Industrial Systems Engineering - International University – VNU-HCM SIMULATION

Chapter 8

Random Variate Generation

Discrete-Event System Simulation

1
 Today question

Random Numbers

Probability Distributions
Random Variates EXPO(..)
Arrival rate TRIA(..)
Time between arrival NORM(..)
Process time WEIB(..)
Lead time empirical
… …

How to use Random Numbers to generate Random Variates?

2
 Recall..

Grocery store with only one checkout counter.

▪ Customers arrive at random times from
1 to 8 minutes apart, with equal ▪ The service times vary from 1 to 6
probability of occurrence: minutes,
with probabilities:

Time between arrival

relation?
follows uniform Service time follows somerelation?
distribution statistical distribution? 3
Techniques to generate Random Variates

– Inverse-transform technique
– Acceptance-rejection technique
– Special properties

4
 1. Inverse-transform Technique
1.1. Continuous distribution
● The concept: Suppose we
wish to generate a Random
Variate X that is continuous,
has distribution function cdf F:
0 < F(X) <1. Denote F-1 the
inverse of function F.
● The algorithm is F(x)
– For cdf function: R = F(X) 1
R = F(x)
1. Generate R from uniform U(0,1) R
2. Return X:

X = F-1(R) X
Fig. 8.1 Inverse Transform Method5 5
a. Exponential Distribution
● Exponential Distribution:
– Exponential

– Algorithm
● Generate R ~U(0,1)
● Solve 1-e-λx = R

● If we want to generate n values of

X:: X1, X2,…, Xn
- Generate uniform RN: R1, R2,…, Rn
- Return:

; i= 1, 2, …, n

6
a. Exponential Distribution (cont)

7
8
b. Uniform Distribution U(a,b)

– Uniform distribution

– Algorithm
●Generate R~U(0,1)
●Return X = a + (b-a)R

9
c. Weibull Distributions

– Weibull distribution

α, β are scale and shape parameters

– Algorithm
– Generate R~U(0,1)
– Return

10
d. Triangular distribution

11
12
Triangular Distribution (cont)

13
 1.2 Empirical Continuous Distributions
● Unable to find theoretical distribution → use empirical distribution
● Procedure:
– Collect data Xi; i= 1, 2, …, n
– Arrange data from smallest to largest
– Assign the probability 1/n to each interval [xi-1, xi]
– Compute the slope of ith line segment
i/n
R

– Generate R
(i-1)/n
– Return
xi-1 x xi

14
1.2 Empirical Continuous Distributions (cont)
● Five observations of fire-crew response times (in mins.):
– 2.76 1.83 0.80 1.45 1.24
– Arranging x0 = 0; x1 = 0.8; x2 = 1.24; x3 = 1.45; x4 = 1.83; x5 = 2.76
– n=5 probability for each interval = 1/n = 0.2

15
1.2 Empirical Continuous Distributions (cont)

Consider R1 = 0.71:

(i-1)/n = 0.6 < R1 < i/n =

0.8

X1 = x(4-1) + a4(R1 – (4-1)/n)

= 1.45 + 1.90(0.71-0.60)
= 1.66

16
 1. 3 Discrete Distribution

● Procedure
P(x)
– With given pmf P(xi) = P(X=xi),
compute cdf

(Assume that X can take only the value x 1<x2<… <xn) X1 X2 Xi-1 Xi
– Generate R~U(0,1) 1
– Determine the smallest integer I:
R≤ F(xI) R
Return X = xI
F2

X1 X2 Xi-1 Xi
17
1. 3 Discrete Distribution

18
1. 3 Discrete Distribution

19
1. 3 Discrete Uniform Distribution

Generate R~U(0,1)
If R satisfy:

Then return X = i
20
2. Acceptance-Rejection technique

The Acceptance-Rejection Method is useful for

generating samples from a target distribution f(x) when
direct inversion is difficult

21
2. Acceptance-Rejection technique

U∼U(0,1) (a uniform random variable for probabilistic

acceptance)

22
 2. Acceptance-Rejection technique

g(x) should fully covers the range of f(x) and is easy to sample from
eg. f(x) Norm(0,1) → g(x) Unif(-3,3)

23
2. Acceptance-Rejection technique

Generate random variates, X ~ U(1/4, 1)

Step 1: Define the Target Density

Step 2: Choose a Proposal Distribution g(x) ~ U(0,1)

Step 3: Find the Acceptance Ratio

24
25
 2. Acceptance-Rejection technique

● Illustration: To generate random variates, X ~ U(1/4, 1)

Generate R
Procedures:
no
Step 1. Generate R ~ U[0,1] Conditi
Step 2a. If R >= ¼, accept X=R. on
yes
Step 2b. If R < ¼, reject R, return
to Step 1 Output R’

● R does not have the desired distribution, but R conditioned (R’) on the
event {R ≥ ¼} does.
● Efficiency: Depends heavily on the ability to minimize the number of
rejections.

26
Mathematical proof

27
28
29
 Poisson Distribution

Based on the relationship between Poisson(λ) and Expo(1/λ)

EXPO:

30
 Poisson Distribution

•Set n = 0 and P =1
•Generate random number Ri+1 and replace P by

P x Ri+1 .
•If P < e-λ then accept X = n. Otherwise, reject current n, increase n
by 1 and return previous step.
31
Example
Generate three Poisson variates with mean α = 0.2.

32
33
 Direct Transformation [Special Properties]

● Approach for normal(0,1):

– Consider two standard normal random variables, Z1 and Z2, plotted as a
point in the plane:

In polar coordinates:
Z1 = B cos φ
Z2 = B sin φ

– B2 = Z21 + Z22 ~ chi-square distribution with 2 degrees of freedom =

Exp(λ = 2). Hence,
– The radius B and angle φ are mutually independent.

34
35
Direct Transformation [Special Properties]

● Approach for normal (μ,σ2):

– Generate Zi ~ N(0,1)

Xi = μ + σ Zi

● Approach for lognormal(μ,σ2):

– Generate X ~ N(μ,σ2)

Yi = eXi

36
Summary

● Principles of random-variate generates of specified

distributions via
– Inverse-transform technique
– Acceptance-rejection technique
– Special properties

● Important for generating samples for different continuous

and discrete distributions as input to simulation modeling.

THANK YOU.
37