Input Data Collection and Analysis.pptx

1
INPUT DATA COLLECTION AND
ANALYSIS
Senir Justin
BITF20M550

Introduction
2
Input data includes gathering, studying, and utilizing input data
in the simulation method
The collection may be gathered from any source.
 A study of input data shows the theoretical distribution of data
the practitioner only gathers a sample of the actual data
distribution when collecting data..

Collecting Input Data
3
There are many ways to collect input data like the following:
 Historical records
 Manufacturer specifications
 Vendor claims
 Operator estimates
 Management estimates
 Automatic data capture
 Direct observation

Collecting Input Data
4
 Input data may be collected manually or with the assistance of
electronic devices
 It is the most difficult part of the simulation process
 While colleting input data, there are different classifications of data

Classification of Data
5
There are two methods for the classification of data:
 Deterministic or probabilistic
 Discrete or continuous

Deterministic /Probabilistic Data
6
Deterministic Data
 Deterministic data are those in which the event affecting the
data occurs consistently or predictably.
Probabilistic Input Data
 A probabilistic process does not occur with the same type of
regularity.
 This implies that since the value of this type of data never
changes, it only has to be gathered once.

Discrete/Continuous Data
7
Discrete Data
 It can take only certain values. Usually, this means a whole number.
 The number of students in class is an example
Continuous Data
 It can take any value in the observed range. This means that
fractional numbers are a definite possibility
 Height of children is an example

Input Data Distributions
8
Bernoulli Distribution:
 Models a random occurrence with one of two possible
outcomes
 Frequently referred to as a success or failure
Uniform Distribution:
 It can be used as a first cut for modeling the input data
 It may be either discrete or continuous
Exponential Distribution :
 Commonly utilized in conjunction with interarrival processes
 Random no. of entities will arrive within a specific time

Input Data Distributions
9
Triangle Distribution:
 Used in situations where the practitioner does not have complete
knowledge about the system
 It has only three parameters:
i) Minimum Possible Value
ii) Most Common Value
iii) Maximum Possible Value

Less Common Distributions
10
Geometric distributions:
The geometric distribution gives the probability of achieving success after
N number of failures
 It is discrete which means that distribution must be whole number
Weibull Distribution:
The Weibull distribution is often used to represent distributions that cannot
have values less than zero
 It has two parameters

Analyzing Input Data
11
Graphics Approach:
Graphic approach is the most fundamental approach to attempting to fit
input data
 It consists of visual qualitative comparison between actual and
theoretical data distribution
Chi-Square:
The chi-square test is based on the comparison of the actual number of
observations
versus the expected number of observations
Commonly accepted as preferred goodness fit technique
Kolmogorov–Smirnov:
The KS test should be utilized only when the number of data points is

Software Implementations for
Data Fitting
12
 Fitting a significant no. of observed data sets to theoretical
distributions is a time consuming task
 For this purpose, practitioners use data-fitting software
 The following two are frequently used to carry out this function:
1. Arena input analyzer
2. Expert fit

Arena input analyzer
13
 Input analyzer is part of ARENA simulation software package
available from Rockwell software
 It has the capability to:
1. Determine the quality of fit of probability distribution functions to
input data
2. Examine a total of 15 distributions for data fitting
3. Calculate Chi-square, KS and square error tests
4. Generate high-quality data plots

Expert Fit
14
 This software is available through Averil M. Law & Associates
 This software has the capability to:
1. Automatically determine best probability distribution for data
sets
2. Fits 40 distributions
3. Conduct Chi-sqaure, KS and Anderson-Darling goodness of fit
tests
4. Provide high-quality plots
5. Analyze a large no. of data sets in batch mode

Input Data Collection and Analysis.pptx

More Related Content

Similar to Input Data Collection and Analysis.pptx (20)

Recently uploaded (20)

Input Data Collection and Analysis.pptx