Pramod Parajuli

Simulation and Modeling, CS-331

Chapter 9
Discrete System Simulation

a l
ep
itn
1
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

1
 Introduction

Event
Change in the state of the system
Discrete Systems
- Systems having discrete nature
- Can be of two types
1. Fixed time step – a timer or clock exists
2. Event-to-event (discrete event simulation) – the event
works as the clock i.e. self clocking
(figure 1.2, Law & Kelton, page 9)

The state of the system is not altered while the

transition is being occurred.

a l
Clock time

ep
Number recorded time

itn
2
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

2
 Introduction

Attributes
Properties of a given entity
Event notice
A record of an event to occur at the current or some
future time, along with any associated data
necessary to execute the event
Event list
A list of event notices for future events, ordered by
time of occurrence; also known as the future event
list (FEL)
Delay (conditional wait)

a l
A duration of time of unspecified indefinite length

ep
which is not known until it ends.

itn
3
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

3
 Introduction

Activity (unconditional wait)

A duration of time of specified length which is known
when it begins
Completion of an activity is an event (primary event)

Completion of a delay is conditional/secondary event.

Such events are not represented by event notices
nor do they appear on the FEL

Duration of activity
- Deterministic (given exact duration)
- Statistical (random draw from specified set with equal

a l
probability of picking an element)

ep
- Functional (defined by a function of any kind)

itn
4
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

4
 Introduction

Significant event simulation

Continuous systems with quiescent periods (interval
between events)
Low order polynomials are used to represent the
quiescent periods

a l
ep
itn
5
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

5
 Fixed time-step vs. Next-event
Start
Start

Read Input Data

Read Input Data

Generate and store random data (if any required)

Generate and store random data (if any required)

Initialize (including time = 0)

Initialize

t = t + dt
Find next potential event, and its time of occurrence

Find all events that occur, if any, during period (t, t + dt)
Let the event occur. Update the system state and clock
Let all these events occur. Update the system state
Extract its effect on statistics being gathered
Extract their effect on statistics being gathered

End of No
End of No simulation?
simulation?

l
Yes

a
Yes
Output desired statistics

ep
Output desired statistics

itn
Stop 6
Stop
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

6
 Components of next-event time-advance approach

System state
The collection of state variables necessary to describe
the system at a particular time
Simulation clock
A variable giving the current value of simulated time
Event list
A list containing the next time when each type of
event will occur
Statistical counters
Variables used for storing statistical information about
system performance

a l
Initialization routine

ep
itn
A subprogram to init the simulation model at time 0 7
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

7
 Components of next-event time-advance approach

Timing routine
A subprogram that determines the next event from the
event list and then advances the simulation clock to
the time when that event is to occur
Event routine
A subprogram that updates the system state when a
particular type of event occurs i.e. one event
routine for each event type
Library routines
A set of subprograms used to generate random
observations from probability distributions that

a l
were determined as part of the simulation model

ep
itn
8
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

8
 Components of next-event time-advance approach

Report generator
A subprogram that computes estimates (from the
statistical counters) of the desired measures of
performance and produces a report when the
simulation ends
Main program
A subprogram that invokes the timing routine to
determine the next event and then transfers
control to the corresponding event routine to
update the system state appropriately.

a l
ep
itn
9
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

9
 Generation of Arrival Patterns

- Generation of exogenous arrivals

- Trace driven simulation - the sequence of inputs may
also be generated from observations on a system
- If there is no interaction between the exogenous arrivals
and the endogenous events of the system, then the
creation of a sequence of arrivals in preparation for the
simulation is permissible otherwise arrivals are generated
as required
- Arrival of exogenous entity is an event and the arrival
time of the next entity is recorded as one of the event
times
- The event of entering the entity into the system is

a l
executed and the arrival time of next entity is calculated

ep
from inter-arrival distribution

itn
10
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

10
 Generation of Arrival Patterns

- Bootstrapping – the previous process of counting the

arrival of next entity as current entity is executed is
known as bootstrapping. It is a process of making one
entity create its successor.
- In bootstrapping, the simulation program have to keep
record of arrival of next entity only. Therefore, it is the
most preferred method of generating arrivals.
- The attributes for arriving entity can be assigned at the
generation time or later on depending on the
requirement.
- If the attributes do not affect the event execution, then
the attributes can be assigned later on.

a l
- But if attributes do affect the event execution, they must

ep
be assigned at the arrival time and retained until the

itn
event execution is finished. 11
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

11
 Simulation Programming Tasks

- 3 basic tasks ENTITIES

ATTRIBUTES
ACTIVITIES

1. Generate model
2. Simulate
GENERATE
3. Report SYSTEM IMAGE ROUTINES
MODEL

SIMULATE SIMULATION ALGORITHMS

REPORT REPORT GENERATORS

System image – Set of numbers created to represent

the state of the system

a l
Simulation algorithms – procedures that execute the

ep
cycle of actions

itn
12
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

12
 Simulation Algorithm
CREATE MODEL
Steps
1. Find the next potential
FIND NEXT POTENTIAL EVENT 1
event
2. Select an activity SELECT ACTIVITY 2

3. Test if the event can

be executed TEST IF IT 3
RECORD
- If direct measure, CONDITIONAL EVENTS N
CAN BE
DONE
either success or
Y
abandon
- If the event can be CHANGE IMAGE 4

executed later, then

it is known as conditional GATHER STATISTICS 5

event
4. Change the system Y

l
MORE?
image

a
ep
5. Gather statistics N

itn
GENERATE REPORT 13
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

13
 Gathering Statistics

Commonly used statistics parameters (included in

report);
1. Counts – no. of entities of particular type or no. of
times some event occurred
2. Summary measures – extreme values, mean values,
standard deviations
3. Utilization – fraction of time some entity is engaged
4. Occupancy – fraction of a group of entities in use on
the average
5. Distributions – queue lengths, waiting times etc.
6. Transit times – time taken for an entity to move from

a l
one part of the system to some other part

ep
itn
14
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

14
 Counters and Summary Statistics

- Counters used to accumulate totals, some level etc.

- Mean of N observations xi,
1 N
mean = ∑ xi
N i =1
- Standard deviation
1/ 2
 1 N
2
s= ∑ (m − xi ) 
 ( N − 1) i =1 

a l
ep
itn
15
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

15
 Measuring Utilization and Occupancy

Measuring the load on some entity is most common

requirement of a simulation
What fraction of the time the item was engaged during the simulation
run?
Measured by utilization factor
Let time tb be the time the item becomes busy, and tf be the
time when the item becomes free

(fig. 8-11) Gordon

The interval tf - tb is added to a counter. At the end, the

utilization is derived as; N

a l
1
U = .∑ (t f − tb ) r

ep
T r =1

itn
16
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

16
 Measuring Utilization and Occupancy

A discrete simulation program, updating time as events occur,

will measure the intervals tf - tb directly
A continuous simulation program updating time in small
intervals need to build up the count by counting the
number of intervals in which the item is busy. For
example, if for ‘x’ number of counts item was busy then
the fact must be preserved.
If the items was busy at the end of the simulation, then add
a count according to the amount of time left in the
current count as the trail of previous run
Same holds for if the item was busy at the beginning of the
simulation run N
1

l
∑

a
For a group of such items; A = . n .(t − t )

ep
r r +1 r
T r =1

itn
17
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

17
 Measuring Utilization and Occupancy

If there is an upper limit on the number of entities, then the

occupancy indicates the average number in use as a
ration to the maximum
If ‘M’ is the maximum number of entities, and the quantity nr
is the number busy in the interval tr to tr+1, the average
occupancy, assuming the number n, changes N times is;
N
1
=
NM
∑ n (t
r =1
r r +1 − tr )

Utilization – timing information for each individual entity

Occupancy – count of a class of entities and last time the
count changed

la
ep
For large number of active entities, utilization cost more (time
and space) as compared with occupancy

itn
18
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

18
 Recording Distributions and Transit Times

To determine the distribution of a variable, we need to count

the no. of times the value of the variable falls within
specific intervals
For each new value, it is compared with predefined intervals
and for that interval, counted up once
All information is kept in a table that requires;
1. The lower limit of tabulation (L)
2. The interval size (dx)
3. The number of intervals (N)
(Fig. 8-13 Gordon)
To measure transit times, the clock is used in the manner of
a time stamp i.e. each entity holds time stamp. When

l
a
entity reaches a point from which transit time is started,

ep
the arrival time is noted. Later when the entity reaches

itn
to end point, the difference is computed 19
(C) Pramod Parajuli, 2005

cs
Source: www.csitnepal.com

19