Chapter 9

Input Modelling

Discrete-Event System Simulation

 Your group project
- Topic selection
- Draft plan
- Weighted will be high
- Presentation before Final Exam
(representative group)
 Outline

1. Data Collection
2. Identifying Distribution
3. Parameter Estimation
4. Goodness-of-Fit Tests
5. Fitting a Non-stationary Poisson Process
6. Selecting Input Modeling w/o Data
7. Multivariate and Time-Series Input
Models
 Overview

• Input models provide the driving force for a

simulation model.
• The quality of the output is no better than the
quality of inputs.
• In this chapter, we will discuss the 4 steps of
input model development:
– Collect data from the real system
– Identify a probability distribution to represent the
input process
– Choose parameters for the distribution
– Evaluate the chosen distribution and parameters
for goodness of fit. 10
 Data Collection

• One of the biggest tasks in solving a real problem. GIGO –

garbage-in-garbage-out
• Suggestions that may enhance and facilitate data collection:
– Plan ahead: begin by a practice or pre-observing session, watch
for unusual circumstances
– Analyze the data as it is being collected: check adequacy
– Combine homogeneous data sets, e.g. successive time periods,
during the same time period on successive days
– Be aware of data censoring: the quantity is not observed in its
entirety, danger of leaving out long process times
– Check for relationship between variables, e.g. build scatter
diagram
– Check for autocorrelation
– Collect input data, not performance data

11
 Identify the Distribution

• Histograms
• Selecting families of distribution
• Parameter estimation
• Goodness-of-fit tests
• Fitting a non-stationary process

12
 Histogram [Identify the Distribution]

• A frequency distribution or histogram is useful in

determining the shape of a distribution
• The number of class intervals depends on:
– The number of observations
– The dispersion of the data
– Suggested: the square root of the sample
size
• For continuous data:
– Corresponds to the probability density
function of a theoretical distribution
• For discrete data:
– Corresponds to the probability mass function
• If few data points are available: combine adjacent
cells to eliminate the ragged appearance of the
histogram
Fig 9.3. Ragged, coarse, and appropriate histograms
a) Original data- too ragged
b) Combining adjacent cells-too coarse
13
c) Combining adjacent cells- appropriate
 Discrete Data

Example 9.5 page 342

The number of vehicles at the northwest corner of an intersection in a 5 minute
period between 7:00 AM and 7:05 AM was monitored for five workday over a
20-week period.

Table 9.1 Number of Arrivals in a 5-Minute Period

Discrete variables
Arrivals per Amble data
Period Frequency  Histogram may
0 12
have a cell for each
1 10
2 19
possible value
3 17
4 10
5 8
6 7
7 5
8 5
9 3
10 3
Fig 9.4 Histogram of number of arrivals per period
11 1
14
Selecting the Family of Distribution [Identify the Distribution]

• A family of distributions is selected based on:

– The context of the input variable

– Shape of the histogram

• Frequently encountered distributions:

– Easier to analyze: exponential, normal and Poisson

– Harder to analyze: beta, gamma and Weibull

15
Selecting the Family of Distribution [Identify the Distribution]
• Use the physical basis of the distribution as a guide, for
example (page 346):
– Binomial: # of successes in n trials
– Poisson: # of independent events that occur in a fixed
amount of time or space
– Normal: dist’n of a process that is the sum of a number of
component processes
– Exponential: time between independent events, or a
process time that is memoryless
– Weibull: time to failure for components
– Discrete or continuous uniform: models complete
uncertainty
– Triangular: a process for which only the minimum, most
likely, and maximum values are known
– Empirical: resamples from the actual data collected 16
Selecting the Family of Distribution [Identify the Distribution]

• Remember the physical characteristics of the

process
– Is the process naturally discrete or continuous valued?

– Is it bounded?

• No “true” distribution for any stochastic input

process
• Goal: obtain a good approximation

17
Quantile-Quantile Plot [Identify the Distribution]

 j - 0.5 
y j is approximately F -1  
 n 

18
Quantile-Quantile Plot [Identify the Distribution]

• The plot of yj versus F-1( (j-0.5)/n) is

– Approximately a straight line if F is a member of an
appropriate family of distributions
– The line has slope 1 if F is a member of an
appropriate family of distributions with appropriate
parameter values

19
Quantile-Quantile Plot [Identify the Distribution]

• Example 9.7 page 348: Check whether the door installation

times follows a normal distribution.

– The observations are now ordered from smallest to largest:

j Value j Value j Value

1 99.55 6 99.98 11 100.26
2 99.56 7 100.02 12 100.27
3 99.62 8 100.06 13 100.33
4 99.65 9 100.17 14 100.41
5 99.79 10 100.23 15 100.47
– yj are plotted versus F-1( (j-0.5)/n) where F has a normal
distribution with the sample mean (99.99 sec) and sample
variance (0.28322 sec2)

20
Quantile-Quantile Plot [Identify the Distribution]

• Example (continued): Check whether the door

installation times follow a normal distribution.

Fig. 8.6 A q-q plot

and histogram of the
installation times

21
Quantile-Quantile Plot [Identify the Distribution]

• Consider the following while evaluating the linearity of a q-

q plot:
– The observed values never fall exactly on a straight line
– The ordered values are ranked and hence not independent,
unlikely for the points to be scattered about the line
– Variance of the extremes is higher than the middle. Linearity
of the points in the middle of the plot is more important.
• Q-Q plot can also be used to check homogeneity
– Check whether a single distribution can represent both
sample sets
– Plotting the order values of the two data samples against
each other

22
Parameter Estimation [Identify the Distribution]

• Next step after selecting a family of distributions

• If observations in a sample of size n are X1, X2, …, Xn (discrete
or continuous), the sample mean and variance are:

i1 X i 
n n 2 2
X i  n X
X S2  i 1

n n 1

• If the data are discrete and have been grouped in a frequency

distribution:
 j 1 f j X j  j 1 j j
n n 2 2
f X  nX
X S2 
n n 1

where fj is the observed frequency of value Xj

23
Parameter Estimation [Identify the Distribution]

• When raw data are unavailable (data are grouped into

class intervals), the approximate sample mean and
variance are:

 j 1 f j X j  j 1 j j
c n 2 2
f m  nX
X S2 
n n 1

where fj is the observed frequency of in the jth class interval

mj is the midpoint of the jth interval, and c is the number of class
intervals

• A parameter is an unknown constant, but an estimator is

a statistic.

24
Parameter Estimation [Identify the Distribution]

• Vehicle Arrival Example (continued): Table in the histogram example

on slide 7 (Table 9.1 in book) can be analyzed to obtain:
n  100, f1  12, X 1  0, f 2  10, X 2  1,...,

 j 1 f j X j  364, and  j 1 f j X 2j  2080

k k
and
– The sample mean and variance are

364
X  3.64
100
2 2080  100 * (3.64) 2
S 
99
 7.63

– The histogram suggests X to have a Poisson distribution

• However, note that sample mean is not equal to sample variance.
• Reason: each estimator is a random variable, is not perfect. 25
Parameter Estimation [Identify the Distribution]

Example 9.10 Exponential Distribution (page 353)

Example 9.11 Weibull Distribution (page 354)
Example 9.12 Poisson Distribution (page 356)
Example 9.13 Lognormal Distribution (page 356)
Example 9.14 Normal Distribution (page 356)
Example 9.15 Gamma Distribution (page 356)
Example 9.16 Beta Distribution (page 358)

26
Goodness-of-fit Test

• Conduct hypothesis testing on input data

distribution using:
– Kolmogorov-Smirnov test
– Chi-square test
• No single correct distribution in a real application
exists.
– If very little data are available, it is unlikely to reject any
candidate distributions
– If a lot of data are available, it is likely to reject all
candidate distributions

27
Chi-square Test [Goodness-of-fit Test]

• Intuition: comparing the histogram of the data to the

shape of the candidate density or mass function
• Valid for large sample sizes when parameters are
estimated by maximum likelihood
• By arranging the n observations into a set of k class
intervals or cells, the test statistics is:

k
(Oi  Ei ) 2 Expected Frequency
 02  
i 1
Ei
Ei = n*pi
where pi is the theoretical
Observed
Frequency prob. of the ith interval.
Suggested Minimum = 5

which approximately follows the chi-square distribution with k-s-1

degrees of freedom, where s = # of parameters of the hypothesized
28
distribution estimated by the sample statistics.
Chi-square Test [Goodness-of-fit Test]

• The hypothesis of a chi-square test is:

H0: The random variable, X, conforms to the distributional
assumption with the parameter(s) given by the estimate(s).
H1: The random variable X does not conform.

• If the distribution tested is discrete and combining

adjacent cell is not required (so that Ei > minimum
requirement):
– Each value of the random variable should be a class interval,
unless combining is necessary, and
pi  p(xi )  P(X  xi )
29
Chi-square Test [Goodness-of-fit Test]

• If the distribution tested is continuous:

ai
pi  
ai 1
f ( x) dx  F (ai )  F (ai 1 )

where ai-1 and ai are the endpoints of the ith class interval
and f(x) is the assumed pdf, F(x) is the assumed cdf.
– Recommended number of class intervals (k):

Sample Size, n Number of Class Intervals, k

20 Do not use the chi-square test
50 5 to 10
100 10 to 20
1/2
– Caution: Different
> 100
grouping of data
n to (i.e.,
n/5 k) can affect the
hypothesis testing result.

30
Chi-square Test [Goodness-of-fit Test]

• Vehicle Arrival Example (continued):

H0: the random variable is Poisson distributed.
H1: the random variable is not Poisson distributed.

xi Observed Frequency, Oi Expected Frequency, Ei (Oi - Ei)2/Ei Ei  np ( x)

0 12 2.6
1 10 9.6
7.87 e   x
2 19 17.4 0.15
n
3 17 21.1 0.8
x!
4 19 19.2 4.41
5 6 14.0 2.57
6 7 8.5 0.26
7 5 4.4
8 5 2.0
9 3 0.8 11.62 Combined because
10 3 0.3
> 11 1 0.1 of min Ei
– Degree of freedom is k-s-1 = 7-1-1 = 5, hence,27.68
100
the hypothesis
100.0

is rejected at the 0.05 level of significance.

 02  27.68   02.05,5  11 .1 31
Kolmogorov-Smirnov [Goodness-of-fit Test]

• Intuition: formalize the idea behind examining a q-q plot

• Recall from Section 7.4.1:
– The test compares the continuous cdf, F(x), of the
hypothesized distribution with the empirical cdf, SN(x), of the N
sample observations.
– Based on the maximum difference statistics (Tabulated in A.8):

D = max| F(x) - SN(x)|

• A more powerful test, particularly useful when:
– Sample sizes are small,
– No parameters have been estimated from the data.
• When parameter estimates have been made:
– Critical values in Table A.8 are biased, too large.
– More conservative, i.e., smaller Type I error than specified.
32
p-Values and “Best Fits” [Goodness-of-fit Test]

• p-value for the test statistics

– The significance level at which one would just reject H0
for the given test statistic value.
– A measure of fit, the larger the better
– Large p-value: good fit
– Small p-value: poor fit

• Vehicle Arrival Example (cont.):

– H0: data is Possion
 02  27.68
– Test statistics: , with 5 degrees of freedom
– p-value = 0.00004, meaning we would reject H0 with
0.00004 significance level, hence Poisson is a poor fit.
33
p-Values and “Best Fits” [Goodness-of-fit Test]

• Many software use p-value as the ranking

measure to automatically determine the “best fit”.
Things to be cautious about:
– Software may not know about the physical basis of the
data, distribution families it suggests may be
inappropriate.
– Close conformance to the data does not always lead to
the most appropriate input model.
– p-value does not say much about where the lack of fit
occurs
• Recommended: always inspect the automatic
selection using graphical methods.
34
Fitting a Non-stationary Poisson Process

• Fitting a NSPP to arrival data is difficult, possible

approaches:
– Fit a very flexible model with lots of parameters or
– Approximate constant arrival rate over some basic
interval of time, but vary it from time interval to time
interval.

• Suppose we need to model arrivals over time

[0,T], our approach is the most appropriate when
we can:
– Observe the time period repeatedly and
– Count arrivals / record arrival times.
35
Fitting a Non-stationary Poisson Process
• The estimated arrival rate during the ith time period is:

1 n
̂ (t )  
nt j 1
Cij

where n = # of observation periods, t = time interval length

Cij = # of arrivals during the ith time interval on the jth observation period

• Example: Consider a 10-hour business day [8am,6pm] allow the

rate to change every half hour, then T = 10, k = 20 and t = ½,
and observe over n =3 days
Number of Arrivals Estimated Arrival
Time Period Day 1 Day 2 Day 3 Rate (arrivals/hr)

8:00 - 8:30 12 14 10 24 For instance,

1/3(0.5)*(23+26+32)
8:30 - 9:00 23 26 32 54
= 54 arrivals/hour
9:00 - 9:30 27 19 32 52

9:30 - 10:00 20 13 12 30

36
Selecting Model without Data
• If data is not available, some possible sources to
obtain information about the process are:
– Engineering data: often product or process has
performance ratings provided by the manufacturer or
company rules specify time or production standards.
– Expert option: people who are experienced with the
process or similar processes, often, they can provide
optimistic, pessimistic and most-likely times, and they
may know the variability as well.
– Physical or conventional limitations: physical limits on
performance, limits or bounds that narrow the range of
the input process.
– The nature of the process.
• The uniform, triangular, and beta distributions are
often used as input models. 37
Selecting Model without Data
• Example 9.20: Production planning simulation.
– Input of sales volume of various products is required,
salesperson of product XYZ-123 says that:
• No fewer than 1,000 units and no more than 5,000 units will be sold.
• Given her experience, she believes there is a 90% chance of selling
more than 2,000 units, a 25% chance of selling more than 3,500
units, and only a 1% chance of selling more than 4,500 units.

– Translating these information into a cumulative probability of

being less than or equal to those goals for simulation input:
i Interval (Sales) Cumulative Frequency, ci
1 1000  x  2000 0.10
2 2000 < x  3500 0.75
3 3500 < x  4500 0.99
4 4500 < x  5000 1.00
38
Multivariate and Time-Series Input Models

• Multivariate:
– For example, lead time and annual demand for an
inventory model, increase in demand results in lead
time increase, hence variables are dependent.

• Time-series:
– For example, time between arrivals of orders to buy
and sell stocks, buy and sell orders tend to arrive in
bursts, hence, times between arrivals are dependent.

39
Multivariate and Time-Series Input Models

Example 9.21: A supply chain simulation the lead time

and annual demand for industrial robots. An increase in
demand results in an increase in lead time (Page 369)

Example 9.22: A simulation of the web-based trading

site of a stock broker includes the time between arrivals
of orders to buy and sell (Page 369)

40
Covariance and Correlation
[Multivariate and Time-Series Input Models]

• Consider the model that describes relationship between X1 and X2:

 is a random
( X 1  1 )   ( X 2   2 )   variable with mean 0
and is independent
of X2

  = 0, X1 and X2 are statistically independent

  > 0, X1 and X2 tend to be above or below their means together
  < 0, X1 and X2 tend to be on opposite sides of their means

• Covariance between X1 and X2 :

cov( X 1 , X 2 )  E[( X 1  1 )( X 2   2 )]  E ( X 1 X 2 )  1 2

= 0, =0
– where cov(X1, X2) < 0, then  <0
> 0, >0
41
Covariance and Correlation
[Multivariate and Time-Series Input Models]

• Correlation between X1 and X2 (values between -1 and 1):

cov( X 1 , X 2 )
  corr ( X 1 , X 2 ) 
 1 2

= 0, =0
– where corr(X1, X2) < 0, then  < 0
> 0, >0
– The closer  is to -1 or 1, the stronger the linear
relationship is between X1 and X2.

42
Covariance and Correlation
[Multivariate and Time-Series Input Models]

• A time series is a sequence of random variables X1, X2, X3, … , are

identically distributed (same mean and variance) but dependent.
– cov(Xt, Xt+h) is the lag-h autocovariance

– corr(Xt, Xt+h) is the lag-h autocorrelation

– If the autocovariance value depends only on h and not on t, the time series is
covariance stationary

43
Covariance and Correlation
[Multivariate and Time-Series Input Models]

• If X1 and X2 are normally distributed, dependence

between them can be modeled by the bivariate normal
distribution with 1, 2, 12, 22 and correlation 
– To Estimate 1, 2, 12, 22, see “Parameter Estimation”
– To Estimate , suppose we have n independent and identically
distributed pairs (X11, X21), (X12, X22), … (X1n, X2n), then:

1 n
côv( X 1 , X 2 )   ( X 1 j  Xˆ 1 )( X 2 j  Xˆ 2 )
n  1 j 1
1  n 
   X 1 j X 2 j  nXˆ 1 Xˆ 2 
n  1  j 1 

côv( X 1 , X 2 )
ˆ 
ˆ1ˆ 2 Sample deviation
44
Covariance and Correlation
[Multivariate and Time-Series Input Models]

• If X1, X2, X3,… is a sequence of identically distributed, but

dependent and covariance-stationary random variables,
then we can represent the process as follows:
– Autoregressive order-1 model, AR(1)
– Exponential autoregressive order-1 model, EAR(1)
• Both have the characteristics that:

 h  corr( X t , X t  h )   h , for h  1,2,...

• Lag-h autocorrelation decreases geometrically as the lag increases,

hence, observations far apart in time are nearly independent

45
AR(1) Time-Series Input Models
[Multivariate and Time-Series Input Models]

• Consider the time-series model:

X t     ( X t 1   )   t , for t  2,3,...
where  2 ,  3 ,  are i.i.d. normally distributed with   0 and variance 2

• If X1 is chosen appropriately, then

– X1, X2, … are normally distributed with mean = , and
variance = /(1-)
– Autocorrelation h = h
• To estimate , 2:
ˆ côv( X t , X t 1 )
ˆ  X , ˆ  ˆ (1  ˆ 2 ) ,
2

2

ˆ 2
46
where côv( X t , X t 1 ) is the lag-1 autocovariance
EAR(1) Time-Series Input Models
[Multivariate and Time-Series Input Models]

• Consider the time-series model:

X t 1 , with probability 
Xt   for t  2,3,...
X t 1   t , with probability 1-φ
where  2 ,  3 ,  are i.i.d. exponentially distributed with  ε  1/λ, and 0    1

• If X1 is chosen appropriately, then

– X1, X2, … are exponentially distributed with mean = 1/
– Autocorrelation h = h , and only positive correlation is
allowed.
• To estimate :
ˆ côv( X t , X t 1 )
ˆ  1 / X , ˆ

ˆ 2
47
where côv( X t , X t 1 ) is the lag-1 autocovariance
 HW/Asignment

- Redo Example 9.7 from slide 13

- Use data from Example 9.5 at slide 07
– Find the best fit of distributions
– Estimate parameters
– Comment on your results and conclusions

48