Evaluate function handle for each value in vectorized form

I initialize a relatively simple function handle that depends on alpha. For simplicity, I will use alpha = 1.

For beta and gamma, I use n=10 random draws, where beta are scalars and gamma are 3 by 1 vectors.

I want to evaluate the function handle for each of the n random draws. This can be achieved e.g. by a for loop.

How can this be achieved in a more Matlab-like (vectorized) way?

Thanks.

alpha = 1;
n=10;

rng default
beta_all = randn(n, 1);
gamma_all = randn(3, n);

res = cell(n, 1);

for k=1:length(res)
    
    beta = beta_all(k);
    gamma = gamma_all(:,k);
    y = @(alpha) (alpha * beta * gamma);
    res{k} = y(alpha);
end

[Jan]

Is it really needed to create y?

res{k} = alpha * beta * gamma;

[Salomon Roth]

Yes, the y has to be created for my real world example, because sometimes I need to evaluate alpha, and sometimes alpha +/- something, which requires y.

[Jan]

Maybe the problem gets more clear, if you post the real code, not a strongly simplificated version. If you need "alpha +- something", you could use:

res{k} = (alpha + x) * beta * gamma;

[Salomon Roth]

Thanks Jan, you were right. Indeed, it is possible to proceed as you have outlined (alpha + x), i.e., the function handle is not necessary. However, I still wonder if the for loop is needed to evaluate the function for each of the n random draws?

Answer 1

[Jan]

alpha = 1;
n=10;
rng default
beta_all = randn(n, 1);
gamma_all = randn(3, n);

res = num2cell(alpha * beta_all.' .* gamma_all, 1).';