You are on page 1of 14

3

Managing the Event


Calendar in a Discrete-
event Simulation Model
In a discrete-event simulation experiment, the simulation is run by managing the event calendar. The
simulation typically starts with an arrival event that brings entity into the system. Each entity marks
a number of events, starting with the next arrival event as it flows through the system and until it
departs. There are a number of methods available to manage the event calendar. A brief description of
each of the two major methods is provided below.

3.1  Time AdvAnce meThods


3.1.1  Fixed Increment Time Advance
Simulation clock is advanced in increments of Dt time units. Selection of the time slice (Dt) is an
issue. After each update, a check is done to determine if any event should have occurred during the
interval of length Dt. If so, these are considered to have occurred at the end of the interval and system
state is updated accordingly.
Error is caused by treating the events occur at the end of each period. Error can be reduced by
reducing the length of the updating interval but that increases computation efforts. Fixed-increment
time advance has very limited use. For example, it may be used to model economic system.

3.1.2  Next-event Time Advance


Copyright © 2013. Pearson India. All rights reserved.

Simulation clock is initialized to zero and the time of occurrence of each future event is marked on
the event calendar. The simulation clock advances to the time of occurrence of the most imminent of
the future events. The system state is updated considering that the event has occurred. New events are
marked as an event is completed. Simulation stops when some specified stopping condition becomes
true. The stopping condition is typically based either on run length or on state of the system. Since all
state changes occur only at event times, periods of inactivity are skipped by jumping the clock from
one event time to the next event time.
The following numerical example shows how an event calendar is managed by the “next-event time
advance” method. The example models a simple system with single server and an infinite queue. The
example includes “arrival” events and “end-of-service” events only. As described earlier, an event
changes the state of the system. A single arrival event either increases number in queue or number in

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
70  System Simulation and Modeling

service by one. Upon arrival the entity checks if the server is busy. The entity joins the queue if the
server is busy. Otherwise, the entity is scheduled for service and an end-of-service event is marked on
the event calendar. Please note that the next arrival event is always created following an arrival event
but not after an end-of-service event. Scheduling a new arrival following an end-of-service event
makes the arrival process dependent on the system state. In case of independent arrivals to a system,
the next arrival must be scheduled right after the current arrival to capture the true sense of an inter-
arrival time distribution.
The next flow chart shows the sequence of steps that follows an arrival event (Fig. 3.1).

No
Entity arrives Schedules the Queue is
Insert into queue
to the system next arrival empty

Yes

Makes resource
busy

Marks end-of-
service event

Clock advances

Figure 3.1 Arrival event

3.1.2.1 Flow Chart – Arrival Event


As the simulation clock advances to an end-of-service event, an entity completes service and checks if
the queue is empty. If this is true, the entity makes the server idle, reports the statistics, and leaves the
system. In case the queue is not empty, the next entity is scheduled for service and an end-of-service
Copyright © 2013. Pearson India. All rights reserved.

event is marked before the current entity leaves the system. The next flow chart shows the sequence
of steps that follows an end-of-service event. Inter-event times are random in a system with random
inter-arrival pattern and random service time. Other events such as machine failure and machine
repair are marked if an unreliable machine is included in the system.

3.1.2.2 Flow Chart – End-of-Service Event


Problem description: An entity arrives to a single server facility by following an inter-arrival time
that follows Exponential (mean = 4) distribution. The service time follows Exponential (mean = 5)
distribution. Time is measured in minutes. Five entities are processed through the system. The
simulation starts at time t = 0. The following inter-arrival and service times have been generated by
using a random number table. The generator for inter-arrival time can be written as iat = (-4 ln R),

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 71

Entity ends Queue is No Schedules for


service empty service

Yes
Marks end-of-
service event

Makes the server


idle

Collect statistic

Leaves the system

Figure 3.2 End-of-service event

and for service time st = (-5 ln R), where R is a U(0, 1) random variable. Discussions on different
generators are included later in this book. Table 3.1 shows the first five inter-arrival and service times.

Table 3.1 Problem parameters

Inter-arrival time, iat = (-4 ln R) Service time, st = (-5 ln R)


R1 = 0.915, iat1 = 0.35 R6 = 0.674, st1 = 1.97
R2 = 0.259, iat2 = 5.40 R7 = 0.527, st2 = 3.20
R3 = 0.958, iat3 = 0.17 R8 = 0.301, st3 = 6.0
Copyright © 2013. Pearson India. All rights reserved.

R4 = 0.340, iat4 = 4.32 R9 = 0.402, st4 = 4.55


R5 = 2.89, iat5= 2.89 R10=0.385, st5 = 4.77

The first arrival (A1) takes place at t = 0, the inter-arrival time is 0.35 min, so the second arrival
occurs at t = 0.35. The first entity is scheduled for service. The first service time = 1.97 min and an
end-of-service event (ES1) is scheduled at t = 1.97 min. The events marked on the event calendar are
second arrival at t = 0.35 and first end-of-service at t = 1.97. The server is made busy. There are no
other activities to be carried out at t = 0 and clock advances to the most imminent future event.
The clock advances to t = 0.35. At this time, the second entity joins the system. Since the server
is busy, the entity joins the queue. The state variables, number in queue and number in system, each

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
72  System Simulation and Modeling

is incremented by one and the entity joins the queue. No end-of-service event is marked at this time.
However, the next arrival, arrival of the third entity is (A3) scheduled at t = 0.35 + 5.40 = 5.75 and an
arrival event is marked on the event calendar.
The clock now advances to t = 1.97 to indicate the end-of-service of the first entity. The second
entity is removed from the queue and scheduled for service. An end-of-service event (ES2) is marked at
t = 1.97 + 3.20 = 5.17. Before the entity leaves the system we must collect the relevant statistics such
as time-in-queue and time-in-system. The state variables are updated and the first entity is disposed.
The clock now advances to t = 5.17 and the second entity completes service. There is no entity
waiting for service and so the server is made idle. No new event is marked. The system is now empty
and server is idle. The second entity is disposed.
The clock advances to t = 5.75 and the third entity joins the Table 3.2 Time stamps data
system. The server is idle. The entity is scheduled for service, the
A1 0
server made busy, and a new end-of-service time (ES3) is marked on
the event calendar at t = 5.75 + 6.0 = 11.75. The next arrival (A4) is A2 0.35
scheduled at t = 5.75 + 0.17 = 5.92. ES1 1.97
The clock advances to t =5.92 and the fourth arrival occurs. The
entity joins the queue since the server is busy. The next arrival (A5) is ES2 5.17
marked on the calendar at t = 5.92 + 4.32 = 10.24. A3 5.75
The simulation continues by advancing the clock to the most
A4 5.92
imminent event time and taking care of the event. New events
are marked and the state variables are updated as necessary. The A5 10.24
simulation stops when any one of the stopping criteria is satisfied.
ES3 11.75
Most often the simulation stops after a specified length of time but a
simulation can terminate when the system reaches a specific state or ES4 16.3
a specified event occurs. Table 3.2 shows the time stamps of different ES5 21.07
events as five entities are processed.
Copyright © 2013. Pearson India. All rights reserved.

0.0 0.35 1.97 5.17 5.75 10.24 11.75 16.30 21.70


A1 A2 ES1 5.92 A5 ES3 ES4 ES5
ES2 A3A4
Time
Figure 3.3 Event time stamps

Ai – Arrival time of ith entity, ESi – End-of-service event time of ith entity; i = 1 … 5 (e.g., A1 – First
entity arives at time = 0, ES1 – First entity ends service at time = 1.97)
Table 3.3 summarizes the history of the first five entities processed by the system.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 73

Table 3.3 Summary of events

Entity no. Arrival time Waiting time End-of-service Time-in-system


time
1 0.0 0.0 1.97 1.97

2 0.35 1.62 5.17 4.82

3 5.75 0.0 11.75 6.0

4 5.92 5.83 16.30 10.38

5 10.24 6.06 21.07 10.83

Table 3.4 shows the state of the system at different event time. For simplicity, the state of the system
is described by the following three elements: (i) number of entities in the queue, (ii) number of entities
in the system, and (iii) status of the server (Busy or Idle). The state of a system is represented by one
or more state variables that collectively define the system completely at an arbitrary point in time.

Table 3.4 System state history

System state
Clock time
Server busy Number in queue Number in system
0 Yes 0 1

0.35 Yes 1 2

5.75 Yes 0 1

5.92 Yes 1 2

10.24 Yes 2 3

The waiting time and time in system are two typical measures of performance used in a discrete-
event system. The simulation ends at t = 21.07 if the objective is to find how long it takes to process five
Copyright © 2013. Pearson India. All rights reserved.

entities through the system. This simple example illustrates the basic mechanism of how a simulation
experiment is performed by a computer. The number of events marked and the book keeping efforts
required for each event increases as system complexity increases.
The following examples are expected to reinforce how events and system states are defined for
different systems. These examples also illustrate how the event calendar is managed.

System state dependent arrival


The following example illustrates how the events are marked on the event calendar when arrival to the
system depends on the state of the system. The following example describes a single machine system
where a new arrival takes place only after completion and departure of the current job. As expected, a
job on arrival will not wait for service leading to no queue build-up.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
74  System Simulation and Modeling

Example 3.1 Entity arrives to single server system with an inter-arrival time that follows exponential
distribution with mean = 11 min. The service time follows uniform distribution with parameters 4
and 6. A new entity cannot arrive to the system unless the previous entity has been serviced and has
left the system. We are modeling a “system state dependent arrival process” in this example. There are
a large number of real world examples where a new job arrives to a machine only when the machine
is idle. With single server in the system, there will be no queue build-up. In this example, the server
will be always busy and with one entity in the system. The objective is to mark first five events on the
event calendar.
Mark four events on the event calendar
The inter-arrival time follows exponential distribution with mean = 10 min.
The service time follows uniform distribution with parameter 4 and 6 min.
The generator for inter-arrival time is
IATi = -10 ln Ri R ~ U(0, 1).
The generator for service time is

STi = 4 + 2Ri R ~ U(0, 1)

Simulation starts at time t = 0. First entity arrives to the system.


Draw a random number R = 0.47, service time for the first entity = 4 + 2 * 0.47 = 4.94 min.
Mark first end-of-service at t = 4.94 (Event ES1).
Second arrival takes place at time t = 4.94 following departure of first entity:
t A1 = 0, t ES1 = 4.49, t A2 = 4.94
Second entity is scheduled for service immediately.
Draw a random number R = 0.18, service time = 4 + 2 * 0.18 = 4.36 min.
Second entity leaves at time = 4.94 + 4.36 = 9.3 min, tES2 = 9.3.
Third entity arrives immediately tA3 = 9.3.
Third entity scheduled for service.
Draw a random number R = 0.63, service time = 4 + 2 * 0.63 = 5.26 min.
Third entity leaves the system at time = 9.3 + 5.26 = 14.56 min, tES3 = 14.56.

tA1 tA2, tES1 tA3, tES2 ...


Copyright © 2013. Pearson India. All rights reserved.

0 4.94 9.30 14.56 ...

In case of system state dependent arrival, it is not required to define the inter-arrival time distribution
explicitly. The next entity arrives as soon as the current entity leaves the system. This example is very
similar to an automotive assembly line where the first machine is never starved and picks up the next
job as soon as the current job is complete the machine is released.
The example described next illustrates how an event calendar may be updated at fixed-time
interval while maintaining the correct order of the events. There are a few real-world applications of
such systems. The challenge lies in accurate processing of the events that occur within an updating
interval.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 75

Example 3.2 The following example illustrates how an investment grows under uncertain rate of
return over a fixed plan period and arbitrary deposit and withdrawal pattern. This problem may also
be viewed as updating an event calendar after a fixed-time interval. The state of the system is defined
as $ value of the investment at the end of every year after completion of all accounting activities. The
accounting activities include addition of interest to the principal and updating the current balance with
deposit and/or withdrawal amount as the case may be. Fixed-time interval updating of event calendar
is difficult due to keeping track of all events in the order it happened between two updates. In this
example, the current balance is updated at the end of each year by adding interest first to beginning of
year balance. This is followed by adding or taking away the specified amount from the current balance
for that year. The updated current balance is the state of the system. The clock now advances to the
next year and the process is repeated until we reach end of the plan period.
The investment plan simulated in this example has the following parameters:
Cash flow diagram
10,000 10,000 10,000 F

F = value of the
1 2 3 4 5 6 investment at the
end of 6 years

100,000 15,000 15,000

(i) Initial investment = $100,000, plan period = 6 years,


(ii) Interest rate of uniform (6%, 8%) in years 1 through 4,
(iii) Interest rate of uniform (7%, 9%) in years 5 and 6,
(iv) $15,000 deposited end of years 3 and 5,
(v) $10,000 withdrawn end of years 1, 2, and 4.

Initial investment $100,000


APR uniform (6, 8)% years 1…4, and uniform (7, 9)% in years 5 and 6
Copyright © 2013. Pearson India. All rights reserved.

Withdrawal $15,000 in years 1, 2, and 4


Deposit $10,000 in years 3 and 5
Find the value of the investment at the end of year 6.
For simplicity, assume that the interest is added at the end of every year.
This is a static simulation problem, if we are interested to estimate the future value of the investment
only. We can mark the system state (current $ value of the investment) at the end of each year.

Initial investment $100,000


Year 1 – draw a random number R = 0.79 interest earned in year 1 = 6 + 0.79(8 - 6) = 7.58%.
Balance at the end of year 1 = 100,000(1 + 0.0758) - 15,000 = 92,580.
Year 2 - R = 0.53 balance end of year 2 after withdrawal = 92,580(1 + 0.0706) - 15,000 = 84,116.15.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
76  System Simulation and Modeling

Year 3 - R = 0.36 balance end of year 3 after deposit = 84,116.15 (1 + 0.0672) + 10,000 =
99,768.75.
Year 4 - R = 0.64, interest earned = 6 + 0.64(8 - 6) = 7.28%, balance end if year 4 after withdrawal
= 92,031.91.
Year 5 - R = 0.16, interest earned = 7 + 0.16(9 - 7) = 7.32%, balance end if year 5 after deposit
= 108,768.65.
Year 6 - R = 0.45, interest earned = 7.9%, balance end of year 6 = 117,361.36.

100,000 92,580 84,116 99,769 92,032 108,769 117,361

0 1 2 3 4 5 6

The value of the investment at the end of year 6 = $117,361.36.

Example 3.3 The following example may be viewed as a static simulation if we are interested to
find the project length only. It may also be viewed as a dynamic simulation if we are interested to find
the status of the project at an arbitrary point in time. We have taken the later approach to mark events
on the event calendar as the project moves toward completion. The network shows the activities A
through J with corresponding activity times. The activities D1 and D2 are dummy activities with zero
activity time for each. They are included in the network to show correct precedence relationships.

2 5 F 8

A ti ∼ N(7, 15)
E
D1 D2
G I 10
Copyright © 2013. Pearson India. All rights reserved.

1 B 3 C 4 D 6 7
H
J

t A ∼ N (5, 1)
t B ∼ N (7, 15)
tC = 8
t E ∼ N ( B, 2)
t D ∼ N (12, 1.5)

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 77

t F ∼ N (10, 0.8)
D1 = 0
D2 = 0
G ∼ N (9, 1)
I ∼ 10
H ∼ N (6, 0.06)
J ∼ N (10, 1)

At any point of time state of the system is defined as the set of tasks already completed.
An event is defined as the project reaches a specific node.
1 – starting node
.
.
.
9 – end node
t1 = start time of the project
t9 = end of the project
A node is reached when every activity leading to the node is complete.
No uncertainty associated with tasks C and I.
We mark event ti as the time to reach node i.
t1 = 0 start the project

Z = 7.01 t A = 5 + (0.17 *1) = 5.17 t 2 = t1 + 5.17 = 5.17


Z = −1.2 t B = 7 − (1.2 *1.5) = 5.2 t3 = Max{t 2 , t B } = Max {5.1 7, 5.2} = 5.2
t 4 = t3 + tC = 5.2 + 8 = 1 32
.

Z = 0.8 t E = 8 + (0.8 * 2) = 9.6 t5 = t 4 + t E = 1 32


. + 9.6 = 2 28
.
Z = 5.
00 t D = 12 + (0.05 *1.5) = 12.1 t6 = t 4 + t D = 13.2 + 12.1 = 25.3
t8 = t5 + t F = 22.8 + 9.28 = 32.1
Copyright © 2013. Pearson India. All rights reserved.

Z = −0.9 t F = 1 0− (0.9 * 0.8) = 9.2 8 t7 = Max {t6 + tG , t8 + 0}


Z = −0.6 tG = 9 − (0.6 *1) = 8.2 = Max {2 53
. + 8.4, 3 21. } = 3 37
.
Z = 2.8 t H = 6 + ( 2.8 * 0.6) = 7.7 t9 = t7 + t H = 3 37
. + 7.7 = 4 14
.
Z = 0.4
t10 = Max{t9 + tJ , t7 + tI }
= Max {41.4 + 10.4, 33.7 + 10}
= Max {51.8, 43.7}
=15 .8

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
78  System Simulation and Modeling

0 5.17 5.2 13.2 22.8 25.3 32.1 33.7 41.4 51.8

t1 t2 t3 t4 t5 t6 t8 t7 t9 t10

The state of the project may be defined as the “tasks already completed”. Each node represents a
point in time as the project is executed. We may think that the project advances from one node to the
next as shown on the event calendar.
The states of the system are summarized in Table 3.5.

Table 3.5 Time vs. state history

Time Node System state (list of tasks completed)


t=0 1 List is empty
t = 5.17 2 (A)
t = 5.2 3 (A, B)
t = 13.2 4 (A, B, C)
t = 22.8 5 (A, B, C, E)
t = 25.3 6 (A, B, C, E, D)
t = 32.1 8 (A, B, C, E, D, F)
t = 33.7 7 (A, B, C, E, D, F, G)
t = 41.4 9 (A, B, C, E, D, F, G, H)
t = 51.8 10 (A, B, C, E, D, F, G, H, I, J)

The project is completed at time t = 51.8 time units.


The state definition for this project is not unique. One may also define the state as the set of tasks
yet to be completed. This example is expected to reinforce the concept of marking events on the event
calendar. This example also indicates that the definition of the state of a system is not unique.
Copyright © 2013. Pearson India. All rights reserved.

3.2  dATA collecTion


The statistics collected from a system can be classified as either a statistic based on observation or a
time-persistent statistic.
Statistic based on observation: Time-in-system and time-in-queue are examples of statistic based
on observation. Each entity records a discrete value and individual observations are summarized and
reported. Report includes mean, standard deviation, maximum value, minimum value, number of
observations, etc.
Time-persistent statistic: Examples of time-persistent statistic include number in queue, resource
utilization, number in system, etc. A time-persistent statistic refers to a state variable that changes
value at discrete points in time. For example, the number of entities in a queue increases or decreases

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 79

as an entity joins a queue or leaves a queue. A resource alternates between a busy state and an idle state
assuming that the resource does not fail.

Summary

 This chapter describes how the event calendar is managed in a discrete-event simulation run.
 Four examples are described to show how events as well as state of a system are defined.
 Three examples relate to updating the event calendar after a random time interval. Only one
example relates to updating after a fixed-time interval.
 The “next-event time” rule is used to manage the event calendar in each example.
 The readers should get an appreciation of the amount of book keeping done by the software
in carrying out a simulation experiment. There is very limited use of “fixed-increment time
advance” method. Maintaining the order of the events and their impact on system states are the
challenges posed by this method.

Exercises

1. Mark the first 10 events on the event calendar for a single server system with following
parameters:
(a) The inter-arrival time distribution follows exponential (mean = 10) distribution,
(b) The service follows uniform (6, 8) distribution, both measured in minutes.
Identify all the events necessary to process six entities through the system. Use number
in system and server’s status as state variables and define the state of the system at each
marking.
2. Redo Problem 1 with the following changes:
(a) Change the inter-arrival time distribution to Expo(mean = 3) minutes,
(b) Change the service time distribution to Expo(mean = 10) minutes,
(c) There is a storage area of capacity 3 units in front of the server. An arriving entity is
turned away if the storage area is full.
Run the simulation long enough so that 10 entities are processed. Estimate: (a) server’s
Copyright © 2013. Pearson India. All rights reserved.

utilization and (b) number of customers turned away.


3. The system under study consists of single machine where a new job can join the system only after
the previous job is processed and leaves the system and the machine is set-up for the next cycle.
(a) The set-up operation requires normal (5, 0.25) minutes. You can start the simulation with
the machine ready to process a job.
(b) The service time follows normal (10, 0.5) distribution. Both inter-arrival time and service
time are measured in minutes.
Process 10 entities through the system and report the following:
(a) Mark each event on the event calendar,
(b) Estimate machine utilization.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
80  System Simulation and Modeling

4. What is the distinctive system behavior that you have noticed in the previous problem?
Describe a real-world example that mimics the condition described in the previous problem.
5. Consider an investment problem with following parameters:
(a) Initial investment = $15,000 and a plan period of 7 years,
(b) Interest rate is fixed at 4% and paid every 6 months,
(c) Annual deposit and withdrawal amounts each follow normal (2000, 100) distribution.
Deposits are made in years 1 through 4 and withdrawal in years 5 and 6. Estimate the
year-end $ value of the investment for years 1 through 7. Recommend another relevant
state variable for this problem. Explain.
6. Redo Problem 5 with the following changes:
(a) Interest rate is 2.5% paid every 3 months in years 1 through 3 and increases to 3% for the
remaining period,
(b) Only deposits of $2000 is made end of years 1, 3, and 5. No withdrawal is made from the
account.
Estimate the year-end $ value of the investment at the end of 7 years. Define an “event”
relevant for this problem. Explain the order of events that take place at the end of each quarter.
Problems 7 and 8 will be based on example Problem #4.
7. Refer to the network for a project shown next and estimate the following:

3
A F
E
D
J
1 2 5 6
B G
H
C I
4

Activity Duration(time units)


A Normal(5, 0.05)
B 4
Copyright © 2013. Pearson India. All rights reserved.

C Normal(8, 10)
D 0(Dummy activity)
E 6
F Normal(4, 0.25)
G 6
H 4
I Normal(6, 0.5)
J 5

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
Managing the Event Calendar in a Discrete-event Simulation Model 81

(a) Mark the time for each node,


(b) The time required to complete the project,
(c) The state of the project at t = 10, 15, and 20.
8. Refer to the project network shown next and answer the following:

Active Duration (time unit)


A 8
B 16
C Normal(10, 0.2)
D 4
E 5
F Normal(12, 1)
G 6
H 10
I 12
J Normal(8, 1)
K Normal(9, 1)

(a) Estimate the time required to reach each node,


(b) The time required to complete the project,
(c) State the project at each node expressed in terms of tasks yet to done.
9. The front lobby of a bank is managed by three servers. Each server is capable of meeting all
the needs of a customer. A customer joins a common queue and goes to an idle server when
the customer reaches at the top the queue. The inter-arrival time is modeled by an exponential
(mean = 5 min) distribution. The service time follows a triangular (5, 8, 12) distribution.
Estimate the following:
(a) The time on the clock when the fourth customer leaves the bank,
(b) Waiting time of the sixth customer,
(c) The time on the clock when all three servers are busy for the first time,
(d) The average utilization of the servers during the first hour of operation.
Copyright © 2013. Pearson India. All rights reserved.

10. Define: (a) The entity, (b) A set of relevant events, and (c) The system’s state for each of the
following systems.
(a) An one shop grocery store that is open 24 hours,
(b) The Secretary of State’s office where you pay car registration fees, renew license, change
titles, etc.
(c) Withdrawing money from an ATM machine in a crowded shopping mall,
(d) Waiting in the emergency department of a hospital,
(e) Quick oil and lubrication service center for automobiles.
Remember an entity arrives to the system, waits for the server to be idle, receives service when
the server is ready and then departs.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.
This page is intentionally left blank.
Copyright © 2013. Pearson India. All rights reserved.

Sengupta, Sankar. System Simulation and Modeling, Pearson India, 2013. ProQuest Ebook Central,
http://ebookcentral.proquest.com/lib/cepune/detail.action?docID=5125336.
Created from cepune on 2023-04-10 06:58:09.

You might also like