Simulation Unit:1: I. Ii. Iii. IV
Simulation Unit:1: I. Ii. Iii. IV
Disadvantages of Simulation:
There are a number of problems with using simulation and these must not be ignored when deciding whether
or not it is appropriate.
i. Expensive: Simulation software is not necessarily cheap and the cost of model development and use may
be considerable, particularly if consultants have to be employed.
ii. Time consuming: It has already been stated that simulation is a time consuming approach. This only
adds to the cost of its use and means that the benefits are not immediate.
iii. Data hungry: Most simulation models require a significant amount of data. This is not always
immediately available and, where it is, much analysis may be required to put it in a form suitable for the
simulation.
iv. Requires expertise: Simulation modelling is more than the development of a computer program or the
use of a software package. It requires skills in, among other things, conceptual modelling, validation and
statistics, as well as skills in working with people and project management. These are the skills that are
discussed in this book. This expertise is not always readily available.
v. Overconfidence: There is a danger that anything produced on a computer is seen to be right. With
simulation this is further exacerbated with the use of an animated display, giving an appearance of reality.
When interpreting the results from a simulation, consideration must be given to the validity of the underlying
model and the assumptions and simplifications that have been made.
Explain with neat block diagram the framework for Conceptual Modelling.
Ans:
The below figure provides an outline of a framework for conceptual modelling. The purpose of this
framework is to provide a modeller with an understanding of how to develop a conceptual model. The
framework consists of four key elements:
Develop an understanding of the problem situation
Determine the modelling objectives
Design the conceptual model: the model content
Design the conceptual model: inputs and outputs
Figure - A Framework for Conceptual Modelling
Starting with an understanding of the problem situation, a set of modelling objectives are determined. These
objectives then drive the derivation of the conceptual model, first by defining the inputs and outputs, and
then by defining the content of the model itself. These elements are described in detail below.
Explain time slicing approach based on Simple Telephone Call Center Simulation.
Ans:
i. The simplest method for modelling the progress of time is the time-slicing approach in which a constant
time-step (Δt) is adopted. This is best explained with an example. In a telephone call centre, calls arrive
every 3 minutes and are passed to one of two operators who take 5 minutes to deal with the customer (below
figure). It is assumed for now that there is no variation in the inter-arrival time and the service time.
ii. Below table shows 24 minutes of simulation of the call centre with t set to 1 minute. Column two shows
the time remaining until a call arrives. Columns three and four show the time remaining until a customer
service is complete. The number of calls completed by each operator is calculated.
iii. It is relatively simple to set up a time-slicing simulation for this situation. The same approach could be
used for more complex situations, although the table would soon become very large and possibly
unmanageable by hand. By devising a flow chart outlining the sequence of activities it would be possible to
develop a computer program to perform the simulation, making larger-scale simulations possible. The time-
slicing approach can also be modelled easily in a spreadsheet.
iv. There are two main problems with the time-slicing approach. First, it is very inefficient. During many of
the time-steps there is no change in the system-state and as a result many computations are unnecessary. In
below table the only points of interest are when a call arrives, when an operator takes a call and when an
operator completes a call. In total there
are 22 such points as opposed to the 72 (24×3) calculations performed in below table. This problem is only
likely to be exacerbated the larger the simulation becomes.
Figure - Time Slicing Approach: Simple Telephone Call Centre Simulation
Table - Time Slicing Approach: Simple Telephone Call Centre Simulation
v. A second problem is determining the value of Δt. Albeit that a one-minute time-step seems obvious for
the example above, in most simulations the duration of activities cannot be counted in whole numbers. Also,
there is often a wide variation in activity times within a model from possibly seconds (or less) through to
hours, days, weeks or more. The discrete-event simulation approach addresses both of these issues.
Explain the requirements of Conceptual Model.
Ans:
There are four main requirements of a conceptual model:
1. Validity
2. Credibility
3. Utility
4. Feasibility
Validity:
a. A valid model is one that is sufficiently accurate for the purpose at hand. However, since the
notion of accuracy is of little meaning for a model that has no numeric output, conceptual model
validity might be defined more precisely as:
A perception, on behalf of the modeller, that the conceptual model will lead to a computer model that
is sufficiently accurate for the purpose at hand.
b. Underlying this notion is the question of whether the model is right. Note that this definition
places conceptual model validity as a perception of the modeller. It also maintains the notion that a
model is built for a specific purpose, which is common to most definitions of validity.
Credibility:
Credibility is similar to validity, but is taken from the perspective of the clients rather than the
modeller. The credibility of the conceptual model is therefore defined as:
A perception, on behalf of the clients, that the conceptual model will lead to a computer model that is
sufficiently accurate for the purpose at hand.
Utility:
The third concept, utility, is defined as:
A perception, on behalf of the modeller and the clients, that the conceptual model will lead to a
computer model that is useful as an aid to decision-making within the specified context.
a. Whereas the definitions of validity and credibility are specific to the modeller and the clients
respectively, utility is seen as a joint agreement about the usefulness of the model.
b. The concept of utility, as defined here, moves away from simply asking if the model is
sufficiently accurate, to whether it is useful. Within any context a range of
models could be designed, all of which might be sufficiently accurate for the purpose at hand. As such, all
these models would be valid and credible.
c. However, if a proposed model is large and cumbersome, albeit sufficiently accurate, it may have
limited utility. Indeed, a less accurate (but still sufficiently accurate), more flexible model that runs
faster may have greater utility by enabling a wider range of experimentation within a time-frame.
Feasibility:
a. The final requirement, feasibility, is defined as follows:
A perception, on behalf of the modeller and the clients, that the conceptual model can be developed
into a computer model.
b. Various factors may make a model infeasible. For instance, it might not be possible to build the
proposed model within the required time-scale, the data requirements of the model may be too
onerous, or there is insufficient knowledge of the real system to develop the proposed model.
Whichever, it is important that the conceptual model can be developed inton a computer model.
A final point to note is that these four concepts are not mutually exclusive. A modeller’s perception of
a model’s accuracy is likely to be highly correlated with the clients’perceptions of the same. Nor is an
infeasible model a useful model. It is useful,
however, to separate these concepts, so a modeller can be cognizant of them when
designing the conceptual model.
I. In order to demonstrate the three-phase approach a slightly more complex call centre example is
now introduced (Figure no. 1). Two types of customer (X, Y) make calls to the centre. Calls arrive
from a customer type X every 5 minutes and from a customer type Y every 10 minutes. Arriving calls
are placed in a queue (denoted by a circle) before the call router (a touch tone menu system) directs
the call to the right operator; an activity that takes 1 minute. There are two operators, the first takes
all customer X calls, the second all customer Y calls. Operator 1 takes exactly 4 minutes to deal with
a call and operator 2 exactly 7 minutes.
II. As a first step all of the B and C events for the system need to be defined. These are shown in
below tables respectively. Note the column that specifies which events are to be scheduled following
an event, for instance, the arrival of a customer type X leads to the next arrival being scheduled
(event B1). Since each C-event represents the start of an activity, they schedule the B-event that
represents the completion of
that activity. For events B4 and B5 the calls are output to the ‘‘world’’. This term means that the
calls are passed out of the model. Also note that for event B4 and B5
statistics are collected on the number of customers served. For each C-event the conditions for it to be
executed are specified.
Figure no. 1 - Discrete-Event Simulation Approach: Telephone Call Centre
Simulation.
Table no. 1 - Telephone Call Centre Simulation: B-Events
III. Having identified all the events, the system can be simulated. Figure no.2 outlines the three-phase
approach. At the start of the simulation the initial state of the model is determined. This may involve
placing work-in-progress in the model in order to create a realistic initial condition. The initial B-
events are also scheduled, for instance, the arrival of the first customers. Scheduled events are placed
into an event list that keeps a record of all future events that have been scheduled. The simulation
then moves into three phases that are continuously repeated.
IV. In the A-phase, which is also known as the simulation executive, the time of the next event is
determined by inspecting the event list. The simulation clock is then advanced to the time of the next
event. In the B-phase all B-events due at the clock time are executed. In the C-phase all C-events are
attempted and those for which the conditions are met are executed. Since the successful execution of
a C-event may mean that another C-event can now be executed, the simulation continues to attempt
C-events until no further events can be executed. The simulation then returns to the A-phase unless it
is deemed that the simulation is complete. Typically a simulation is run for a predetermined run-
length or possibly a set number of arrivals.
What are the three different methods of simulation modeling ? Explain discrete event based
modeling.
Ans:
1. By method in simulation modeling, we mean a general framework for mapping a real world system
to its model. A method suggests a type of language, or “terms and conditions” for model building.
To date, there exist three methods:
System dynamics
Discrete event modeling
Agent based modeling
Software models:
Unlike system dynamics, discrete event modeling is supported by tens (if not
hundreds) of software tools. There is no uniformly accepted language for specifying
discrete event models; interoperatibilty is not possible, and not even planned by
software vendors.