QMT634 – SIMULATION

Topic 2: Static Simulation Model on Spreadsheets

Static Simulation (Monte Carlo Simulation

1. In Topic 1, we introduced the concept of a model as a useful representation for analyzing

the possible consequences of decisions and the concept of a simulation as a sampling
experiment concerning the outcomes of the model.

2. In this topic, we are going to develop the idea of simulation and to explore using static
models.

Static simulation (Monte Carlo simulation) – sampling observations and

transforming them according to formulas or rules that compose the model. Repeated
independently many times.

3. Recall that, static model is one in which we do not record observations on the system over
time (not observing the behavior of system’s model over time).

4. The simulation consists of:

 Generating random variates and combining them according to the formulas or

rules of the model to produce an observation for the output parameter or
performance measure.

 Repeat this process a specific number of times or a specific number of replications

in order to produce a collection of independent, identically distributed observations.

 Then can be analyzed using standard statistical and graphical methods to show
the distribution of the data and estimate parameters of this distribution, such as the
mean performance measure.

5. Under static simulation the objective here is to demonstrate that a simulation can
provide a reliable estimate of the desired output parameter if it is allowed to produce
enough data.

6. Two other aspects of simulation are generation of random variates from some common
distributions and analysis of output data.

Five step in Monte Carlo Simulation

STEP 1: Setting up a probability distribution for important variables.

STEP 2: Building a cumulative probability distribution for each variable in STEP 1.

STEP 3: Establish interval of random numbers for each variable.

STEP 4: Generating random numbers.

STEP 5: Simulating a series of trials.

1
 QMT634 – SIMULATION

Example 2.1:

Harry’s Auto Tire sells all types of tires, but Tire A accounts for a large portion of Harry overall
sales. Harry wishes to simulate the daily demand for 10 days. Find the average daily demand.

STEP 1: Setting up a probability distribution for important variables.

The daily demand for Tire A over the past 200 days is as follows:

Demand for Tire A Frequency Probability

(days)
0 10 10/200 = 0.05
1 20 0.10
2 40 0.20
3 60 0.30
4 40 0.20
5 30 0.15
Total 200 1.00

STEP 2: Building a cumulative probability distribution for each variable in STEP 1.

STEP 3: Establish interval of random numbers for each variable.

Demand for Tire A Probability Cumulative Random Number

Probability Interval
0 0.05 0.05 01 – 05
1 0.10 0.15 06 – 15
2 0.20 0.35 16 – 35
3 0.30 0.65 36 – 65
4 0.20 0.85 66 – 85
5 0.15 1.00 86 - 00

STEP 4: Generating random numbers.

 Random numbers may be generated for simulation problems in several ways such as using
computers or by using random number table as follows.

 For this example, we can generate random numbers using the following random number
table.

2
 QMT634 – SIMULATION

Partial table of random numbers (upper left corner)

52 06 50 88 53 30 10 47 99 37 66 91 35
37 63 28 02 74 35 24 03 29 60 74 85 90

82 57 68 28 05 94 03 11 27 79 90 87 92
69 02 36 49 71 99 32 10 75 21 95 90 94
98 94 90 36 06 78 23 67 89 85 29 21 25
96 52 62 87 49 56 49 23 78 71 72 90 57
33 69 27 21 11 60 95 89 68 48 17 89 34
50 33 50 95 13 44 34 62 63 39 55 29 30
88 32 18 50 62 57 34 56 62 31 15 40 90
90 30 36 24 60 82 51 74 30 35 36 85 01
50 48 61 18 85 23 08 54 17 12 80 69 24
27 88 21 62 69 64 48 31 12 73 02 68 00
45 14 46 32 13 49 66 62 74 41 86 98 92

Source: Reprinted from A Million Random Digits with 100000 Normal Deviates, Rand (New
York: The Free Press, 1995).

STEP 5: Simulating a series of trials.

 Simulate the demand for Tire A for 10 days.

Day Random Number Daily Demand

1 52 3
2 06 1
3 50 3
4 88 5
5 53 3
6 30 2
7 10 1
8 47 3
9 99 5
10 37 3

Total 29

29
Average daily demand  2.9  3 tires
10

3
 QMT634 – SIMULATION

Exercises

Past year Jan 2018, Jun19

2.1 The AAA Health Insurance Company is concern with its cash outflows on a weekly
basis. AAA is being considered for a large group policy. If AAA wins the contract and insures the
group, the daily frequency of claims is estimated as follows:

Number of claims Probability

0 0.05
1 0.06
2 0.12
3 0.34
4 0.25
5 0.11
6 0.07

Cost per claim (RM) Probability

1000 0.30
1100 0.24
1200 0.22
1300 0.18
1400 0.06

i) Calculate the expected number of claim per day and the expected cost per claim. Hence,
estimate the weekly (5 working-day) cash outflow.
ii) Simulate for 5 days to estimate weekly cash outflow. Compare your simulation result
with (i) above. Are they the same? Explain.

Use the following random numbers for number of claim and cost per claim.

Number of claim: 44 06 61 93 59 42 39 87 34 63 88 54

Cost per claim : 35 44 57 99 66 47 43 52 31 72 15 68 80 87 98 20 16 80

43 62

2.2 Pantas Emergency Rescue Squad has gained reputation over the years from the public for
its prompt and quality service. From previous experience, the company establishes the following
probability distribution regarding receipt of number of emergency calls at night.

Number of calls Probability

0 0.06
1 0.15
2 0.18
3 0.30
4 0.24
5 0.07

Pantas categorizes each call as either Minor, Normal or Major, and has found the following
distribution for these categories:

4
 QMT634 – SIMULATION

Category Probability
Minor 0.28
Normal 0.62
Major 0.10

Pantas sends a team comprising 3, 5 or 7 persons for Minor, Normal or Major calls respectively.
Use random numbers given below.

Number of Calls: 47 31 05 76 18 59 35 16 72 60

Type of Calls : 08 56 37 71 92 74 17 13 50 41 27 55
93 10 32 72 99 65 33 07 42 88 22

a) Simulate the emergency calls received by Pantas for 10 nights.

b) Determine the average number of calls of each category per night.
c) Determine the average number of calls received per night.
d) Determine the average crew size required per night.

Example 2.2

Dave’s Candies is a small family-owned business that offers gourmet chocolates and ice cream
fountain service for special occasions such as Teacher’s Day, the store must place an order for
special packaging several weeks in advance from their supplier. One product, Sweet Chocolate
is bought for RM7.50 a box and sells for RM12. Any boxes that are not sold by May 16 are
discounted by 50% and can always be sold easily. Historically Dave’s Candies has sold
between 40 and 80 boxes each year with no apparent trend (either increasing or decreasing). If
demand is exceed the purchase quantity, Dave loses profit opportunity. On the other hand, if too
many boxes are purchased, he will lose money by discounting them below cost. Dave’s
dilemma is deciding how many boxes to order for the Teacher’s Day customers.

Solution:
 If Q boxes are purchased and sales demand is D:

12D  7.50Q  6(Q  D) , if D  Q (2.1)

Profit  
 12D  7.50Q , if D  Q (2.2)

 Input to a simulation model of this situation would be:

1. The order quantity, Q (decision variable)

2. The various revenue and cost factors (constants)
3. The demand, D (uncontrollable and probabilistic)

 The model output we seek is the net profit. If we know the demand, we can use
equation (2.1) or (2.2) to compute the profit. Since demand is probabilistic, we need to
be able to ‘sample’ a value from the probability distribution of demand.

5
 QMT634 – SIMULATION

 For now, we simplify this problem by assuming that demand will be either 40, 50, 60, 70
or 80 boxes with equal probability (1/5 or 0.2).

 Simulate this model for 10 replications with the number of order is either 40, 60 or 80
boxes. From the net profit (simulation results), determine which order quantity will be the
best.

 Use the following random numbers:

Order 40 50 48 61 18 85 23 08 54 17 12
Order 60 27 88 21 62 69 64 48 31 12 73
Order 80 45 14 46 32 13 49 66 62 74 41

Example 2.3

A large catalog merchandiser is planning to have a special furniture promotion a year from now.
To do this, the company must place its order for the furniture now. It plans to sign a contract
with the manufacturer for 3000 chairs at a cost of RM175 per unit, which the company plans to
offer initially RM250 per unit. The promotion will last for eight weeks, after which all remaining
units will be offered at half the initial price or RM125 per unit. The company believes that 2000
units will be sold during the first eight weeks. Determine the profit.

P the profit from the promotion

C the per unit cost for the chairs (RM175)
R the initial price per unit for the chairs (RM250)
S the number of units ordered (3000)
V the number of units sold during the first eight weeks the promotion (2000).

P = (R – C)V + (R/2 – C)(S – V)

What is the value of P = ?

6
 QMT634 – SIMULATION

Exercise 2.4 (Based on Example 2.3)

When dealing with probabilistic models we must consider this:

 Two of the four inputs, C and S are fixed because the company will sign a contract with
the manufacturer. The other two inputs, V and R are uncertain.
 Let say the demand, has a symmetric triangular distribution between 500 and 3500 units,
with a peak at 2000 units. The initial price, R will be between RM200 and RM300 and
uniformly distributed.
 Since these quantities are random variables and the net profit depends on them, the net
profit is also a random variable.

** y-axis: probability density function

 U represents a uniformly distributed random variate between 0 and 1; that is, U = RAND() in
Excel.

Distribution Parameters Formula

Symmetric triangular a<b V = a + ((b-a)/2)(U1 + U2)
Uniform a<b R = a + (b-a)U

 Simulate for 10 replications. Determine the average net profit.

7
 QMT634 – SIMULATION

 Use the following random numbers.

Demand 0.36 0.99 0.14 0.38 0.91 0.55 0.95 0.55 0.66 0.23
0.12 0.26 0.43 0.72 0.66 0.44 0.22 0.66 0.68 0.22
Initial
Price 0.86 0.24 0.47 0.57 0.19 0.38 0.93 0.21 0.20 0.75

Replication Demand Initial Price Net Profit