# Interval variables representation in uqlab

Good Morning,

Thank you all for the useful software. I have a question regarding interval variables. How does uqlab handle interval variables? For example, I have a limit state function with three variables where two are random and the last an interval variable with just bounds.

g(X)= X_1X_2-X_3

in the above expression, X_1 and X_2 are random variables and X_3 an interval variable.

X1 and X2 are standard normal. For the interval variable, the bounds are [3, 8]. Is there a way to represent the interval variable X_3 in uqlab given that it has no distribution but just bounds as stated?

uqlab
MOpts.mString ='X(:,1).*X(:,2) - X(:,3)';
MOpts.isVectorized = 1;
myModel = uq_createModel(MOpts);
iOpts.Marginals(1).Type = 'Gaussian';
iOpts.Marginals(1).Moments = [0 1];
iOpts.Marginals(2).Type = 'Gaussian';
iOpts.Marginals(2).Moments = [0 1];

1 Like

Hi @aokoro

I see two ways to handle this:

1. Assume that X_3 follows a uniform distribution \mathcal{U}(3,8)
2. Perform two analyses: The first with x_3=3 and the second with x_3=8

Hope this helps

Thank you so much @paulremo for the response and explanation.