
420
480
540
600
Figure 4-7: Plot of observed and expected distribution of service times at counter no. 2
Figure 4-8: Congestion distribution at Bank, it indicates that about 29% people come
for withdrawal and depositing cash whereas rest of the people visit bank primarily
for passbook updating, draft, buying forms etc.
Let
19
For counter no 1:
1= 0.4651 customers per minute
E[t1] =121.96 sec
12 E [ t 21 ]
Average queue length , L1=
=9.778 customers
2 ( 11 E [ t 1 ] )
L1
=21.02 min
1
Total time spent by customer for getting served at counter no 1 =23.94 min
For counter no 2:
2= 0.2262 customers per minute
E2 (t) =238.3625 sec
2
2
2 E [t 2]
Average queue length , L2=
=5.0141 customers
2 ( 12 E [ t 2 ] )
L2
=22.10 min
2
Total time spent by customer for getting served at counter no 2 =26.02 min
20
For case of M/G/2 the Table 5-11: Descriptive statistics of service times for
one M/G/2 queue and consequently Table 5-12: Frequency table of service times
for one M/G/2 queue and Figure 5-10: Plot of service times for one M/G/2 queue.
From Table 5-11
3=
1
1
=
=0.384 customers per minute
E[t 3] 156
Std.Dev.
96
Table 5-11: Descriptive statistics of service times for one M/G/2 queue
Category
-100.000<x<=0.000000
0.000000<x<=100.0000
100.0000<x<=200.0000
200.0000<x<=300.0000
300.0000<x<=400.0000
400.0000<x<=500.0000
500.0000<x<=600.0000
Missing
% of all
Cases
0.0
25.6
52.0
15.8
3.7
1.1
1.8
0.0
Table 5-12: Frequency table of service times for one M/G/2 queue
Cumulative %
of All
0
26
78
93
97
98
100
100
21
Histogram: MG2
K-S d=.13808, p<.01 ; Lil iefors p<.01
Expected Normal
160
140
120
No. of obs.
100
80
60
40
20
0
-100
100
200
300
400
500
600
22
60
No. of observations
50
40
30
20
10
-60
0
-30
60
30
120
90
180
150
240
210
300
270
360
330
420
390
480
450
540
600
510
570
630
Figure 5-11: Plot of observed and expected distribution of service times for one M/G/2
queue
Table 5-13 and Figure 5-11: Plot of observed and expected distribution of service
times for one M/G/2 queue respectively.
3= = 0.658 customers per minute
E[t3]=156.07 sec
2 E [ t 23 ] E [t 3 ]
2(2E [ t 3 ]) [
( E [ t 3 ])
n!
=7.316 minute s
( E [ t 3 ])
( 2E [ t 3 ])
23
if there are some training costs for employees or some more resources to be used at
each counter.
24
Chapter 7 References
Introduction to probability models by Sheldon M. Ross