Simulation & System Modelling

Course No. MIS 514

HANNAN MIA
[email protected]
29/03/2014
1
 Learning outcomes and Reading
• The learning outcomes of this lecture are the following
– Introduce the importance of simulation modelling
and its advantages.
– Present the basics of simulation modelling
– Present the basic elements of simulation models.
• Reading
– Pidd M (2004) “Computer Simulation in
Management Science” 5th edition, John Wiley and
sons, Chichester. Chapters 1, 2.
– Robinson S (2004) “Simulation: The practice of
model development and use” John Wiley and sons,
Chichester. Chapter 1
2
 What is Simulation?
• The word “simulate” means “create a
likeliness”.
• Therefore, “simulation” means the creation
of some kind of “models” which replicate,
to some extent, the functioning of real
world systems.
• These systems may be biological,
physical, natural, human, or
organisational.
3
 Example
• Assume that you are managing an airport
terminal. From your observations, you noticed
that many passengers get late to airplanes
causing delays to flights.
You would like to know the following
-What are the sources of passengers’ delay.
-How many check in desks should be open
during the different periods of the day. Where
should these desks located.
-Where should the security checks carried out.
How many security check points you would
need.
4
 Example (Continued)

-How many security officers you would need. How they

should be allocated to different security check points.
-Where should the gates located and so on.
• The only way to answer these questions and assess the
possible changes to passengers’ delays in to build a
simulation model of the terminal.

• Therefore, the definition of simulation is

“Experimentation with a simplified imitation (on a
computer) of an operations system as it progresses
through time, for the purpose of better understanding
and/or improving the system” (Robinson 2004)

5
 Advantages of Simulation (1)
• Cost: Direct experimentation in the real world is
generally very expensive. An airport manager will not
build a new floor in a terminal to see how this will impact
passengers’ delay.
• Time: Direct experimentation is time consuming. It
takes years to build a new motorway to assess whether
it reduces road congestion or not.
• Replication: Events are not controllable in the real
world. Some events happen only once or it takes a long
period of time for an event to occur again. To evaluate
the impact of a Tsunami wave on the fish population, you
have to wait a long period of time to have the opportunity
to do it (and hope it will not cause damage)
6
 Advantages of Simulation (2)

• Safety: Many experiments in the real world are

dangerous or even illegal. So it is safer to use a
simulation model. If an air traffic control manager would
like to assess the implications of reducing vertical
separation between aircrafts, you would not try this in the
real world as the consequences may be disastrous.
• Legality: In many situations, organisations would like to
investigate the impact of actions which are not allowed
by the law. For example, an airline company would like
to assess the impact of introducing night flights on its
performance. The police may investigate the impact of
increased criminal activities on the effectiveness of the
police force employees. 7
Simulation and Mathematical Modelling
• Many systems are represented by mathematical
models, which are used to determine optimal or
good decisions to improve the system.
• However, mathematical models have some
serious limitations
– They cannot deal effectively with time related
dynamic systems
– They are over simplified and includes generally
unrealistic assumptions
– They are too complex to be solved analytically,
• Therefore, simulation is a better, or sometimes
the only, methodology to solve such problems.
8
 Simulation Modelling General
Considerations
• Building a simulation model requires a
knowledge about the following
– The nature of the system being simulated.
– The objectives of the study and the expected results.
• The nature of the system is important because it
affects the selection of the simulation modelling
technique to be used.
• The objective of the study is important because
it indicates the level of details to be included in
the simulation model.
9
 Elements of Simulation Modelling
• Before building a simulation model, the analyst
has to make decisions about the following
– Time handling
– Stochastic or deterministic durations
• Time handling indicates how the state of the
system is updated during the simulation time
frame.
• Stochastic or deterministic durations indicate
whether the events durations in the simulation are
known (deterministic) or random (stochastic)

10
 Time Handling

• In simulation, the change in the system state is

modelled through time.
• A great advantages of simulation is that the
speed of time elapse can be controlled by the
modeller.
• For example, the functioning of a factory over a
one year period can be simulated in few minutes
or less.
• Therefore, it is important to know how the time
flow is to be handled in the simulation model.
11
 Time Slicing
• Time slicing is the simplest way to represent
time in simulation.
• Under this approach, the flow of time is moved
forward in equal time intervals, for example
every one hour, one day, two weeks, five years.
• In the case of time slice of one day, for example,
the changes in the system state between day T
and day (T+1) is updated at day (T+1).
• The main advantage of this approach is its
simplicity. The period to observe and record the
changes in the system state is the same over
the simulation time frame.
12
 Next Event Technique
• In many situations, the system remains idle for
some period of times when no event occurs, In this
situation, there is no need to update the system
state.
• In this situation, it is preferable that the system is
updated when it is known that a change is the
system is about to occur. These changes are
known as events.
• Therefore, this approach is known as next event
technique.
• under this method, the time intervals for system
update are different and are no longer pre- 13
determined
 Deterministic Systems
• A system is deterministic if its behaviour is completely
known and determined.
• The future behaviour of such systems can be predicted
accurately as there are no variations in the system’s
elements.
• For example, if the number of passengers arriving at a
train station platform per unit time is fixed and the trains
inter-arrival times do not change, it is possible to
determine exactly the number of passengers in the
platform at any time (provided that we know the train
capacity).
• In this case, the results of the simulation of the system’s
behaviour have no variability.
14
 Stochastic Systems
• A system is stochastic if its behaviour cannot be
completely predicted.
• However, in most cases, it is possible to have some
information regarding the times at which the stochastic
events may occur.
• This information is captured, generally, through probability
distributions.
• For example, if the treatment duration for tuberculosis is
Normal with a mean of 60 days and a standard deviation
of 6 days, we would expect that most patients will be
treated between 42 and 78 days.
• The decision regarding deterministic or stochastic
systems depend on the system characteristics and the
level of knowledge we have about the system.
15
 Simulation of Deterministic and
Stochastic Systems
• The simulation of any system is done by updating the
information about the elements of the system over time.
• These elements are represented by variables.
• The functioning of the system is captured through
equations linking the variables of the system.
• If the system is discrete, its variables are completely
determined. The simulation of this system is done
trough allocating the values to the variables and
updating them through the equations.
• If the system is stochastic, some of its variables are
random and have to be defined by their probability
distributions. The simulation of this system is done
through generating a sample of the random variables in
the system and use the sample values to update the
information in the system through the equations. 16
 Terminating and Non Terminating
Systems
• In a simulation, it is important to determine the
time frame in which the system is operating.
• This allows us to decide about the length of the
simulation period necessary to have
representative results.
• In general, there are two types of systems:
– Terminating systems; In these systems some natural
events starts, ends, or starts and ends the system’s
operations.
– Non terminating systems: In these systems, such
starting or ending events do not exist. The system
operates continuously over time.
17
 Examples
• A flight from London to Frankfurt is a terminating
system. The system’s operations start by the
aircraft leaving the gate at London and end by
the aircraft arriving at its determined gate at
Frankfurt airport.
• The air traffic control at an airport is a non
terminating system. Activities in this system do
not start and end by an event as they are carried
out continuously over time. A simulation of this
system over a period of time is only a partial
representation of a continuous stream of
activities.
18