SIMULATION

Simulation is a powerful and widely used management science technique for analysis and study of
complex problems. It is one of the most widely used quantitative approaches for decision-making.

Simulation may be defined as a technique that imitates the operations of a real-world system as it
evolves overtime. This is normally done by developing simulation model. Simulation model usually
takes the form of a set of assumptions about the operation of the system, expressed as
mathematical or logical relations between objects of interest in the system.

The goal of simulation is to create an environment in which it is possible to obtain information about
alternative actions via experimentation. Simulation is used in many applied experiments, for
example:

1. Testing of medicines on animals in laboratories. In this case, the animal responses simulate
human responses.
2. Driving automobiles on test tracks. The test track simulates the actual environment in which
automobiles are driven
3. Testing wing designs of airplanes in wind tunnels, which stimulate flight conditions
4. Training airplane pilots in actual cabins under simulated conditions

Characteristics of Simulation

1. Simulation is not strictly a type of model because models in general represent reality,
whereas simulation imitates reality
2. Simulation is a technique for conducting experiments. It involves the testing of specific
values of the decision variables in the model and observing the impact on the output
variables
3. Simulation is a descriptive rather than a normative tool. There is usually no automatic search
for an optimal solution. Instead, a simulation describes or predicts the characteristics of a
given system under different circumstances. Once, these characteristics are known, the best
policy can be selected.
4. The simulation process often consists of the repetition of an experiment many, many times
to obtain an estimate of the overall effect of certain actions
5. Simulation is usually called for only when the problem under investigation is too complex to
be treated by analytical models or by numerical optimization techniques.

Advantages & Disadvantages of Simulation:

Advantages

1. Simulation is relatively straightforward and flexible

2. Simulation can be used to analyse large and complex real world situations that can not be
solved by conventional quantitative analysis models
3. Simulations allows ‘What-if?’ types of questions
4. Simulation do not interfere with the real world system
5. Simulation allow us to study the interactive effect of individual components or variables to
determine which ones are important
6. Simulation allows ‘time compression’
7. Simulation allows for inclusion of real world complications that most quantitative analysis
models can not permit
Disadvantages

1. Good simulation models can be very expensive

2. Development of simulation models can be time consuming
3. Simulation does not generate optimal solution to problems. It is a trial and erro approach to
arrive at solutions to problems
4. Each simulation model is unique

Types of Simulation:

1. Probabilistic Simulation: In this type of simulation, one or more of the independent

variables is probabilistic ie., it follows a certain probability distribution. Two sub categories
of probabilistic distribution are:
Discrete distribution which involve a situation with a limited number of events
Continuous distribution which refer to a situation involving variables with an unlimited
number of possible values that follow density functions such as normal distribution
2. Time dependent and time independent simulation: Time independent simulation refers to
a situation where it is not important to know exactly when the event occurred. Example, we
may know that the demand is 3 units per day but we donot care when during the day the
item was demanded. On the other hand, in waiting line problem, it is important to know the
precise time of arrival. In this case, the simulation is referred to as time dependent
simulation.
3. Visual Simulation: it is one of the most interesting developments in computer graphics. This
technique is also known as visual interactive modelling or visual interactive problem solving.
It uses computer graphic displays to present the impact of various managerial decisions
4. Business games: these are simulations of competitive decision making which may also
involve probabilistic simulation. Business games involve several participants who are
engaged in a game that simulates a realistic situation. For example, each team participating
in a business games tries to maximize its own profit by making periodic decisions in areas
such as production, inventory, marketing, investment, maintenance
5. Large system simulation: these involves complex simulations of corporations or even
national economics. Example, simulation may be applied in corporate planning, especially
the financial aspects.

Steps Involved in Monte Carlo Simulation:

1. Setting up a probability distribution for important variables

2. Building a cumulative probability distribution for each variable in step1
3. Establishing an interval of random numbers for each variable
4. Generating random numbers
5. Actually simulating a series of trials

The basic idea in Monte Carlo Simulation is to generate values for the variables which made up of
the model under study. Real world systems consists of a lot variables that are probabilistic in nature
and that need to be simulated.

Historical outcomes are examined to establish a probability distribution for a given variable. The
frequency of observation is divided by the total number of observations to determine the relative
frequency or the probability

The regular probability distribution is converted to a cumulative probability distribution

After establishing a cumulative probability distribution for each variables included in the simulation,
a set of numbers must be assigned to represent each possible value or outcome. These are referred
as random number interval, that have been selected by a totally random process

Using the cumulative probability distribution computed in step 2, we can set the interval of random
numbers for each level of variable.. the top end of each interval is always equal to the cumulative
probability percentage

There are many ways of generating random numbers. If the problem is very large and the process
being studied involves thousands of simulation trials, computer programs are available to generate
the random numbers needed. The random numbers may be picked by lottery method or choosing
numbers from a table of random digits which itself is generated by a computer program. It has the
characteristic that every digit or number in it has an equal chance of occurring.

The outcomes of an experiment may be simulated by simply selecting random numbers from the
table of random numbers. We can begin from anywhere in the random table and note the interval
into which the number falls and then select the variable corresponding to that interval.

Application of Simulation

1. New Product Development: The objective of this simulation is to determine the probability
that a new product will be profitable. A model relating profit to various probabilistic inputs
such as demand, cost is developed. The only controllable input is whether to introduce the
product or not. A variety of possible values will be generated for the probabilistic inputs and
the resulting profit will be computed using simulation model.
2. Airline Reservation: The objective is to determine the number of reservations an airline
should accept for a particular flight. A simulation model is developed relating profit for the
flight to a probabilistic input( the number of passengers with reservations who arrive at the
airport in time and use their reservation) and a controllable input, a variety of possible
values will be generated for the number of passengers who show up and the resulting profit
can be computed
3. Inventory policy: The objective is to choose an inventory policy that will provide good
customer service at a reasonable cost. A model is developed relating two output measures,
total inventory cost and the service level to probabilistic inputs such as product demand and
delivery lead time from vendors and controllable inputs such as the order quantity and the
reorder point. For each set of controllable inputs, a variety of possible values would be
generated for the probabilistic inputs and the resulting cost and service levels would be
computed.
4. Traffic Flow: The objective of this simulation is to determine the effect of installing a left turn
signal on the flow of traffic through a busy interaction. A model is developed relating waiting
time for the vehicles to get through the intersection to probabilistic inputs such as the
number of vehicle arrivals and the fraction that want to make a left turn, and controllable
inputs such as the length of time the left turn signal is on. For each setting of the controllable
inputs, values would be generated for the probabilistic inputs and the resulting vehicle
waiting times would be computed.
5. Waiting Lines: The objective is to determine the waiting times for customers at a bank’s
automated teller machine. A model is developed relating customer waiting times to
probabilistic inputs such as customer arrivals and service times and a controllable input, the
 number of ATM machines installed. For each value of the controllable input, a variety of
values would be generated for the probabilistic inputs and the customer waiting times
would be computed.
6. Estimating the probability of completing a project by the Dead-line: PERT network analysis
involves probabilistic estimates of three time estimates and a rough estimate of the
probability of meeting the dead line for completion of the project is determined analytically.
But the time estimate may be overly optimistic or pessimistic to a considerable amount.
7. Design and Operation of Manufacturing System: In scheduling jobs to machines or work
centres in manufacturing systems some complications such as occasional machine
breakdowns, defective items needing to be reworked, and multiple type jobs, etc., make it
difficult to use queuing models. Such complications are no problem for simulation.