C h a p t e r 19

Simulation
“The difference between management and administration (which is what the bureaucrats used
to do exclusively) is the difference between choice and rigidity.”
– Heller, Robert

PREVIEW
Simulation is one of the widely used technique by corporate managers as an aid for decision-making.
This technique uses a computer to simulate (imitate) the operation of any system or process. It is also
used to analyse systems that operate indefinitely. In such a case, the computer randomly generates and
records the occurrence of the events that drive the system as if it was physically operating. Recording the
performance of the simulated operation of the system for a number of alternative options of operating
procedures enables us to evaluate and compare these alternatives to choose the most desired one.

LEARNING OBJECTIVES
After studying this chapter you should be able to
z make distinction between analytical and simulation models.
z appreciate what simulation is and how it can be used.
z know several definitions of simulation and appreciate the importance of simulation modelling.
z understand the advantages and disadvantages of simulation.
z apply Monte Carlo simulation technique for solving various types of problems.
z develop random number intervals and use them to generate outcomes.
z know several types of computer languages that are helpful in the simulation process.
z know certain causes of simulation analysis failure and how these can be avoided.

CHAPTER OUTLINE
19.1 Introduction 19.11 Simulation of PERT Problems
19.2 Simulation Defined 19.12 Role of Computers in Simulation
19.3 Types of Simulation 19.13 Application of Simulation
19.4 Steps of Simulation Process • Conceptual Questions
19.5 Advantages and Disadvantages of Simulation • Self Practice Problems
19.6 Stochastic Simulation and Random Numbers ‰ Chapter Summary
19.7 Simulation of Inventory Problems ‰ Chapter Concepts Quiz
19.8 Simulation of Queuing Problems ‰ Case Study
19.9 Simulation if Investment Problems ‰ Appendix : The Seven Most Frequent Causes of
19.10 Simulation of Maintenance Problems Simulation and How to Avoid Them
674 Operations Research: Theory and Applications

19.1 INTRODUCTION
Mathematical models discussed in earlier chapters help decision-makers to choose a decision alternative
from the given list of decision alternatives to reach an optimal solution to a problem. Simulation that is
not an optimizing technique, helps decision-makers to perform experiment with new values of variables and/
or parameters in order to understand the changes in the performance or effectiveness of a real system and
to make better decision. Such ‘experiments’ allow to answer ‘what if ’ questions relating to the effects of
changes in the value of variables and/or parameters on the model response. Comparing the payoffs or
outcomes due to these changes into the model is referred to as simulating the model.
The following few examples illustrate scope of applications of simulation.
1. Aircraft designers use wind tunnels to simulate the effect of air turbulence on various structural
parts of an airplane before finalizing its design.
2. Aircraft pilots or Astronauts are trained in a simulator to expose them with various problems that
they are likely to face in the sky while flying real aircraft.
3. Hospital management may simulate alternative scheduling rules of the ambulances, their locations,
the response time to an emergency call and, of course, the overall service quality and the costs
incurred if the ambulances were to be configured in a certain way (in types, number, location,
scheduling and staffing).
4. Computer designers simulate with a computer system configuration (its speed, size, computing
qualities, memory and so on) in terms of costs of such configurations and resulting computational
service that it provides to users.
5. Managers simulate alternative work flows and use of new manufacturing technologies (such as
Just-in-Time manufacturing, flexible manufacturing, etc.) to design ‘new’ shop floor to get an
experience and a learning tool that would give greater confidence in the productivity and the
management of future system, at a relatively small cost.
6. A queuing system decision-makers simulate the effect of probabilistic nature of arrival rate of
customers and the service rate of the server to serve the customer on the cost of waiting against
the cost of idle time of service facilities in the queuing system.
Other problems, such as location of bank branches, the deployment of fire stations, routing and
dispatching when roads are not secured (where materials sent might not, potentially, reach their destination),
the location and the utilization of recreational facilities (such as parks, public swimming pools, etc.) and
many other problems could be studied through simulation.

19.2 SIMULATION DEFINED

Simulation is one of the operations research technique used for representing real-life problems through
numbers and mathematical symbols that can be readily manipulated. For example, games such as chess to
simulate battles, backgammon to simulate racing, and other games to simulate hunting and diplomacy were
already invented where decision-makers used simulation to gain the ability to experiment with a situation
under controlled conditions.
Today, a modern game like monopoly simulates the competitive arena of real estate. Many have played
baseball with a deck of cards which has hits, strikeouts and walks with a cardboard, diamond and plastic chips
as runners. The distribution of hits, runs and outs, etc., in a deck of cards serves as a realistic reflection of the
overall average with which each would occur in real life.
Now the availability of computer software makes it possible to deal with large quantity of details that can
be incorporated into a model and also the ability to conduct many ‘experiments’ (i.e. replicating all the
possibilities). Mathematician Von Neumann and Ulam, in the late 1940s, developed the term Monte Carlo
analysis while trying first to ‘break’ the Casino at Monte Carlo and subsequently, applying it to the solution
of nuclear shielding problems that were either too expensive for physical experimentation, or too complicated
for treatment by the already known mathematical techniques.
Few definitions of simulation are stated below:
z A simulation of a system or an organism is the operation of a model or simulator which is a

representation of the system or organism. The model is amenable to manipulation which would
 Simulation 675

be impossible, too expensive or unpractical to perform on the entity it portrays. The operation
of the model can be studied and for it, properties concerning the behaviour of the actual system
can be inferred. — Shubik
z Simulation is the process of designing a model of a real system and conducting experiments with Simulation model
this model for the purpose of understanding the behaviour (within the limits imposed by a criterion represents a
system using
or set of criteria) for the operation of the system. — Shannon
number and
z Simulation is a numerical technique for conducting experiments on a digital computer, which symbols that can be
involves certain types of mathematical and logical relationships necessary to describe the readily manipulated.
behaviour and structure of a complex real-world system over extended periods of time.
— Naylor et al.
z ‘X simulated Y’ is true if and only if
(i) X and Y are formal systems,
(ii) Y is taken to be the real system,
(iii) X is taken to be an approximation to the real system, and
(iv) The rules of validity in X are non-error-free, otherwise X will become the real system.
z Simulation is the use of a system model that has the designed characteristics of reality in order
to produce the essence of actual operation. — Churchman
These definitions pointed out that simulation can be equally applied to military war games, business games,
economic models, etc. Also simulation involves logical and mathematical modeling that involves the use
of computers to test the behaviour of a system using iterations or successive trials under realistic
conditions.
For operations research practioners, simulation is a problem solving technique that uses a
computer-aided experimental approach to study problems which otherwise is not possible through
analytical methods. Table 19.1 highlights what simulation is and what it is not.

It is It is not

• A technique which uses computers. • An analytical technique which provides exact

solution.
• An approach for reproducing the processes • A programming language (but it can be progra-
by which events of chance and change are ammed into a set of commands that can form a
created in a computer. language to facilitate the programming of
simulation)
• A procedure for testing and experimenting Table 19.1
on models to answer what if . . ., then so Simulation –
and so . . . types of questions. What it is/not

19.3 TYPES OF SIMULATION

There are several types of simulation. A few of them are listed below:
1. Deterministic versus probabilistic simulation The deterministic simulation involve cases in which a specific
outcome is certain for a given set of inputs. Whereas probabilistic simulation deals with cases that involves
random variables and obviously the outcome can not be known with certainty for a given set of inputs.
2. Time dependent versus time independent simulation In time independent simulation it is not important
to known exactly when the event is likely to occur. For example, in an inventory control situation, even
if decision-maker knows that the demand is three units per day, but it is not necessary to know when
demand is likely to occur during the day. On the other hand, in time dependent simulation it is important
to know the exact time when the event is likely to occur. For example, in a queuing situation the exact
time of arrival should be known (to know that the customer will have to wait).
3. Interactive simulation Interactive simulation uses computer graphic displays to present the conse-
quences of change in the value of input variation in the model. The decisions are implemented
interactively while the simulation is running. These simulations can show dynamic systems that evolve
over time in terms of animation. The decision-maker watches the progress of the simulation in an
animated form on a graphics terminal and can alter the simulation as it progresses.
676 Operations Research: Theory and Applications

4. Business games Business game simulation model involves several participants who need to play a
role in a game that simulates a realistic competitive situation. Individuals or teams compete to achieve
their goals, such as profit maximization, in competition or cooperation, with the other individuals or
teams. The few advantages of business games are:
(i) participants learn much faster and the knowledge and experience gained are more memorable
than passive instruction.
(ii) complexities, interfunctional dependencies, unexpected events, and other such factors can be
introduced into the game for evoking special circumstances.
(iii) the time compression – allowing many years of experience in only minutes or hours – lets the
participants try out actions that they would not be willing to risk in an actual situation and see
the result in the future.
(iv) provide insight into the behaviour of an organization. The dynamics of team decision-making
style highlight the roles assumed by individuals on the teams, the effect of personality types
and managerial styles, the emergence of team conflict and cooperation, and so on.
5. Corporate and financial simulations The corporate and financial simulation is used in corporate
planning, especially the financial aspects. The models integrate production, finance, marketing, and
possibly other functions, into one model either deterministic or probabilistic when risk analysis is
desired.

19.4 STEPS OF SIMULATION PROCESS

The process of simulating a system consists the following steps:
1. Defining the problem Define the scope of study and the level of details that is required to derive
desired results. Thus decision-makers should have clear idea about what is to be accomplished. For example,
if a simulation study is to be done for arrival patterns of customers in a queuing system, then scope of
study should decide certain hours of the day.
2. Identifying the decision variables and setting performance criterion Once problem is
defined, the next step is to understand objectives for using simulation and degree (extent) with which these
objectives shall be measured. In other words, before the start of simulation study, the decision-maker must
ascertain how a system will behave given a set of input variables (conditions). For example, in an inventory
control situation, the demand (consumption rate), lead time and safety stock are identified as decision
variables. These variables shall be responsible to measure the performance of the system in terms of the
total inventory cost under the decision rule – when to order.
3. Developing a simulation model For developing a simulation model, an understanding of the
relationships among the elements of the system being studied is required. For this purpose the influence
diagram (drawn in a variety of different ways) is useful. This is because simulation models for each of
these diagrams may be formulated until one seems better or more appropriate than the other. Even after
one has been chosen, it may be modified again and again before an acceptable version is arrived at.
4. Testing and validating the model The purpose of validation is to check whether the model
adequately reflects performance of real system. This requires comparing a model with the actual system –
a validation process. A validated model should behave similar to the system under study. Discrepancies
(if any) should be rectified in order to achieve objectives of simulation.
The validation process requires (i) determining whether the model is internally correct in a logical and
programming sense called internal validity and (ii) determining whether it represents the system under
study called external validity. The first step involves checking the equations and procedures in the model
for accuracy, both in terms of mistakes (or errors) and in terms of properly representing the system under
study.
After verifying internal validity, the model is tested by putting different values to variables into the
model and observing whether it replicates what happens in reality. The decision-maker can make changes
in the assumptions or input data and see the effect on the outputs. If the model passes this test, extreme
values of the input variables are entered and the model is checked for the expected output.
 Simulation 677

Fig. 19.1
Steps of
Simulation
Process

5. Designing of the experiment Experimental design refers to controlling the conditions of the study
such as the variables to be included and recording the effect on the output.
The design of experiment requires determining (i) the parameters and variable in the model, (ii) levels
of the parameters to use, (iii) the criterion to measure performance of the system, (iv) number of times the
model will be replicated, (v) the length of time of each replication, and so on. For example, in a queuing
simulation we may consider the arrival and service rates to be constant but the number of servers and the
customers waiting time may vary (dependent variable).
6. Run the simulation model Run the model using suitable computer software to get the results in
the form of operating characteristics.
7. Examine the outputs Examine the outputs of the experiments and their reliability. If the simulation
process is complete, then select the best course of action (or alternative), otherwise make desired changes
in model decision variables, parameters or design, and return to Step 3.
The steps of simulation process are also shown in Fig. 19.1.

19.5 ADVANTAGES AND DISADVANTAGES OF SIMULATION

Advantages
1. This approach is suitable to analyse large and complex real-life problems that cannot be solved by the
analytical methods.
2. It facilitates to study the interactive system variables, and the effect of changes that take place in these
variables, on the system performance in order to determine the desired result.
3. Simulation experiments are done on the model, not on the system itself. Experimentation takes into
consideration additional information during analysis that most quantitative models do not permit. In
other words, simulation can be used to ‘experiment’ on a model of a real situation, without incurring
the costs of operating on the system.
678 Operations Research: Theory and Applications

4. Simulation can be used as a pre-service test to try out new policies and decision rules for operating
a system before running the risk of experimentation in the real system.

Disadvantages
1. Simulation models are expensive and take a long time to develop. For example, a corporate planning
model may take a long time to develop and may also prove to be expensive.
2. It is the trial and error approach that produces different solutions in repeated runs. This means it does
not generate optimal solutions to problems.
3. The simulation model does not produce answers by itself. The user has to provide all the constraints
for the solutions that he wants to examine.

19.6 STOCHASTIC SIMULATION AND RANDOM NUMBERS

In simulation, probability distributions are used to quantify the outcomes in numerical terms by assigning
a probability to each of the possible outcomes. For example, if you flip a coin, the set of possible outcomes
is {H, T}. A random variable assigns a number to the possible occurrence of each outcome. In simulation,
random variables are numerically controlled and are used to simulate elements of uncertainty that are
defined in a model. This is done by generating (using the computer) outcomes with the same frequency
as those encountered in the process being simulated. In this manner many experiments (also called
simulation runs) can be performed, leading to a collection of outcomes that have a frequency (probability)
distribution, similar to that of the model under study.
To use simulation, it is necessary to generate the sample random events that make up the model. This
helps to use a computer to reproduce the process through which chance is generated in the actual situation.
Thus, a problem that involves many interrelationships among random variables can be evaluated as a
function of given parameters. Process generation (simulating chance processes) and modelling are therefore
the two fundamental techniques that are needed in simulation.
The most elementary and important type of process is the random process. This requires the selection
of samples (or events) from a given distribution so that the repetition of this selection process would yield
a frequency distribution of sample values that match the original distribution. These samples are generated
through some mechanical or electronic device – called pseudo random numbers. Alternately, it is possible
to use a table of random numbers where the selection of number in any consistent manner would yield
numbers that behave as if they were drawn from a uniform distribution.
Random numbers can also be generated using random number generator (which are inbuilt feature
of spread sheets and many computer languages) tables (see Appendix), a roulette wheel, etc.
Random numbers between 00 and 99 are used to obtain values of random variables that have a known
discrete probability distribution in which the random variable of interest can assume one of a finite number
of different values. In some applications, however, the random variables are continuous, that is, they can
assume any real value according to a continuous probability distribution.

19.6.1 Monte Carlo Simulation

The monte Carlo simulation approach is used to incorporate the random behaviour of variable(s) of interest
in a model. A formal definition of Monte Carlo Simulation is as follows:
• The Monte Carlo simulation technique involves conducting repetitive experiments on the model
of the system under study, with some known probability distribution to draw random samples
(observations) using random numbers.
The Monte Carlo simulation approach consists of following steps:
1. Setting up a probability distribution for variables to be analysed.
2. Building a cumulative probability distribution for each random variable.
3. Generating random numbers and then assigning an appropriate set of random numbers to represent
value or range (interval) of values for each random variable.
4. Conducting the simulation experiment using random sampling.
5. Repeating Step 4 until the required number of simulation runs has been generated.
6. Designing and implementing a course of action and maintaining control.
 Simulation 679

19.6.2 Random Number Generation

Monte Carlo simulation requires the generation of a sequence of random numbers where (i) all numbers
are equally likely, and (ii) no patterns appear in sequence of numbers. This sequence of random numbers
help in choosing random observations (samples) from the probability distribution.
Arithmetic computation The nth random number rn, consisting of k-digits, generated by using
multiplicative congruential method is given by:
rn ≡ p.rn – 1 (modulo m)
where p and m are positive integers, p < m, rn – 1 is a k-digit number and modulo m means that rn is the
remainder when p.rn – 1 is divided by m. This means, rn and p.rn – 1 differ by an integer multiple of m. To
start the process of generating random numbers, the first random number (also called seed) r0 is specified
by the user. Then, using above recurrence relation, a sequence of k-digit random number with period
h < m, at which point the number r0 occurs can be generated again.
For illustration, let p = 35, m = 100 and arbitrarily start with r0 = 57. Since m – 1 = 99 is a 2-digit number,
therefore, it will generate 2-digit random numbers:
r1 = p r0 (modulo m) = 35 × 57 (modulo 100)
= 1,995/100 = 95, remainder
r2 = p r1 (modulo m) = 35 × 95 (modulo 100)
= 3,325/100 = 25, remainder
r3 = p r2 (modulo m) = 35 × 25 (modulo 100)
= 875/100 = 75, remainder
The choice of r0 and p for any given value of m, requires great care, and the method used is also not
a random process because a sequence of numbers generated, is determined by the input data for the method.
Thus, the numbers generated through this process are pseudo random numbers because these are reprod-
ucible and hence, not random.
The recurrence relation can also be used to generate random numbers as decimal fraction between 0
and 1, with desired number of digits. For this, the recurrence relation un = r n /m is used to generate
uniformly distributed decimal fraction between 0 and 1.
Computer generator The random numbers that are generated by using computer software are uniformly
distributed decimal fractions between 0 and 1. The software works on the concept of cumulative
distribution function for random variables for which we seek to generate random numbers.
For example, for the negative exponential function, with density function f (x) = λe – λx, 0 < x < ∞ ,
the cumulative distribution function is given by:

or
F(x) = z
0
x
λ e − λ x dx = 1 − e −λ x

e −λ x = 1− F ( x )
Taking logarithm on both sides, we have:
−λ x = log [1 – F(x)]
or x = – (1/λ) log [1 – F(x)]
If r = F(x) is a uniformly distributed random decimal fraction between 0 and 1, then the exponen-
tial variable associated with r is given by:
xn = – ( 1/λ) log (1 – r) = – (1/λ) log r.
This is an exponential process generator since 1 – r is a random number and can be replaced by r.
Remark While drawing random numbers from the random number table, we may start with any number
in any column or row, and proceed in the same column or row to the next number. But a consistent, unvaried
(i.e. we should not jump from one number to another indiscriminately) pattern should be followed in drawing
random numbers. If random numbers are to be taken for more than one variable, then different random
numbers for each variable should be used.
A number of process generators for use with a digital computer are shown in Table 19.2.
680 Operations Research: Theory and Applications

Theoretical Probability Parameters Process Generators

Distribution for Random Variable, x

(a) Discrete Random Variables

x−a x − a +1
Uniform a, b = x where < r ≤
b−a b − a +1
a ≤ x ≤ b , r = random number

Binomial
n
n, p = Σ x i , where x i =
RS1, ri ≤ p
i =1 T 0, ri > p

p = prob. of success; n = number of trials

k −1 − log ri k − log ri
Poisson λ = k – 1, where Σ ≤ 1 ≤ Σ
i =1 λ i =1 λ
λ = mean arrival rate per unit of time
(b) Continuous Random Variables
Uniform a, b = a + (b, a ) r
Exponential λ = (–1/λ) log r

R| a , u≤a
Table 19.2
Normal µ, σ, a , b = Su, ha < u < b ; u = [( −2 log r1 ) 1/ 2 (cos 6 . 283 r2 ) σ + µ )]
Some Process
Generators
|T b , u≥b ; µ = mean, σ = standard deviation

19.7 SIMULATION OF INVENTORY PROBLEMS

Example 19.1 Using random numbers to simulate a sample, find the probability that a packet of 6
products does not contain any defective product, when the production line produces 10 per cent defective
products. Compare your answer with the expected probability.
Solution Given that 10 per cent of the total production is defective and 90 per cent is non-defective,
if we have 100 random numbers (0 to 99), then 90 or 90 per cent of them represent non-defective products
and the remaining 10 (or 10 per cent) of them represent defective products. Thus, the random numbers 00
to 89 are assigned to variables that represent non-defective products and 90 to 100 are assigned to variables
that represent defective products.
If we choose a set of 2-digit random numbers in the range 00 to 99 to represent a packet of 6 products
as shown below, then we would expect that 90 per cent of the time they would fall in the range 00 to 89.

Sample Number Random Number

A 86 02 22 57 51 68
B 39 77 32 77 09 79
C 28 06 24 25 93 22
D 97 66 63 99 61 80
E 69 30 16 09 05 53
F 33 63 99 19 87 26
G 87 14 77 43 96 43
H 99 53 93 61 28 52
I 93 86 52 77 65 15
J 18 46 23 34 25 85

It may be noted that out of ten simulated samples 6 contain one or more defectives and 4 contain no defectives.
Thus, the expected percentage of non-defective products is 40 per cent. However, theoretically the probability
that a packet of 6 products containing no defective product is (0.9)6 = 0.53144 = 53.14%.
 Simulation 681

Example 19.2 A bakery keeps stock of a popular brand of cake. Previous experience shows the daily
demand pattern for the item with associated probabilities, as given below:
Daily demand (number) : 0 10 20 30 40 50
Probability : 0.01 0.20 0.15 0.50 0.12 0.02
Use the following sequence of random numbers to simulate the demand for next 10 days.
Random numbers: 25, 39, 65, 76, 12, 05, 73, 89, 19, 49.
Also estimate the daily average demand for the cakes on the basis of the simulated data.
Solution Using the daily demand distribution, we first obtain a probability distribution as shown in Table 19.3.

Daily Demand Probability Cumulative Random Number

Probability Intervals

0 0.01 0.01 00
10 0.20 0.21 01–20
20 0.15 0.36 21–35
30 0.50 0.86 36–85 Table 19.3
40 0.12 0.98 86–97 Daily Demand
50 0.02 1.00 98–99 Distribution

Next to conduct the simulation experiment for demand take a sample of 10 random numbers from a
table of random numbers, which represent the sequence of 10 samples. Each random sample number
represents a sample of demand.
The simulation calculations for a period of 10 days are given in Table 19.4.

Days Random Number Simulated

Demand

1 40 30 ← because random number 40 falls in the interval 36–85

2 19 10 ← because random number 19 falls in the interval 01-20
3 87 40 and so on
4 83 30
5 73 30
6 84 30
7 29 20
8 09 10
9 02 10
10 20 10
Table 19.4
Total = 220 Simulation
Expected demand = 220/10 = 22 units per day Experiments

Example 19.3 A company manufactures around 200 mopeds. Depending upon the availability of raw
materials and other conditions, the daily production has been varying from 196 mopeds to 204 mopeds,
whose probability distribution is as given below:
Production/day : 196 197 198 199 200 201 202 203 204
Probability : 0.05 0.09 0.12 0.14 0.20 0.15 0.11 0.08 0.06
The finished mopeds are transported in a specially designed three-storied lorry that can accommodate
only 200 mopeds. Using the following 15 random numbers: 82, 89, 78, 24, 53, 61, 18, 45, 23, 50, 77, 27, 54
and 10, simulate the mopeds waiting in the factory?
(a) What will be the average number of mopeds waiting in the factory?
(b) What will be the number of empty spaces in the lorry? [PT Univ., BTech, 2001]
682 Operations Research: Theory and Applications

Solution (a) Using production per day distribution, the daily production distribution is shown in Table 19.5.
Production/day Probability Cumulative Random Number
Probability Intervals

196 0.05 0.05 00 – 04

197 0.09 0.14 05 – 13
198 0.l2 0.26 14 – 25
199 0.14 0.40 26 – 39
200 0.20 0.60 40 – 59
201 0.15 0.75 60 – 74
Table 19.5 202 0.11 0.86 75 – 85
Daily Production 203 0.08 0.94 86 – 93
Schedule 204 0.06 1.00 94 – 99

Based on the given 15 random numbers, simulation experiment of the production per day is show in Table 19.6.
Days Random Production Number of Mopeds Empty Space
Number per Day Waiting in the Lorry

1 82 202 2 —
2 89 203 3 —
3 78 202 2 —
4 24 198 — 2
5 53 200 0 —
6 61 201 1 —
7 18 198 — 2
8 45 200 — —
9 04 196 — 4
10 23 198 — 2
11 50 200 — —
Table 19.6 12 77 202 2 —
Result of 13 27 199 1 —
Simulation 14 54 200 — —
Experiment 15 10 197 — 3

1
Average number of mopeds waiting in the factory = [2 + 3 + 2 + 1 + 2 + 1] = 1 moped (approx..)
15
13
Average number of empty spaces in the lorry = = 0.86
15
Example 19.4 A book store wishes to carry a particular book in stock. The demand of the book is not
certain and there is a lead time of 2 days for stock replenishment. The probabilities of demand are given below:
Demand (units/day) : 0 1 2 3 4
Probability : 0.05 0.10 0.30 0.45 0.10
Each time an order is placed, the store incurs an ordering cost of Rs 10 per order. The store also incurs
a carrying cost of Re 0.5 per book per day. The inventory carrying cost is calculated on the basis of stock
at the end of each day. The manager of the book store wishes to compare two options for his inventory
decision.
A : Order 5 books when the present inventory plus any outstanding order falls below 8 books.
B : Order 8 books when the present inventory plus any outstanding order falls below 8 books.
Currently (beginning of 1st day) the store has a stock of 8 books plus 6 books ordered two days ago
and are expected to arrive the next day. Carryout simulation run for 10 days to recommend an appropriate
option. You may use random numbers in the sequences, using the first number for day one.
89, 34, 78, 63, 61, 81, 39, 16, 13, 73 [AMIE, 2005]

Solution Using the daily demand distribution, we obtain a probability distribution, as shown in Table
19.7.
 Simulation 683

Daily Demand Probability Cumulative Random Number

Probability Intervals

0 0.05 0.05 00–04

1 0.10 0.15 05–14
2 0.30 0.45 15–44
Table 19.7
3 0.45 0.90 45–89 Daily Demand
4 0.10 1.00 90–99 Distribution

The stock in hand is of 8 books and stock on order is 5 books (expected next day).

Random Daily Opening Stock Receipt Closing Order Closing

Number Demand in Hand Stock in Hand Quantity Stock

89 3 8 – 8 – 3= 5 5 5
34 2 5 5 5 + 5 – 2= 8 – 8
78 3 8 – 8 – 3= 5 5 5
63 3 5 5 10 – 3 = 7 – 7
61 3 7 – 7– 3 = 4 5 4
81 3 4 5 9 – 3= 6 – 6
39 2 6 – 6 – 2= 4 5 4
16 2 4 5 9 – 2= 7 – 7
13 1 7 – 7 – 1= 6 – 6 Table 19.8
73 3 6 – 6 – 3= 3 – 3 Optimal A

55

Since 5 books have been ordered four times as shown in Table 19.8, therefore, the total ordering cost
is Rs (4 × 10) = Rs 40.
Closing stock of 10 days is of 55 books. Therefore, the holding cost at the rate of Re 0.5 per book
per day is Rs 55 × 0.5 = Rs 27.5
Total cost for 10 days = Ordering cost + Holding cost = Rs 40 + 27.5 = Rs 67.5

Random Demand Opening Stock Receipt Closing Order Closing

Number Daily in Hand Stock in Hand Quantity Stock

89 3 8 – 8 – 3 = 5 8 5
34 2 5 8 8 + 5 – 2 = 11 – 11
78 3 11 – 11 – 3 = 8 – 8
63 3 8 – 8 – 3 = 5 8 5
61 3 5 8 13 – 3 = 10 – 10
81 3 10 – 10 – 3 = 7 – 7
39 2 7 – 7 – 2 = 5 8 5
16 2 5 8 13 – 2 = 11 – 11
13 1 11 – 11 – 1 = 10 – 10
73 3 10 – 10 – 3 = 7 – 7 Table 19.9
Optimal B
71

Eight books have been ordered three times, as shown in Table 19.9, when the inventory of books
at the beginning of the day plus outstanding orders is less than 8. Therefore, the total ordering cost is:
Rs (3 × 10) = Rs 30.
Closing stock of 10 days is of 71 books. Therefore, the holding cost, Re 0.5 per book per day is
Rs 71 × 0.5 = Rs 35.5
The total cost for 10 days = Rs (30 + 35.5) = Rs 65.5. Since option B has a lower total cost than option
A, therefore, the manager should choose option B.
684 Operations Research: Theory and Applications

Example 19.5 A company trading in motor vehicle spare parts wishes to determine the levels of stock
it should carry for the items in its range. The demand is not certain and there is a lead time for stock
replenishment. For an item A, the following information is obtained:
Demand (units/day) : 3 4 5 6 7
Probability : 0.10 0.20 0.30 0.30 0.10
Carrying cost (per unit/day) : Rs 2
Ordering cost (per order) : Rs 50
Lead time for replenishment : 3 days
Stock on hand at the beginning of the simulation exercise was 20 units.
Carry out a simulation run over a period of 10 days with the objective of evaluating the inventory rule:
Order 15 units when present inventory plus any outstanding order falls below 15 units.
You may use random numbers in the sequence of: 0, 9, 1, 1, 5, 1, 8, 6, 3, 5, 7, 1, 2, 9, using
the first number for day one. Your calculation should include the total cost of operating this inventory rule
for 10 days. [AMIE, 2004]

Solution Let us begin simulation by assuming that:

(i) Orders are placed at the end of the day and received after 3 days, at the end of a day.
(ii) Back orders are accumulated in case of short supply and are supplied when stock is available.
The cumulative probability distribution and the random number range for daily demand is shown in
Table 19.10.

Daily Probability Cumulative Random Number

Demand Probability Intervals

3 0.10 0.10 00
4 0.20 0.30 01–02
Table 19.10 5 0.30 0.60 03–05
Daily Demand 6 0.30 0.90 06–08
Distribution 7 0.10 1.00 09

The results of the simulation experiment conducted are shown in Table 19.11.
Days Opening Random Resulting Closing Order Order Average Stock
Stock Number Demand Stock Placed Delivered in the Evening

1 20 0 3 17 – – 18.5
2 17 9 7 10 15 – 13.5
3 10 1 4 6 – – 8
4 6 1 4 2 – – 4
5 2 5 5 0 (– 3)* 15 15 1
6 12 1 4 8 – – 10
7 8 8 6 2 – – 6
Table 19.11 8 2 6 6 0 ( – 4)* 15 15 1
Simulation 9 11 3 5 6 – – 8.5
Experiments 10 6 5 5 1 – – 3.5

* Negative figures indicate back orders.

Average ending stock = 78/10 = 7.8 units/day
Daily ordering cost = (Cost of placing one order) × (Number of orders placed per day)
= 50 × 3 = Rs l50
Daily carrying cost = (Cost of carrying one unit for one day) × (Average ending stock)
= 2 × 7.8 = Rs l5.60
Total daily inventory cost = Daily ordering cost + Daily carrying cost = 150 + 15.60 = Rs 165.60.
Example 19.6 The manager of a warehouse is interested in designing an inventory control system for
one of the products in stock. The demand for the product comes from numerous retail outlets and the
orders arrive on a weekly basis. The warehouse receives its stock from a factory but the lead time is not
 Simulation 685

constant. The manager wants to determine the best time to release orders to the factory so that stockouts
are minimized, yet the inventory holding costs are at acceptable levels. Any order from retailers, not supplied
on a given day, constitute lost demand. Based on a sampling study, the following data are available:
Demand per Week Probability Lead Time Probability
(in thousand)

0 0.20 2 0.30
1 0.40 3 0.40
2 0.30 4 0.30
3 0.10
The manager of the warehouse has determined the following cost parameters: Ordering cost (C0) per
order equals Rs 50, carrying cost (Ch) equals Rs 2 per thousand units per week, and shortage cost (Cs)
equals Rs 10 per thousand units.
The objective of inventory analysis is to determine the optimal size of an order and the best time to
place an order. The following ordering policy has been suggested.
Policy : Whenever the inventory level becomes less than or equal to 2,000 units (reorder level), an order
equal to the difference between current inventory balance and the specified maximum replenish-
ment level, is equal to 4,000 units, is placed.
Simulate the policy for a week’s period assuming that the (i) the beginning inventory is 3,000 units,
(ii) no back orders are permitted, (iii) each order is placed at the beginning of the week, as soon as the
inventory level is less than or equal to the reorder level, and (iv) the replenishment orders are received at
the beginning of the week. [AMIE, 2005]

Solution Using weekly demand and lead time distributions, assign an appropriate set of random numbers
to represent value (range) of variables as shown in Tables 19.12 and 19.13, respectively.

Weekly Demand Probability Cumulative Probability Random Number

(in thousand) Interval
Table 19.12
0 0.20 0.20 00–19 Probabilities and
1 0.40 0.60 20–59 Random number
2 0.30 0.90 60–89 Interval for
3 0.10 1.00 90–99 Weekly Demand

Lead Time Probability Cumulative Probability Random Number

(weeks) Interval Table 19.13
Probabilities and
2 0.30 0.30 00–29 Random number
3 0.40 0.70 30–69 Interval for Lead
4 0.30 1.00 70–99 Time

The simulation experiment conducted for a 10 week period is shown in Table 19.14. The simulation process
begins with an inventory level of 3,000 units. The following four steps occur in the simulation process:
1. Begin each simulation week by checking whether any order has just arrived. If it has, increase the
beginning (current) stock (inventory) by the quantity received.
2. Generate a weekly demand from the demand probability distribution in Table 19.12 by selection of a random
number. This random number is recorded in column 4. The demand simulated is recorded in column 5.
The random number 31 generates a demand of 1,000 units when it is subtracted from the initial inventory
level value of 3,000 units. It yields an ending inventory of 2,000 units at the end of the first week.
3. Compute the ending inventory every week and record it in column 7.
Ending inventory = Beginning inventory – Demand = 3,000 – 1,000 = 2,000
If on hand inventory is not sufficient to meet the week’s demand, then record the number of units
short in column 6.
4. Determine whether the week’s ending inventory has reached the reorder level. If it has, and if there
is no outstanding order (back orders), then place an order.
Since the ending inventory of 2,000 units is equal to the reorder level, therefore, an order for
4,000 – 2,000 = 2,000 units is placed.
686 Operations Research: Theory and Applications

5. The lead time for the new order is simulated by first choosing a random number and recording it in
column 8. Finally, this random number is converted into a lead time (column 9) by using the lead time
distribution in Table 19.10.
The random number 29 corresponds to a lead time of 2 weeks, with 2,000 units to be held (carried) in
stock. Therefore, the holding cost of Rs 4 is paid and since there were no shortages, there is no shortage
cost. Summing these cost yields a total inventory cost (column 10) for week one of Rs 54.
The same step-by-step process is repeated for the remaining 10 weeks of the simulation experiment.
Analysis of Inventory Cost
1,000 total unit
Average ending inventory = = 100 units per week.
10 weeks
2 orders
Average number of orders placed = = 0.2 order per week.
10 weeks
7,000
Average number of lost sales = = 7 units per week.
1,000
Total average inventory cost = Ordering cost + Holding cost + Shortage cost
= (Cost of placing one order) × (Number of orders placed per week)
= + (Cost of holding one unit for one week) × (Average ending
inventory) + (Cost per lost sale) × (Average number of lost sales per
week)
100 16 70
= + + = 10 + 1.6 + 7 = Rs 18.6
10 10 10
Maximum Inventory Level = 4,000 units Reorder Level = 2,000 units

Week Order Beginning Random Demand Ending Quantity Random Lead Total Cost (TC)
Receipt Inventory Number Inventory Ordered Number Time C0 + Ch + Cs = TC (Rs)

1 0 3,000 31 1,000 2,000 2,000 29 2 50 4 – = 54

2 0 2,000 70 2,000 0 0 – – – – – –
3 0 0 53 1,000 (– 1,000) 0 – – 0 0 10 = 10
4 2,000 2,000 86 2,000 0 4,000 83 4 50 – – = 50
5 0 0 32 1,000 (– 1,000) 0 10 = 10
6 0 0 78 2,000 (– 2,000) 0 20 = 20
7 0 0 26 1,000 (– 1,000) 0 10 = 10
8 0 0 64 2,000 (– 2,000) 0 20 = 20
Table 19.14 9 4,000 4,000 45 1,000 3,000 0 6 – = 06
Inventory
10 0 3,000 12 0 3,000 0 6 – = 06
Simulation
Experiments Total 1,000 100 16 70

The negative figures in Table 19.14 enclosed in brackets indicate loss of sales.

19.8 SIMULATION OF QUEUING PROBLEMS

Example 19.7 A dentist schedules all his patients for 30-minute appointments. Some of the patients take
more 30 minutes some less, depending on the type of dental work to be done. The following summary shows
the various categories of work, their probabilities and time actually needed to complete the work:

Category of Time Required Probability

Service (minutes) of Category

Filling 45 0.40
Crown 60 0.15
Cleaning 15 0.15
Extraction 45 0.10
Checkup 15 0.20

Simulate the dentist’s clinic for four hours and determine the average waiting time for the patients as
well as the idleness of the doctor. Assume that all the patients show up at the clinic at exactly their
 Simulation 687

scheduled arrival time starting at 8.00 a.m. Use the following random numbers for handling the above
problem: 40 82 11 34 25 66 17 79 [AMIE, 2005]

Solution The cumulative probability distribution and random number interval for service time are shown
in Table 19.15.

Category Service Time Required Probability Cumulative Random Number

of Service (minutes) Probability Interval

Filling 45 0.40 0.40 00–39

Crown 60 0.15 0.55 40–54
Cleaning 15 0.15 0.70 55–69
Extraction 45 0.10 0.80 70–79
Checkup 15 0.20 1.00 80–99 Table 19.15

The various parameters of a queuing system such as arrival pattern of customers, service time, waiting
time, in the context of the given problem, are shown in Tables 19.16 to 19.18.
Patient Scheduled Random Category of Service Time
Number Arrival Number Service (minutes)

1 8.00 40 Crown 60
2 8.30 82 Checkup 15
3 9.00 11 Filling 45
4 9.30 34 Filling 45
5 10.00 25 Filling 45 Table 19.16
6 10.30 66 Cleaning 15 Arrival Pattern
7 11.00 17 Filling 45 and Nature of
8 11.30 79 Extraction 45 Service

Time Event Patient Number Waiting

(Patient Number) (Time to Exit) (Patient Number)

8.00 1 arrive 1 (60) –

8.30 2 arrive 1 (30) 2
9.00 1 departs; 3 arrive 2 (15) 3
9.15 2 depart 3 (45) –
9.30 4 arrive 3 (30) 4
10.00 3 depart; 5 arrive 4 (45) 5
10.30 6 arrive 4 (15) 5, 6 Table 19.17
10.45 4 depart 5 (45) 6 Computation of
11.00 7 arrive 5 (30) 6, 7 Arrivals,
11.30 5 depart; 8 arrive 6 (15) 7, 8 Departures and
11.45 6 depart 7 (45) 8 Waiting of
12.00 End 7 (30) 8 Patients

The dentist was not idle even once during the entire simulated period. The waiting times for the patients
were as follows:
Patient Arrival Time Service Starts at Waiting Time (minutes)
1 8.00 8.00 0
2 8.30 9.00 30
3 9.00 9.15 15
4 9.30 10.00 30
5 10.00 10.45 45
6 10.30 11.30 60 Table 19.18
7 11.00 11.45 45 Computation of
8 11.30 12.30 60 Average Waiting
280 Time

The average waiting time = 280/8 = 35 minutes.

688 Operations Research: Theory and Applications

Example 19.8 The management of ABC company is considering the question of marketing a new product.
The fixed cost required in the project is Rs 4,000. Three factors are uncertain, viz., the selling price, variable cost
and the annual sales volume. The product has a life of only one year. The management has the data on these
three factors as under:

Selling Price Probability Variable Cost Probability Sales Volume Probability

(Rs) (Rs) (Units)
3 0.2 1 0.3 2,000 0.3
4 0.5 2 0.6 3,000 0.3
5 0.3 3 0.1 5,000 0.4

Considering the following sequence of thirty random numbers: 81, 32, 60, 04, 46, 31, 67, 25, 24, 10, 40, 02, 39,
68, 08, 59, 66, 90, 12, 64, 79, 31, 86, 68, 82, 89, 25, 11, 98, 16.
Using the sequence (First 3 random numbers for the first trial, etc.) simulate the average profit for the above
project on the basis of 10 trails.
Solution The cumulative probability distribution and random number interval for selling price, variable cost
and sales volume are shown below:

Selling Price (Rs) Probability Cumulative Random Numbers

Probabilities Interval
3 0.2 0.2 00—19
4 0.5 0.7 20—69
5 0.3 1.0 70—99
Variable cost (Rs)
1 0.3 0.3 00—29
2 0.6 0.9 30—89
3 0.1 1.0 90—99
Sales volumes (Units)
2,000 0.3 0.3 00—29
3,000 0.3 0.6 30—59
5,000 0.4 1.0 60—99

The simulation experiment sheet for finding average profit is shown in Table 19.19.

Number of Random Selling Random Variable Random Sales Volume

Trials Number Price (Rs) Number Cost (Rs) Number (‘000 units)

1 81 5 32 2 60 5
2 04 3 46 2 31 3
3 67 4 25 1 24 2
4 10 3 40 2 02 2
5 39 4 68 2 08 2
6 59 4 66 2 90 5
7 12 3 64 2 79 5
Table 19.19 8 31 4 86 2 68 5
Simulation 9 82 5 89 2 25 2
Experiment Sheet 10 11 3 98 3 16 2
 Simulation 689

Trial Number Profit = (Selling Price – Variable Cost) × Sales Volume – Fixed Cost
1 (5 – 2) × 5,000 – 4,000 = 11,000
2 (3 – 2) × 3,000 – 4,000 = (– 1000)
3 (4 – 1) × 2,000 – 4,000 = 2,000
4 (3 – 2) × 2,000 – 4,000 = (– 2,000)
5 (4 – 2) × 2,000 – 4,000 = 0
6 (4 – 2) × 5,000 – 4,000 = 6,000
7 (3 – 2) × 5,000 – 4,000 = 1,000
8 (4 – 2) × 5,000 – 4,000 = 6,000
9 (5 – 2) × 2,000 – 4,000 = 2,000 Table 19.20
10 (3 – 3) × 2,000 – 4,000 = (– 4,000) Simulated Profit
21,000 in 10 Trials

Average profit per trial = 21,000/10 = Rs 21,00

Example 19.9 A firm has a single channel service station with the following arrival and service time
probability distributions:
Interarrival Time Probability Service Time Probability
(minutes) (minutes)

10 0.10 15 0.08
15 0.25 10 0.14
20 0.30 15 0.18
25 0.25 20 0.24
30 0.10 25 0.22
30 0.14

The customer’s arrival at the service station is a random phenomenon and the time between the arrivals
varies from 10 to 30 minutes. The service time varies from 5 minutes to 30 minutes. The queuing process
begins at 10 a.m. and proceeds for nearly 8 hours. An arrival immediately, goes to the service facility if it
is free. Otherwise it waits in a queue. The queue discipline is first-come first-served.
If the attendant’s wages are Rs 10 per hour and the customer’s waiting time costs Rs 15 per hour, then
would it be an economical proposition to engage a second attendant? Answer using Monte Carlo simulation
technique.
Solution The cumulative probability distributions and random number interval, both for interarrival time
and service time, are shown in Tables 19.21 and 19.22, respectively.

Interarrival Time Probability Cumulative Random Number

(minutes) Probability Interval

10 0.10 0.10 00–09

15 0.25 0.35 10–34
20 0.30 0.65 35–64
25 0.25 0.90 65–89
30 0.10 1.00 90–99 Table 19.21

Interarrival Time Probability Cumulative Random Number

(minutes) Probability Interval
5 0.08 0.08 00–07
10 0.14 0.22 08–21
15 0.18 0.40 22–39
20 0.24 0.64 40–63
25 0.22 0.86 64–85
30 0.14 1.00 86–99 Table 19.22
690 Operations Research: Theory and Applications

The simulation worksheet developed to the given problem is shown in Table 19.23.

Arrival Random Interarrival Arrival Service Waiting Random Service Exit Time in
Number Number Time (min.) Time (min.) Starts (min) Time Number Time Time System
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) =
(6) + (8)

1 20 15 10.15 10.15 0 26 15 30 15
2 73 25 10.40 10.40 0 43 20 60 20
3 30 15 10.55 11.00 5 98 30 90 35
4 99 30 11.25 11.30 5 87 30 120 35
5 66 25 11.45 12.00 15 58 20 140 35
6 83 25 12.10 12.20 10 90 30 170 40
7 32 15 12.25 1.05 35 84 25 195 60
8 75 25 12.50 1.30 40 60 20 215 60
9 04 10 1.00 1.50 50 08 10 225 60
10 15 15 1.15 2.00 45 50 20 245 65
Table 19.23 11 29 15 1.30 2.20 50 37 15 260 65
Single Server 12 62 20 1.50 2.35 45 42 20 280 65
Queuing 13 37 20 2.10 2.55 45 28 15 295 60
Simulation for 15 14 68 25 2.35 3.10 35 84 25 320 60
Arrivals
15 94 30 3.05 3.35 30 65 25 345 55

From the 15 samples of waiting time, 225 minutes, and the time spent, 545 minutes, by the customer
in the system, we compute all average waiting time in the system and average service time as follows:
Average waiting time = 380/15 = 25.3 minutes.
Average service time = 545/15 = 36.33 minutes
Thus, the average cost of waiting and service is given by:
Cost of waiting = 15 × (25.30/60) = Rs 6.32 per hour
Cost of service = 10 × (36.33/60) = Rs 6.05 per hour
Since the average cost of service, per hour, is less than the average cost of waiting per hour, therefore
second attendant may be hired.
Example 19.10 Observations of past data show the following patterns in respect of interarrival
durations and service durations in a single channel queuing system. Using the random number table below,
simulate the queue behaviour for a period of 60 minutes and estimate the probability of the service being
idle and the mean time spent by a customer waiting to be served.

Interarrival Time Service Time

Minutes Probability Minutes Probability

2 0.15 1 0.10
4 0.23 3 0.22
6 0.35 5 0.35
8 0.17 7 0.23
10 0.10 9 0.10

Random numbers (start at North-West corner and proceed along the row)

93 14 72 10 21
81 87 90 38 10
29 17 11 68 99
51 40 30 52 71

Solution The cumulative probability distributions and random number interval for interarrival time and
service time are shown in Table 19.24.
 Simulation 691

Arrival Time Cumulative Random Service Time Cumulative Random

Minutes Probability Probability Number Minutes Probability Probability Number
Interval Interval

2 0.15 0.15 00–14 1 0.10 0.10 00–09

4 0.23 0.38 15–37 3 0.22 0.32 10–31
6 0.35 0.73 38–72 5 0.35 0.67 32–66
8 0.17 0.90 73–89 7 0.23 0.90 67–89
10 0.10 1.00 90–99 9 0.10 1.00 90–99 Table 19.24

The simulation worksheet developed for the given problem is shown in Table 19.25.

Random Inter- Arrival Service Random Service Service Waiting Time Line
Number arrival Time Time Starts Number Time Ends Attendant Customer Length
(1) (min.) (min.) (min.) (2) (min.) (min.) (min.) (min.)

93 10 9.10 9.10 71 7 9.17 10 – –

14 2 9.12 9.17 63 5 9.22 – 5 1
72 6 9.18 9.22 14 3 9.25 – 4 1
10 2 9.20 9.25 53 5 9.30 – 5 1
21 4 9.24 9.30 64 5 9.35 – 6 1
81 8 9.32 9.35 42 5 9.40 – 3 1
87 8 9.40 9.40 07 1 9.41 – – –
90 10 9.50 9.50 54 5 9.55 9 – –
38 6 9.56 9.56 66 5 10.01 1 – –
Total 56 41 20 23 5 Table 19.25

(i) Average queue length = 5/9 = 0.56 = 1 customer (approx.).

(ii) Average waiting time of customer before service = 23/9 = 2.56 minutes.
(iii) Average service idle time = 20/9 = 2.22 minutes.
(iv) Average service time = 41/9 = 4.56 minutes.
(v) Time a customer spends in the system = (4.56 + 2.56) = 7.12 minutes.
(vi) Percentage of service idle time = 20/(20 + 41) = 0.33.

19.9 SIMULATION OF INVESTMENT PROBLEMS

Example 19.11 The Investment Corporation wants to study the investment projects based on three
factors, namely, market demand in units price per unit minus cost per unit, and investment required. These
factors are believed to be independent of each other. In analyzing a new consumer product, the Corporation
estimates the following probability distributions:

Annual Demand Price minus Cost per Unit Investment Required

Units Probability Rs Probability Rs Probability

20,000 0.05 3.00 0.10 17,50,000 0.25

25,000 0.10 5.00 0.20 20,00,000 0.50
30,000 0.20 7.00 0.40 25,00,000 0.25
35,000 0.30 9.00 0.20
40,000 0.20 10.00 0.10
45,000 0.10
50,000 0.05

Using the simulation process, repeat the trial 10 times, compute the return on investment for each trial
taking these three factors into account. What is the most likely return?
692 Operations Research: Theory and Applications

Solution The return per annum can be computed by the following expression
(Price – Cost) × Number of units demanded
Return (R) =
Investment
Developing a cumulative probability distribution, corresponding to each of the three factors, an
appropriate set of random numbers is assigned to represent each of the three factors, as shown in Tables
19.26, 19.27 and 19.28.

Annual Demand Probability Cumulative Probability Random Number

20,000 0.05 0.05 00–04

25,000 0.10 0.15 05–14
30,000 0.20 0.35 15–34
35,000 0.30 0.65 35–64
40,000 0.20 0.85 65–84
45,000 0.10 0.95 85–94
Table 19.26 50,000 0.05 1.00 95–99

Price minus Probability Cumulative Random Number

Cost per Unit Probability
3.00 0.10 0.10 00–09
5.00 0.20 0.30 10–19
7.00 0.40 0.70 20–69
9.00 0.20 0.90 70–89
Table 19.27 10.00 0.10 1.00 90–99

Investment Probability Cumulative Random Number

Required Probability

17,50,000 0.25 0.25 00–24

20,00,000 0.50 0.75 25–74
Table 19.28 25,00,000 0.25 1.00 75–99

The simulation worksheet is prepared for 10 trials. The simulated return (R) is also calculated by using
the formula for R, as stated before. The results of simulation are shown in Table 19.29.

Trials Random Simulated Random Simulated Random Simulated Simulated return

Number for Demand Number for Profit Number for Investment (%): Demand ×
Demand (’000) Profit (Price – Investment (’000) Profit per Unit
Cost) per Unit × 100
Investment

1 28 30 19 5.00 18 1,750 8.57

2 57 35 07 3.00 61 2,000 5.25
3 60 35 90 10.00 16 1,750 20.00
4 17 30 02 3.00 71 2,000 4.50
5 64 35 57 7.00 43 2,000 12.25
6 20 30 28 5.00 68 2,000 7.50
7 27 30 29 5.00 47 2,000 7.50
8 58 35 83 9.00 24 1,750 18.00
9 61 35 58 7.00 19 1,750 14.00
Table 19.29 10 30 30 41 7.00 97 2,500 8.40
 Simulation 693

As shown in Table 19.29, the highest likely return is 20 per cent, which corresponds to the annual
demand of 35,000 units yielding a profit of Rs 10 per unit and investment required is Rs 17,50,000.

19.10 SIMULATION OF MAINTENANCE PROBLEMS

Example 19.12 A plant has a large number of similar machines. The machine breakdown or failure is
random and independent.
The shift incharge of the plant collected the data about the various machines breakdown times, the repair
time required on hourly basis, and the record for the past 100 observations. This is shown below was:

Time Between Recorded Probability Repair Time Probability

Machine Breakdowns (hours) Required (hours)

0.5 0.05 1 0.28

1 0.06 2 0.52
1.5 0.16 3 0.20
2 0.33
2.5 0.21
3 0.19

For each hour that one machine is down due to being, or waiting to be, repaired, the plant loses
Rs 70 by way of lost production. A repairman is paid at Rs 20 per hour.
(a) Simulate this maintenance system for 15 breakdowns.
(b) How many repairmen should the plant hire for repair work.
Solution The random numbers coding for the hourly breakdowns and the repair times are shown in
Tables 19.30 and 19.31.

Time Between Probability Cumulative Random Number

Breakdowns (hours) Probability Range

0.5 0.05 0.05 00–04

1 0.06 0.11 05–10
1.5 0.16 0.27 11–26 Table 19.30
2 0.33 0.60 27–59 Random Number
2.5 0.21 0.81 60–80 coding for
3 0.19 1.00 81–99 Breakdowns

Repair Time Probability Cumulative Random Number

Required (hours) Probability Range
Table 19.31
1 0.28 0.28 00–27 Random Number
2 0.52 0.80 28–79 Coding for
3 0.20 1.00 80–99 Repairs

The simulation worksheet is shown in Table 19.32. It is assumed that the first day begins at midnight
(00.00 hours) and also that the repairman begins work at 00.00 hours. The first breakdown occurred at 2.30
a.m and the second occurred after 3 hours, at clock time of 5.30 a.m.
694 Operations Research: Theory and Applications

Breakdown Random Time Time of Repair Random Repair Repair Total Idle Waiting
Number Number for Between Break- Work Number for Time Work Time Time
Break- Break- down Begins Repair Required Ends at (hours) (hours)
downs downs at Time
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1 61 2.5 02.30 02.30 87 3 05.30 3.00 –
2 85 3 05.30 05.30 39 2 07.30 2.00 –
3 16 1.5 07.00 07.30 28 2 09.30 2.30 0.30
4 46 2 09.00 09.30 97 3 12.30 3.30 0.30
5 88 3 12.00 12.30 69 2 14.30 2.30 0.30
6 08 1 13.00 14.30 87 3 17.30 4.30 1.30
7 82 3 16.00 17.30 52 2 19.30 3.30 1.30
8 56 2 18.00 19.30 52 2 21.30 3.30 1.30
9 22 1.5 19.30 21.30 15 1 22.30 3.00 2.00
10 49 2 21.30 22.30 85 3 01.30 4.00 1.00
11 44 2 23.30 01.30 41 2 03.30 4.00 2.00
12 33 2 01.30 03.30 82 3 06.30 5.00 2.00
13 77 2.5 04.00 06.30 98 3 09.30 5.30 2.30
Table 19.32 14 87 3 07.00 09.30 99 3 12.30 5.30 2.30
Simulation 15 54 2 09.00 12.30 23 2 14.30 5.30 3.30
Worksheet 38.30 36 57.30 21.30

Total current maintenance cost = Idle time cost + Repairman’s wage

= + (Repair time + Waiting time) × Hourly rate + Total hours ×
= Hourly wages
= 57.30 × 70 + 38.30 × 20 = Rs 4,777
Maintenance Cost with Additional Repairmen
If the plant hires two more repairmen, then no machine will wait to be repaired. Thus, the total idle time
would only be the repairing time of 36.00 hours. Therefore,
Total cost = 36 × 70 + (38.30 × 2) × 20 = Rs 4,052
This shows that hiring more than two repairmen would only increase the total maintenance cost. Hence,
the plant should ideally hire one additional repairman.
Example 19.13 Two persons X and Y work on a two-station assembly line. The distributions of activity
at their stations are
Time (in seconds) Time Frequency for X Time Frequency for Y

10 4 2
20 7 3
30 10 6
40 15 8
50 35 12
60 18 9
70 8 7
80 3 3

(a) Simulate operation of the line for eight items.

(b) Assuming Y must wait until X completes the first item before starting work, will he has to wait to
process any of the other seven items? What is the average waiting time of items. Use the following
random numbers :
For X : 83 70 06 12 59 46 54 and 04.
For Y : 51 99 84 81 15 36 12 and 54.
(c) Determine the inventory of items between the two stations.
(d) What is the average production rate? [Kuru. Univ., BE (Mech.), 2000]
 Simulation 695

Solution (a) The cumulative frequency distribution for X is shown in Table 19.33

Time (in seconds) Time Frequency Cumulative Random Numbers

for X Frequency Range

10 4 4 00–03
20 7 11 04–10
30 10 21 11–20
40 15 36 21–35
50 35 71 36–70 Table 19.33
60 18 89 71–88 Cumulative
70 8 97 89–96 Frequency
80 3 100 97–99 Distribution for X

Random Number : 83 70 06 12 59 46 54 04
Time Taken by X : 60 50 20 30 50 50 50 20
Thus the eight times for X are: 60, 50, 20, 30, 50, 50, 50 and 20 seconds respectively. Likewise, the eight
times for Y are derived from cumulative distribution shown in Table 19.34

Time (in seconds) Time Frequency Cumulative Random Numbers

for Y Frequency Range
(a) (b) = 2 × (a)

10 2 2 4 00 – 03
20 3 5 10 04 – 09
30 6 11 22 10 – 21
40 8 19 38 22 – 37
50 12 31 62 38 – 61 Table 19.34
60 9 40 80 62 – 79 Cumulative
70 7 47 94 80 – 93 Frequency
80 3 50 100 94 – 99 Distribution for Y

Thus the eight times for Y are: 62, 100, 94, 94, 22, 38, 22 and 62 seconds respectively. The cumulative
frequency has been multiplied by 2 to make it 100.
Random Number : 51 99 84 81 15 36 12 54
Time Taken by Y : 62 100 94 94 22 38 22 62
(b) The times for persons X and Y are used to calculate the waiting time as shown in Table 19.35

Person X Person Y Waiting Time Waiting Time

Time in Time out Time in Time out on the Part of Y on the Part of X

0 60 60 112 60 —
60 110 112 212 — —
110 130 212 306 — 82
130 160 306 416 — 146
160 210 328 432 — 118
210 260 366 470 — 106 Table 19.35
260 310 388 492 — 78 Waiting Time for
310 330 450 554 — 120 Part X

Thus person X will not have to wait for the remaining seven items.
0 + 0 + 82 + 146 + 118 + 106 + 178 + 120 530
Average waiting time of items = = = 67 seconds (approx.)
8 8
(c) In all there are 6 items waiting between the two stations.
(d) Total time taken to process 8 items = 554 seconds = 9 minutes.
Average production rate = 9/8 = 1 item/minute (approx.)
696 Operations Research: Theory and Applications

Example 19.14 Popa Ltd trade in a perishable commodity. Each day Popa Ltd. receives supplies of the
goods from a wholesaler but the quantity supplied is a random variable, as is the subsequent retail customer
demand for the commodity. Both supply and demand are expressed in batches of 50 units and over the past
working year (300 days), Popa Ltd. has kept records of supplies and demands. The results are given below:

Wholesaler Number of Days Customer’s Number of Days

Supplies Occurring Demand Occurring

50 60 50 60
100 90 100 60
150 90 150 150
200 60 200 30

Popa Ltd. buys the commodity at Rs 6 per unit and sells at Rs 10 per unit. At present unsold units
at the end of the day are worthless and there are no storage facilities. Popa Ltd. estimates that each unit
of unsatisfied demand on any day costs them Rs 2. Using the random numbers: 8, 4, 8, 0, 3, 3, 4, 7, 9, 6,
1 and 5
(a) simulate six days trading and estimate the annual profit.
(b) repeat the exercise to estimate the value of storage facilities. [ICWA, 2000 ]
Solution Wholesaler’s supplies are simulated for 6 days as shown in Table 19.36

Supplies Number of Total Number Number of Days Random

Days of Days Out of 10 Number Range

50 60 60 2 00 – 01
100 90 150 5 02 – 04
150 90 240 8 05 – 07
Table 19.36 200 60 300 10 08 – 09

Random Number : 8 4 8 0 3 3 4 7 9 6 1 5
Supplies : 200 100 200 50 100 100 100 150 200 150 50 150
Thus, wholesaler’s supplies in the next 6 days are: 200, 100, 200, 50, 100 and 100 units respectively.
Similarly, customer’s demand over the next 6 days is simulated as shown in Table 19.37.

Supplies Number Total Number Number of Days Random

of Days of days Out of 10 Number Range

50 60 60 2 00 – 01
100 60 120 4 02 – 03
150 150 270 9 04 – 08
Table 19.37 200 30 300 10 09

Customer’s demand in the next 6 days is: 150, 150, 200, 150, 50 and 150 units.
The shortage and net profit or loss in the 6 days is now calculated as shown in Table 19.38.

Days Supply Demand Shortage Net Profit

Unit Cost Unit Cost Unit Cost or Loss
(Rs) (Rs) (Rs) (Rs)

1 200 1,200 150 1,500 — — 1,500 – 1,200 = 30

2 100 600 150 1,000 50 100 1,000 – 600 – 100 = 300
3 200 1,200 200 2,000 — — 2,000 – 1,200 = 800
4 50 300 150 500 100 200 500 – 300 – 200 = 0
5 100 600 50 500 — — 500 – 600 = –100
Table 19.38 6 100 600 150 1,000 50 100 1000 – 600 – 100 = 300

(a) Net profit from Table 22.47 for 6 days = Rs 1,600

1, 600 × 300
Annual profit = = Rs 80,000
6
 Simulation 697

(b) The value of storage facilities has been shown in Table 19.39.

Days Supply Demand Storage

1 200 150 50
2 100 150 Nil Shortage cost of Rs 100 for 50 units could be avoided.
3 200 200 Nil
4 50 150 Nil
5 100 50 50
6 100 150 Nil Shortage cost of Rs 100 for 50 units could be avoided. Table 19.39

If the storage facilities, then Rs 200 of shortage cost could be avoided by storing 50 + 50 = 100 units.
These 100 units could be then sold, yielding a profit of Rs 4 × 100 = Rs 400. The value of storage facilities
= 200 + 400 = Rs 600

19.11 SIMULATION OF PERT PROBLEMS

Example 19.15 A project consists of eight activities A to H. The completion time for each activity is
a random variable. The data concerning probability distribution, along with completion times for each
activity, is as follows:

Activity Immediate Time (day)/Probability

Predecessor(s) 1 2 3 4 5 6 7 8 9

A – – – – 0.2 – 0.4 0.4 – –

B – – – – – – 0.5 – 0.5 –
C A – – 0.7 0.3 – – – – –
D B, C – – – – 0.9 – – 0.1 –
E A – – – – 0.2 – – – 0.8
F D, E – – – 0.6 0.4 – – – –
G E – – 0.4 0.4 – 0.2 – – –
H F – 0.4 – – – – 0.6 – –

(a) Draw the network diagram and identify the critical path using the expected activity times.
(b) Simulate the project to determine the activity times. Determine the critical path and project expected
completion time.
(c) Repeat the simulation four times and state the estimated duration of the project in each of the trials.
Solution (a) The network diagram based on the precedence relationships is shown in Fig. 19.2. The
expected completion time of each activity is obtained by using the formula:

Fig. 19.2
Network Diagram

Expected time = Σ (Activity time × Probability)

= 4 × 0.2 + 6 × 0.4 + 7 × 0.4 = 6 days (activity A)
The critical path of the project is: 1 – 2 – 3 – 4 – 5 – 6 – 7, with expected completion time of 23.6 days.
The random number coding for each of the activities expected time is shown in Table 19.40.
698 Operations Research: Theory and Applications

Activity Time Probability Cumulative Random Number

Probability Range

A 4 0.20 0.20 00–19

6 0.40 0.60 20–59
7 0.40 1.00 60–99
B 6 0.50 0.50 00–49
8 0.50 1.00 50–99
C 3 0.70 0.70 00–69
4 0.30 1.00 70–99
D 5 0.90 0.90 00–89
8 0.10 1.00 90–99
E 5 0.20 0.20 00–19
9 0.80 1.00 20–99
F 4 0.60 0.60 00–59
5 0.40 1.00 60–99
G 3 0.40 0.40 00–39
4 0.40 0.80 40–79
Table 19.40
Random Number 6 0.20 1.00 80–99
Coding for Activity H 2 0.40 0.40 00–39
Times 7 0.60 1.00 40–99

The simulation worksheet for four simulation runs is shown in Table 19.41. For each run the project
time is obtained as follows:
Total time = Highest times for activities A, B and C + Highest times for activities D and E +
= Highest times for activities F and G + Time for activity H.
Using the data given in Table 19.41, the simulation results that we have are shown in Table 19.42.

Run Activity Times (days)

A B C D E F G H
R. No. Time R. No. Time R. No. Time R. No. Time R. No. Time R. No. Time R. No. Time R. No. Time

1 22 6 17 6 68 3 65 5 84 9 68 5 95 6 23 2
2 92 7 35 6 61 3 09 5 43 9 95 5 06 3 87 7
3 02 4 22 6 57 3 51 5 58 9 24 4 82 6 03 2
4 47 6 19 6 36 3 27 5 59 9 46 4 13 3 79 7
5 93 7 37 6 66 3 85 5 52 9 05 4 30 3 62 7
Table 19.41
Simulation Total 30 30 15 25 45 22 21 25
Worksheet Average 6 6 3 5 9 4.4 4.2 5

Simulation Activity Time Project Duration Longest (Critical) Path

Run (days)

1 6 + 9 + 6 + 2 23 1 – 2 – 3 – 4 – 5 – 6 – 7
1 – 3 – 4 – 5 – 6 – 7
2 7 + 9 + 5 + 7 28 1 – 3 – 4 – 6 – 7
3 6 + 9 + 6 + 2 23 1 – 2 – 3 – 4 – 5 – 6 – 7
4 6 + 9 + 4 + 7 26 1 – 2 – 3 – 4 – 6 – 7
1 – 3 – 4 – 6 – 7
Table 19.42
Simulation 5 7 + 9 + 4 + 7 27 1 – 3 – 4 – 6 – 7
Results 127
 Simulation 699

Here it may be noted that simulated mean project completion time of 25.4 days is almost two days
longer than the 23.6 days completion time, indicated using expected values alone.

CONCEPTUAL QUESTIONS

1. Distinguish between solutions derived from simulation models control, where the demand is probabilistic and lead time is
and solutions derived from analytical models? random.
2. What are random numbers? Why are random numbers useful 11. Discuss the Monte Carlo method of solving a problem, illustrat-
in simulation models and in solutions derived from analytical ing it by outlining a procedure to solve a specified problem of
models? your choice.
3. (a) What are the advantages and limitations of simulation 12. Describe the kind of problems for which Monte Carlo will be an
models? [Delhi Univ., MBA, 2003] appropriate method of solution.
(b) What are the advantages and disadvantages of simulation 13. Explain what factors must be considered when designing a
over the use of analytical models? Is the use of computers simulation experiment.
in simulation absolutely essential? 14. Draw a flow chart to describe the simulation of a simple system.
4. What is Monte Carlo simulation? Describe the idea of experi- 15. State the considerations involved in trading-off costs, with
mentation (Random sampling) in simulation. reliability in designing a simulation experiment.
5. ‘Monte Carlo technique has been used to tackle a variety of
16. ‘. . . simulation is a quantitative technique developed for studying
problems involving stochastic situations and mathematical prob-
alternative courses of action by building a model of that system
lems which cannot be solved with mathematical techniques and
and then conducting a series of repeated trial and error experi-
where physical experimentation with the actual system is im-
ments to predict the behaviour of the system over a period of
practicable.’ Discuss. [AMIE, 2004 ]
time.’ Discuss.
6. ‘When it becomes difficult to use an optimization technique for 17. Why is a computer necessary in conducting a real-world
solving a problem, one has to resort to simulation technique.’ simulation?
Discuss. [Delhi Univ., MBA, 2003, AMIE, 2005 ] 18. Do you think the application of simulation will strongly increase in
7. ‘Simulation typically is nothing more or less than the technique the next ten years? Give reasons for your answer.
of performing sampling experiments on the model of the system.’ 19. What types of problems can be solved more easily by quanti-
Discuss. tative techniques other than simulation?
8. State two major reasons of using simulation. Explain the basic 20. Why would an analyst ever prefer a general purpose language
steps of Monte Carlo simulation. Briefly describe its application such as FORTRAN or BASIC in a simulation when there are
in Finance and Accounting. advantages of using special purpose languages such as GPSS
9. Define simulation. Why is simulation used? Give one application or SIMSCRIPT?
area when this technique is used in practice.
21. Explain the methods of gathering statistical observations in
10. Explain how simulation can be applied in the case of inventory simulation modelling. [AMIE, 2004 ]

SELF PRACTICE PROBLEMS

1. Bharat Transport Company is considering discontinuing its
Daily Demand Probability
leasing of 12 pickups and delivery trucks. If it does this, it will
have to buy 12 trucks right away and buy replacement trucks 25,001–45,000 0.30
in the future, one by one, as and when the old ones wear out.
45,000–55,000 0.30
The question that the Bharat Transport Company’s manager
wants to answer is how many trucks will they have to buy during 55,000–65,000 0.40
the next five years (including the initial 12 trucks) in order to
(a) What is the expected daily demand?
keep 12 trucks in operation all the time. At the moment, they are
(b) Construct a model that can be used to simulate the
not concerned with the fact that at the end of five years they
company’s daily receiving, storage and shipping activities.
will have 12 trucks in hand, some of which will be relatively new
and some of which will be old. The only question is how many 3. XYZ company operates an automatic car-wash facility in a city.
will have to be purchased so that they can plan their cash The manager is concerned about the long lines of cars that build
requirements accordingly. The following table shows a history of up while waiting for service. The service time for the system is
truck life: machine-paced and thus constant. The manager has the oppor-
tunity to decrease the service time by increasing the speed of
Truck life (months) : 12 15 18 21 24 27 30 the conveyor that pulls the cars through the system. Of course,
Percentage of trucks the quicker the pace, the lower is the quality of the car-wash.
which have worn out : 5 10 20 25 30 5 5 The manager wants to study the effects of setting the system
2. A gas transport company controls pipe-lines between several for a 2-minute car-wash. The following data on customer arrivals
natural gas fields and out of state distributors. The company has have been gathered.
a 1,00,000 unit storage capacity. Because of certain govern- Interarrival time (minutes) : 1 3 3 4 5
ment regulations, the company receives either 40,000 or 60,000 Number of occurrences : 136 34 102 51 17
units per day but the probability of receiving such quantity is not (a) Compare the traffic density for a service rate of 2 minutes
equal. The actual demand for natural gas is given by the per car.
following table: (b) Simulate the arrival of 20 customers. The doors open at
8.00 a.m.
700 Operations Research: Theory and Applications

Compute the average waiting time per customer, the aver- 7. A retail store distributes catalogues and takes orders by tele-
age time spent in the system, the percentage of time the system phone. Distributions for intervals between incoming calls and the
is ideal, and the maximum queue length. length of time required to complete each call are given below.
4. The customers of State Distribution Corporation send their own The store management has determined that the probability that
purchase orders. In the past, the arrival of these purchase a caller will have to wait for more than 10 seconds for a call to
orders, per day, has approximated a normal distribution with a be answered should not be more than 5 per cent. Use simulation
mean of 50 and a standard deviation of 6. In terms of the to determine how many sales representatives should be avail-
probability of occurrence, the following is being indicated: able to answer incoming calls.

Interval between Probability Length of Call Probability

Number of Probability of Incoming Calls (seconds)
Purchase Orders Occurrence (seconds)
26–32 00.5 10 0.08 60 0.07
32–38 02.0 12 0.11 65 0.12
38–44 13.0 14 0.14 70 0.18
44–50 36.0 16 0.16 75 0.16
50–56 33.0 18 0.14 80 0.15
56–62 13.0 20 0.12 85 0.12
62–68 02.0 22 0.08 90 0.08
68–74 00.5 24 0.07 95 0.06
26 0.04 100 0.06
Develop a Monte Carlo simulation for the number of pur- 28 0.04
chase orders per day to be expected for a particular month. If 30 0.02
the firm can purchase only 41 orders per day, how many days
8. A firm has single channel service station with the following arrival
in that month would the firm be behind schedule?
and service time probability distributions:
5. The materials manager of a firm wishes to determine the
expected (mean) demand for a particular item in stock during
Arrivals Probability Service Time Probability
the reorder lead time. This information is needed to determine
(min) (min)
how far in advance he should reorder, before the stock level is
reduced to zero. However, both the lead time (in days) and the 1.0 0.35 1.0 0.20
demand per day, for the item, are random variables, described 2.0 0.25 1.5 0.35
by the probability distribution given below: 3.0 0.20 2.0 0.25
4.0 0.12 2.5 0.15
Lead Time Probability of Demand/Day Probability 5.0 0.08 3.0 0.05
(days) Occurrence (units)
The customer’s arrival at the service station is a random
1 0.50 1 0.10 phenomenon and the time between the arrival varies from one
2 0.30 2 0.30 minutes to five minutes. The service time varies from one minute
3 0.20 3 0.40 to three minutes. The queuing process begins at 10.00 a.m. and
4 0.20 proceeds for nearly 2 hours. An arrival goes to the service
Manually simulate the problem for 30 reorders to estimate the facility immediately, if it is free, otherwise it waits in a queue. The
demand during lead time. [Delhi Univ., MBA, 1995, 2000 ] queue discipline is first-come first-served.
If the attendant’s wages are Rs 8 per hour and the customer’s
6. An automatic machinery company receives a different number of waiting time costs Rs 9 per hour, then would it be an economical
orders each day and the orders vary in terms of the time required proposition to engage a second attendant? Answer on the basis
to process them. The firm is interested in determining how many of Monte Carlo simulation technique.
machines it should have in the departments in order to minimize
9. A certain maintenance facility is responsible for the upkeep of
the combined cost of machine idle and order waiting time. The firm
five machines. The machines, which fail frequently, must be
knows, from past experience, the average number of orders per
repaired as soon as possible in order to maintain as high a
day and the average number of hours per order, which are as
productive capacity of the production system as possible. The
follows:
management is concerned about the average down time per
machine and is considering an increase in the capacity of the
Number of Probability Hours/Order Probability maintenance facility. From historical data, the following distribu-
Orders/Day tions have been developed:
0 0.10 5 0.05
1 0.15 10 0.05 Time between Probability Repair Time Probability
2 0.25 15 0.10 Breakdown
3 0.30 20 0.10 (days)
4 0.15 25 0.20
2 0.05 1 0.40
5 0.05 30 0.25
3 0.10 2 0.50
35 0.15
4 0.15 3 0.10
40 0.10
5 0.40
Cost/hour of idle machine time = Rs 4.00 Cost-hour for orders 6 0.20
waiting = Rs 6.00 7 0.10
Assuming 24 hours working in three shifts, solve the problem
Simulate the failure and repair of 10 machines. Begin by
using simulation.
determining the time of the first breakdown by each of the 5
 Simulation 701

machines. Sequence the machines through the repair facility on 13. A confectioner sells confectionery items. Past data of demand
a first-come first-served basis. If there is more than one per week (in hundred kilograms), with frequency, is given below:
machine waiting to be repaired, arbitrarily choose one to repair Demand/week : 05 10 15 20 25
the next. After a machine has been repaired, determine its next
time of breakdown and continue until you have repaired 10. Frequency : 211 8 21 5 3
[Delhi Univ., MBA, 2003] Using the following sequence of random numbers, generate the
demand for the next 10 weeks. Also find the average demand
10. A trader has studied his varying monthly sales and monthly
per week:
expenses (including the value of goods) and has arrived at the
35, 52, 90, 13, 23, 73, 34, 57, 35, 83, 94, 56, 67, 66, 60.
following empirical distributions:
14. An owner of a petrol pump with a single attendant wishes to
(a) The trader at the beginning of the year has Rs 2,000 in the
perform a simulation of his operations to see whether any
bank. Simulate his sales and expenses over a year (two
improvement is possible. He studied the system and found that
times). Assume that the trader can avail temporary over-
an average of 6 customers arrive for service with random arrival
draft facilities to cover any negative balance.
times and form a queue, and the attendant provides service for
(b) How much money does the trader have at the end of the
exactly 9 minutes. For simulating the arrival times of customers,
year?
he has selected 10 random numbers with expected length of
interval equal to one as: 3.62, 1.78, 1.84, 1.31, 1.27 0.14, 1.71,
Monthly Sales Probability Monthly Expenses Probability 0.77, 0.97, 1.32. Find:
(’000) Rs (’000) Rs
(a) The total idle time for the attendant
15 0.30 12 0.15 (b) Total waiting time for the customers; and
16 0.25 13 0.20 (c) Maximum queue length during this period.
17 0.15 14 0.25 If the service time is reduced to 6 minutes what is the quality
18 0.15 15 0.20 of the service?
19 0.10 16 0.15 15. A machine shop has 30 machines. The following is the distribu-
20 0.05 18 0.05 tion sample of 73 breakdown of machines:
Time between : 10 11 12 13 14 15 16 17 18 18 19
11. The management of a company is considering the problem of breakdowns (hrs)
marketing a new product. The investment or the fixed cost Frequency : 4 10 14 16 12 6 4 3 3 3 1
required in the project is Rs 25,000. There are three factors that (Total = 73)
are uncertain – selling price, variable cost and the annual sales A study of time required to repair the machines by one mechanic
volume. The product has a life of only one year. The manage- yields the following distribution:
ment has past data regarding the possible levels of these three Repair time (hrs) : 8 9 10 11 12 13 14 15 16 17 18
factors.
Frequency : 2 3 8 16 14 12 8 5 3 1 1
(Total = 73)
Unit Probability Unit Probability Sales Probability
(a) Convert the distribution to cumulative probability distribu-
Selling Variable Volume
Price (Rs) Cost (Rs) (units)
tions. (b) Using a simulated sample of 20, estimate the
average per cent machine waiting time and the average per
40 0.30 20 0.10 3,000 0.20 cent idle time of the mechanic.
50 0.50 30 0.60 4,000 0.40 16. A gas transport company controls pipelines between several
60 0.20 40 0.30 5,000 0.40 natural gas fields and out of state distributors. The company has
a 1,00,000 unit storage capacity. Because of federal regulations,
Using Monte Carlo simulation technique, determine the average the company receives either 40,000 or 60,000 units per day.
profit from the said investment on the basis of 20 trials. There is no equal probability of either quantity being shipped on
12. A company manufactures 200 motor cycles per day. Depending a given day. The actual demand for natural gas is given by the
upon the availability of raw materials and other conditions, the following table of relative frequencies:
daily production has been varying from 196 motor cycles to 204
motor cycles, whose probability distribution is as given below: Daily Demand Probability

Production/day Probability 25,001–45,000 0.30

45,000–55,000 0.30
196 0.05 55,000–65,000 0.40
197 0.09
198 0.12 (a) What is the expected daily demand?
199 0.14 (b) Construct a model that can be used to simulate the
200 0.20 company’s daily receiving, storage and shipping activities.
201 0.15 Attempt a simulation for 12 months to establish a cashflow
202 0.11 pattern using random numbers from the following table:
203 0.08 9,548 5,099 9,747 3,755 4,162
204 0.06 4,552 6,291 1,830 7,263 7,010
9,969 8,265 1,572 7,705 1,352
The motor cycles are transported in a specially designed three- 3,512 4,191 4,570 4,826 3,140
storied lorry that can accommodate only 200 motor cycles.
Using the following random numbers: 82, 89, 78, 24, 52, 53, 61, State assumptions.
18, 45, 04, 23, 50, 77, 27, 54, 10, simulate the process to find 17. The materials manager of a company is interested in determin-
out: ing the reorder point for an item, the pattern of demand for which
(a) The average number of motor cycles waiting in the factory? is given below:
(b) The average number of empty spaces on the lorry? No. of units per day : 0 1 2 3 4 5
No. of days on which
702 Operations Research: Theory and Applications

the demands occurred : 5 9 16 38 23 9 calculation should include the total cost of operating this inven-
Past experience indicates that there are fluctuations in the lead tory rule for 10 days.
time for procurement of items. The following data are available 19. The materials manager of a company is interested in deter-
from the records for the last 30 orders: mining the reorder point for an item, the pattern of demand for
which is as given below:
Lead time (weeks) : 1 2 3
No. of times the specified
No. of Units No. of Days on Which the
lead time occurred : 18 7 5
per Day Demand Occurred
The management policy is to ensure that the proportion of
stockouts should not exceed 5 per cent. Illustrate how the 0 5
simulation approach can be used to determine the reorder point. 1 9
(Analysis for 20 orders is adequate.) 2 16
18. A company trading in motor vehicle spares wishes to determine 3 38
the level of stock it should carry for the items in its range. 4 23
Demand is not certain and there is a lead time for stock 5 9
replenishment of one item N X. The following information is
obtained: Past experience indicates that there are fluctuations in the lead
time for procurement of the item. The following data are available
Demand (units/day) : 3 4 5 6 7
from records for the last 30 orders:
Probability : 0.10 0.20 0.30 0.30 0.10
Carrying cost (per unit per day) = 20 paise. Lead Time No. of Times the Specified
(weeks) Lead Time Occurred
Ordering cost (per order) = Rs 5
Lead time for replenishment = 3 days 1 181
2 7
Stock in hand at the beginning of the simulation exercise was
3 5
20 units.
You are required to carry out a simulation run over a period The management policy is to ensure that the proportion of
of ten days with the objective of evaluating the following inven- stockouts should not exceed 5 per cent. Illustrate how the
tory rule: Order 15 units when present inventory plus any simulation approach can be used to determine the reorder point.
outstanding order falls below 15 units. (Analysis for 20 orders is adequate.) [Delhi Univ., MBA, 2005]
The sequence of random numbers used is: 0, 9, 1, 1, 5, l, 8,
6, 3, 5, 7, 1, 2, 9, using the first number for day one. Your

CHAPTER SUMMARY
Simulation is a powerful and intuitive technique and uses computer to simulate the operation of an entire system or process.
Random numbers are generated using a probability distribution to generate various outcomes over a period of time. These
outcomes provide at a glance view of different configurations of the system at a least cost in comparison of actually operating
the system. Hence, many alternative system configurations can be investigated and compared before selecting the most
appropriate one to use.
Simulation approach has applications to a wide variety of areas such as queuing system, inventory system, replacement, PERT
projects, financial risk analysis, health care system, distribution system, etc.
Spread sheet software is increasingly being used to perform basic computer simulations. The availability of such a software
enables decision makers to use simulation approach for solving real-life decision problems.

CHAPTER CONCEPTS QUIZ

True or False
1. The purpose of using simulation technique is to reduce the cost 8. Biased random sampling is made from among alternative which
of experiment on a real-life model. have equal probability.
2. The results of simulation experiment should be viewed as exact. 9. To assign random numbers in Monte-Carlo simulation, it is
3. One of the causes of simulation analysis failure is incomplete necessary to assign the particular appropriate random numbers.
mix of essential skills. 10. Simulation should not be applied in all cases because it provides
4. The step required for simulation approach in solving a problem at best approximate solution to problem.
is to design an experiment.
5. The general purpose system simulation language needs a set Fill in the Blanks
of equation to describe a system.
11. Mathematical models can not be ________ to destroy its
6. Simulation is defined as a technique that uses computer. acceptability as a reasonable representation of a system under
7. Analytical results are taken into consideration before a simula- study.
tion study so as to identify suitable values of the system 12. As the complexity of a model increases, simulation seeks to
parameters. ________ the uncertainty in the model.
 Simulation 703

13. Simulation is the process of a model of a real system and 27. Biased random sampling is made from among alternatives which
conducting ________. have
14. Simulation is an approach for reproducing the processes by (a) equal probability
which ________ and change are created in a computer. (b) unequal probability
15. Determining whether the model is internally correct in a logical (c) probability which do not sum to 1
and programming sense called ________. (d) none of the above
16. The random numbers generated by a computer software are 28. Large complicated simulation models are appreciated because
uniformly distributed fractions between ________ and ________. (a) their average costs are not well-defined
17. Using simulation for queuing problem would be appropriate if the (b) it is difficult to create the appropriate events
________ follows a Poisson distribution. (c) they may be expensive to write and use as an experimental
18. Special-purpose simulation languages includes ________ and device
________. (d) all of the above
19. The Monte-Carlo method of simulation is developed through the 29. Simulation should not be applied in all cases because it
use of ________ and ________. (a) requires considerable talent for model building and exten-
20. Validation is the process of ________ a model to the real system sive computer programming efforts
that it represents to make sure that it is ________. (b) consumes much computer time
(c) provides at best approximate solution to problem
Multiple Choice (d) all of the above
30. Simulation is defined as
21. An advantage of simulation as opposed to optimization is that
(a) a technique that uses computers
(a) several options of measure of performance can be exam-
(b) an approach for reproducing the processes by which
ined
events by chance and changes are created in a computer
(b) complex real-life problems can be studied
(c) it is applicable in cases where there is an element of (c) a procedure for testing and experimenting on models to
randomness in a system answer what if . . ., then so and so . . . types of questions
(d) all of the above (d) all of the above
22. The purpose of using simulation technique is to 31. The general purpose system simulation language
(a) imitate a real-world situation (a) requires programme writing
(b) understand properties and operating characteristics of (d) does not require programme writing
complex real-life problems (c) requires predefined coding forms
(c) reduce the cost of experiment on a model of real situation (d) needs a set of equations to describe a system
(d) all of the above 32. Special simulation languages are useful because they
23. Which of the following is not the special purpose simulation (a) reduce programme preparation time and cost
language (b) have the capability to generate random variables
(a) BASIC (b) GPSS (c) require no prior programming knowledge
(c) GASP (d) SIMSCRIPT
(d) all of the above
24. As simulation is not an analytical model, therefore, the result of 33. Few causes of simulation analysis failure are
simulation must be viewed as
(a) inadequate level of user participation
(a) unrealistic (b) exact
(c) approximation (d) simplified (b) inappropriate levels of detail
25. While assigning random numbers in Monte Carlo simulation, it is (c) incomplete mix of essential skills
(a) not necessary to assign the exact range of random number (d) all of the above
interval as the probability 34. To make simulation more popular, we need to avoid
(b) necessary to develop a cumulative probability distribution (a) large cost over runs
(c) necessary to assign the particular appropriate random (b) prolonged delays
numbers (c) user dissatisfaction with simulation results
(d) all of the above
(d) all of the above
26. Analytical results are taken into consideration before a simula- 35. The important step required for simulation approach in solving
tion study so as to a problem is to
(a) identify suitable values of the system parameters
(a) test and validate the model
(b) determine the optimal decision
(b) design the experiment
(c) identify suitable values of decision variables for the specific
choices of system parameters (c) conduct the experiment
(d) all of the above (d) all of the above

Answers to Quiz
1. T 2. F 3. T 4. T 5. F 6. T 7. F 8. F 9. F 10. T
11. manipulated 12. replicate 13. designing, experiments 14. events of chance 15. internal validity
16. zero, one 17. arrival rate 18. GPSS, BASIC 19. probability distribution, random number
20. Comparing, accurate
21. (d) 22. (d) 23. (a) 24. (c) 25. (b) 26. (c) 27. (b) 28. (c) 29. (d) 30. (d) 31. (b)
32. (d) 33. (d) 34. (d) 35. (d)
704 Operations Research: Theory and Applications

CASE STUDY

Case 19.1: Manisha Enterprise

Manisha Enterprise desire to evaluate cash flows for planning purposes. This firm receives orders the beginning of
the month for delivery at the end of the month. Monthly sales and the associated probabilities are as follows:
Monthly sales (units) : 100 125 150
Probability : 0.60 0.30 0.10
The unit selling price is Rs 250. Cash is received either in the month of sale or the following month. From previous
experience the firm has found that if collection is delayed until the following month, 5 per cent of the receivables are
not collectable. Collection and associated probabilities are:
Collection made : Current month Following month
Probability : 0.70 0.30
Fixed cash disbursements are Rs 1.5 million a month. Variable cost disbursements are Rs 100 per unit produced
(sold). If overtime (OT) is required, the estimated cash disbursements are:

Production (units) Overtime (OT ) Probability

100 OT not required 0.80

OT required 0.20
125 OT not required 0.70
OT required 0.30
150 OT not required 0.50
OT required 0.50

Overtime costs are 50 per cent more on variables costs. You are expected to suggest management of the firm a
cash flow pattern to improve total revenue.

Case 19.2: JK Oil Mills

The manager of JK oil mills warehouse is interested in designing an inventory control system for one of the products
in stock. The demand for the product comes from numerous retail outlets and the orders arrive on a daily basis. The
warehouse receives its stock from a factory but the lead time is not constant. The manager wants to determine the
best time to release orders to the factory so that stockouts are minimized yet inventory holding costs are at acceptable
levels. Any orders from retailers, not supplied on a given day, constitute lost demand. Based on a sampling study,
the following data are available:

Daily Probability Lead Time Probability

Demand (days)
1 0 1 0.20
2 0.20 2 0.50
3 0.50 3 0.30
4 0.20
5 0.10

Two alternative ordering policies have been proposed:

Policy 1 : Whenever the inventory level drops below 6 units (reorder point), order 10 units (lot size).
Policy 2 : Whenever the inventory level drops below 10 units, order 10 units.
(a) Simulate each of these two policies for 20 days. Assume that you have 12 units in inventory at the start of the simulation.
(b) Compare and contrast the outcomes for each of these two policies.

APPENDIX: THE SEVEN MOST FREQUENT CAUSES OF SIMULATION

ANALYSIS FAILURE AND HOW TO AVOID THEM*
Over the years we have observed that the following reasons for simulation failure crop up repeatedly . . .
And that they can be avoided!

* Based on, Joseph S. Annino and E.C. Russell The Seven Most Frequent Causes of Simulation Analysis Failure and How to
Avoid Them, Interfaces, Vol. 11, No. 3, 1981.