Matlab wrong:number of permutations too high: would require 1.859578e+04 MB, while the specified max is 8192 MB

% for the coefficient calculation strategy:
MetaOpts.Method = 'LARS';  
% Specify a sparse truncation scheme (hyperbolic norm with $q = 0.75$):
MetaOpts.TruncOptions.qNorm = 0.75; 
% Specify the range of the degrees to be compared
% by the adaptive algorithm:
MetaOpts.Degree = 2:10;
%%
% UQLab will automatically determine the appropriate quadrature level.
    X_ED = X;
    % Evaluate the full model on the current experimental design
    Y_ED = Y;
    % Use the sample as the experimental design for the PCE
    MetaOpts.ExpDesign.X = X_ED;
    MetaOpts.ExpDesign.Y = Y_ED;
% Create the quadrature-based PCE metamodel:
myPCE = uq_createModel(MetaOpts);

Dear professors, when I was creating the agent model with UQlab, I encountered an error from matlab. The error command is: myPCE = uq_createModel(MetaOpts) .Is it because I have too many variables? My variables are 217. Matlab Error:number of permutations too high: would require 1.859578e+04 MB, while the specified max is 8192 MB.
How can I solve it? Thank you

Dear @aishenma

I assume that the error occurs when UQLab calculates the permutations needed to build the polynomial basis. So you need to either reduce the number of permutations needed or reduce the dimensionality of the input. You can try to set MetaOpts.TruncOptions.MaxInteraction = 2 to reduce the interaction, because the q-norm truncation is not sufficient for this high number of input variables. You may also want to have a look at the section about truncation in the manual:

Best regards
Styfen