Simulation Concepts and Methods doc

1
Chapter 9: Simulation
Concepts and Methods
Project risk analysis by
simultaneous adjustment
of forecast values.
2
Introduction

Simulation allows the repeated
solution of an evaluation model.

Each solution randomly selects
values from predetermined
probability distributions.

All solutions are summarized into
an overall distribution of NPV
values.

This distribution shows
management how risky the project
is.
3
Simulation Terminology

The treatment of risk by using simulation
is known as ‘stochastic’ modeling.

Other names for our term ‘Simulation’,
are - ‘Risk Analysis’, ‘Venture

Analysis’,’Risk Simulation’, ‘Monte Carlo
Simulation’.

The name ‘Monte Carlo Simulation’
helps visualization of repeated spins of
the roulette wheel, creating the selected
values.

Each execution of the model is known as
a ‘replication’ or ‘iteration’.
4
The Role of Simulation

Follows the initial creation and basic
testing of the representative model.

Is sometimes used as a test of the model.

Emphasizes the need for formal
forecasting, and requires close
specification of the forecast variables.

Draws managements attention to the
inherent risk in any project.

Focuses attention on accurate model
building.
5
Probability Distributions
of Forecast variables


Uniform: upper and lower
bounds required.
6
Probability Distributions
of Forecast variables

Uniform: upper and lower
bounds required.

Triangular: pessimistic,
most likely, and optimistic
values required
7
Probability Distributions
of Forecast variables

Uniform: upper and lower
bounds required.

Triangular: pessimistic,
most likely, and optimistic
values required

Normal: mean and variance
required.
8
Probability Distributions of Forecast
Variables


Uniform: upper and lower
bounds required.

Triangular: pessimistic, most
likely, and optimistic values
required

Normal: mean and variance
required.

Exponential: initial value and
growth factor required.
9
Process of Computation
per Replication

A value of a variable is selected from
its distribution using a random
number generator.

For example: Sales 90 units; selling
price per unit $2,350; component cost
per unit $1,100; labour cost per unit
$280.

These values are incorporated into
the model, and an NPV is calculated
for this replication.

The NPV for this replication is stored,

and later reported as one of many in
an overall NPV distribution.
10
Making the Replications

Each replication is unique.

Selection of values from the
distribution is made according to
the particular distributions

The automated process is driven
by a random number generator.

Excel add-ons such as ‘@Risk’ and
‘Insight’ can be used to streamline
the process.

About 500 replications should give
a good picture of the project’s risk.
11
Using the Output

Management can view the risk
of the project.

Probability of generating an
NPV between two given values
can be calculated.


Probability of loss is the area to
the left of a zero NPV.
12
Benefits and Costs of Simulation

Focuses on a detailed definition and analysis
of risk.

Sophisticated analysis clearly portrays the
risk of a project

Gives the probability of a loss making project

Allows simultaneous analysis of variables



Requires a significant forecasting effort.

Can be difficult to set up for computation.

Output can be difficult to interpret.