I am trying to scale my UQlab Inversion for larger problems with multiple (space dependent) outputs. I have stacked two output variables to end up with N_{outputs} = 1534.
Before doing any advanced analysis, I wanted to use the default settings of the Inversion tool to test before making any adjustments for advanced sampling. I have created a PCE model as my forward model with 1534 coefficients. All other settings are the default. However, when I try to start the inversion, I get the error “Number of dimensions must be a positive scalar integer between 1 and 1111”. The forward model itself is dependend on only 5 parameters.
How can I proceed with fixing this small error? When I try to look under the hood, I get val = 1539 line 197 of the qrandset function file. My assumption is, it is trying to sample 1539 parameters. I would appreciate any assistance to proceed properly. thank you.
Hi @Uche_A,
Could you paste your code that produces the error?
I am sure it is simply a dimension problem, but it would be easier to identify it with some code.
Thank you 
Gian
1 Like
I am sorry for the very late response, yes it was a dimension problem. in case anyone experiences this problem too.