UTK IE 406 - Statistical Models in Simulation

Unformatted text preview:

IE406, I & IEUTKSimulationStatistical Models in SimulationDr. Xueping LiUniversity of TennesseeBased onBanks, Carson, Nelson & NicolDiscrete-Event System Simulation2Purpose & Overview The world the model-builder sees is probabilistic rather than deterministic.  Some statistical model might well describe the variations. An appropriate model can be developed by sampling the phenomenon of interest: Select a known distribution through educated guesses Make estimate of the parameter(s) Test for goodness of fit We will: Review several important probability distributions Present some typical application of these models3Review of Terminology and Concepts In this section, we will review the following concepts: Discrete random variables Continuous random variables Cumulative distribution function Expectation4Discrete Random Variables [Probability Review] X is a discrete random variable if the number of possible values of X is finite, or countably infinite. Example: Consider jobs arriving at a job shop. Let X be the number of jobs arriving each week at a job shop. Rx= possible values of X (range space of X) = {0,1,2,…} p(xi) = probability the random variable is xi= P(X = xi) p(xi), i = 1,2, … must satisfy: The collection of pairs [xi, p(xi)], i = 1,2,…, is called the probability distribution of X, and p(xi) is called the probability mass function (pmf) of X.∑∞==≥11)( 2. allfor ,0)( 1.iiixpixp5Continuous Random Variables [Probability Review] X is a continuous random variable if its range space Rxis an interval or a collection of intervals. The probability that X lies in the interval [a,b] is given by: f(x), denoted as the pdf of X, satisfies: PropertiesXRXRxxfdxxfRxxfXin not is if ,0)( 3.1)( 2.in allfor , 0)( 1.==≥∫∫=≤≤badxxfbXaP )()(0)( because ,0)( 1.000dxxfxXPxx===∫)()()()( .2 bXaPbXaPbXaPbXaP ≺≺≺≺ =≤=≤=≤≤6Continuous Random Variables [Probability Review] Example: Life of an inspection device is given by X, a continuous random variable with pdf: X has an exponential distribution with mean 2 years Probability that the device’s life is between 2 and 3 years is:⎪⎩⎪⎨⎧≥=−otherwise ,00 x,21)(2/xexf14.021)32(322/==≤≤∫−dxexPx7Cumulative Distribution Function [Probability Review] Cumulative Distribution Function (cdf) is denoted by F(x), where F(x) = P(X <= x) If X is discrete, then  If X is continuous, then  Properties All probability question about X can be answered in terms of the cdf, e.g.:∑≤=xxiixpxF all)()(∫∞−=xdttfxF )()(0)(lim 3.1)(lim 2.)()( then , If function. ingnondecreas is 1.==≤−∞→∞→xFxFbFaFbaFxx≺baaFbFbXaP ≺≺ allfor ,)()()(−=≤8Cumulative Distribution Function [Probability Review] Example: An inspection device has cdf: The probability that the device lasts for less than 2 years: The probability that it lasts between 2 and 3 years:2/02/121)(xxtedtexF−−−==∫632.01)2()0()2()20(1=−==−=≤≤−eFFFXP145.0)1()1()2()3()32(1)2/3(=−−−=−=≤≤−−eeFFXP9Expectation [Probability Review] The expected value of X is denoted by E(X) If X is discrete If X is continuous a.k.a the mean, m, or the 1stmoment of X A measure of the central tendency The variance of X is denoted by V(X) or var(X) or σ2 Definition: V(X) = E[(X – E[X]2] Also, V(X) = E(X2) – [E(x)]2 A measure of the spread or variation of the possible values of X around the mean The standard deviation of X is denoted by σ Definition: Square root of V(X) Expressed in the same units as the mean∑=iiixpxxE all)()(∫∞∞−= dxxxfxE )()(10Expectations [Probability Review] Example: The mean of life of the previous inspection device is: To compute variance of X, we first compute E(X2): Hence, the variance and standard deviation of the device’s life are:22/21)(02/002/=+−==∫−∫∞−∞∞−dxexdxxeXExxxe82/221)(02/002/22=+−==∫−∫∞−∞∞−dxexdxexXExxex428)(2=−=XV2)( == XVσ11Useful Statistical Models In this section, statistical models appropriate to some application areas are presented. The areas include: Queueing systems Inventory and supply-chain systems Reliability and maintainability Limited data12Queueing Systems [Useful Models] In a queueing system, interarrival and service-time patterns can be probablistic. Sample statistical models for interarrival or service time distribution: Exponential distribution: if service times are completely random Normal distribution: fairly constant but with some random variability (either positive or negative) Truncated normal distribution: similar to normal distribution but with restricted value. Gamma and Weibull distribution: more general than exponential (involving location of the modes of pdf’s and the shapes of tails.)13Inventory and supply chain [Useful Models] In realistic inventory and supply-chain systems, there are at least three random variables:  The number of units demanded per order or per time period The time between demands The lead time Sample statistical models for lead time distribution: Gamma Sample statistical models for demand distribution:  Poisson: simple and extensively tabulated. Negative binomial distribution: longer tail than Poisson (more large demands). Geometric: special case of negative binomial given at least one demand has occurred.14Reliability and maintainability [Useful Models] Time to failure (TTF) Exponential: failures are random Gamma: for standby redundancy where each component has an exponential TTF Weibull: failure is due to the most serious of a large number of defects in a system of components Normal: failures are due to wear15Other areas [Useful Models] For cases with limited data, some useful distributions are: Uniform, triangular and beta  Other distribution: Bernoulli, binomial and hyperexponential.16Discrete Distributions Discrete random variables are used to describe random phenomena in which only integer values can occur. In this section, we will learn about: Bernoulli trials and Bernoulli distribution Binomial distribution Geometric and negative binomial distribution Poisson distribution17Bernoulli Trials and Bernoulli Distribution[Discrete Dist’n] Bernoulli Trials:  Consider an experiment consisting of n trials, each can be a


View Full Document

UTK IE 406 - Statistical Models in Simulation

Download Statistical Models in Simulation
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Statistical Models in Simulation and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Statistical Models in Simulation 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?