10.06.2016 Views

eldo_user

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Monte Carlo Analysis<br />

Correlation Between Gaussian Input Variables<br />

Correlation Between Gaussian Input Variables<br />

You can specify linear correlations between random variables in Eldo under some conditions:<br />

the variables must have Gaussian distribution and only LOT variations. We review some basic<br />

notions on multivariate probabilities.<br />

Basic Concepts on Multivariate Distributions<br />

A multivariate random variable is a vector of random variables. Notation is made more compact<br />

by writing X = (X 1 , …, X k ) T . Each of the components X i is a random variable in its own right,<br />

but specification of the properties of X as a whole requires information about influence of every<br />

variable on each of the others. Knowledge that one component is large therefore increases the<br />

probability that the other components are large.<br />

Generalizing the single variable case, the joint distribution function of X is defined by:<br />

where x = (x 1 , …, x k ) T . When the X i are continuous random variables, and provided it exists, the<br />

joint density function is given by:<br />

In this case:<br />

The probability density functions of each of the individual X i , termed marginal density<br />

functions, are obtained by integrating out the other components. For example:<br />

is the marginal probability density function of the component X 1 . Similarly:<br />

Eldo® User's Manual, 15.3 437

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!