dot indexing in a function handle

Hi all,

Is it possible to use dot indexing in a function handle? Or at least ignore an output?

Please look at the code below.

clear
clc
close

format shortG

P = 100;
vel = 1.5066e+03;
omegab = 0;
L = 30480;
E = 200000;
I = 1.6253e6;
b = 2438;
h = 20;
A = b * h;
rho = 7.8E-9;
jmax = 5;

varMean = [P vel];
varStd = abs([0.15 0.15] .* varMean);
varDist = ["normal","normal"];

g = @(X) 1.76 - beamUnderMovingLoad(X(1),X(2),L,E,I,omegab,rho,A,jmax);

inputs = struct('means', varMean, 'stds', varStd, 'dists', varDist, 'g_func', g, 'N', 10000);

output = monteCarloSimulation(inputs)

output.Pf
output.Beta

It runs properly when the function 'beamUnderMovingLoad()' has only one output, but Id like to select the output in the 'g' function.

Trust all clear!

Answer 1

[Rik]

You will have to write a wrapper that can select an output for you. There is no built-in functionality to do so.

Some like this should work:

function out=output_selector(fun,k,varargin)
% Select the k-th output variable given some input.
% The first input should be a function handle. Any inputs required for the
% function can provided as extra arguments.
out = cell(1,k);
[out{:}]=fun(varargin{:});
out = out{end};
end

[Geovane Gomes]

Thanks Rik, but it did not work.

I'd like to select for example the 2nd output of 'beamUnderMovingLoad()' to use in the 'g' function handle

[Rik]

What code did you write?

[Geovane Gomes]

I edited this line

g = @(X) 1.76 - output_selector(beamUnderMovingLoad(X(1),X(2),X(3),E,I,omegab,rho,A,jmax),2,varargin)

[Rik]

It might have helped if you had read the documentation for varargin, which might have given you the idea for how to use it.

The problem is that you provided the output of the function, not the function itself to output_selector.

The line below should do the trick.

g = @(X) 1.76 - output_selector(@beamUnderMovingLoad,2,X(1),X(2),X(3),E,I,omegab,rho,A,jmax)

[Geovane Gomes]

This way did not work also.

Note the error message: Operator '-' is not supported for operands of type cell.

If I remove the term '1.76 -', it shows 'Conversion to double from cell is not possible'.

Both errors happen when 'g_func' is evaluated in the monteCarloSimulation function

Here is the line:

X = randomNumberGenerator(N, n_vars, dists, means, stds);
G = zeros(N, 1);                % Pre-allocate memory for the limit state function
for i = 1:N
    G(i) = g_func(X(i,:));      % Evaluate limit state function for each sample
end

[Rik]

Ah, that is the downside of writing that function on mobile. It should have been out = out{end}; instead of out = out(end);.

I'll edit my answer.

[Geovane Gomes]

Wow!!!

Now it works perfect!

i really appreciate your help

[Rik]

You're welcome