eFAST Method

GlobalSensitivity.eFASTType
eFAST(; num_harmonics::Int = 4)
  • num_harmonics: the number of harmonics to sum in the Fourier series decomposition, this defaults to 4.

Method Details

eFAST offers a robust, especially at low sample size, and computationally efficient procedure to get the first and total order indices as discussed in Sobol. It utilizes monodimensional Fourier decomposition along a curve exploring the parameter space. The curve is defined by a set of parametric equations,

\[x_{i}(s) = G_{i}(sin ω_{i}s), ∀ i=1,2 ,..., N\]

where s is a scalar variable varying over the range $-∞ < s < +∞$, $G_{i}$ are transformation functions and ${ω_{i}}, ∀ i=1,2,...,N$ is a set of different (angular) frequencies, to be properly selected, associated with each factor for all $N$ (samples) number of parameter sets. For more details on the transformation used and other implementation details you can go through A. Saltelli et al..

API

gsa(f, method::eFAST, p_range::AbstractVector; samples::Int, batch = false,
         distributed::Val{SHARED_ARRAY} = Val(false),
         rng::AbstractRNG = Random.default_rng(), kwargs...) where {SHARED_ARRAY}

Example

Below we show use of eFAST on the Ishigami function.

using GlobalSensitivity, QuasiMonteCarlo

function ishi(X)
    A= 7
    B= 0.1
    sin(X[1]) + A*sin(X[2])^2+ B*X[3]^4 *sin(X[1])
end

lb = -ones(4)*π
ub = ones(4)*π

res1 = gsa(ishi,eFAST(),[[lb[i],ub[i]] for i in 1:4],samples=15000)

##with batching
function ishi_batch(X)
    A= 7
    B= 0.1
    @. sin(X[1,:]) + A*sin(X[2,:])^2+ B*X[3,:]^4 *sin(X[1,:])
end

res2 = gsa(ishi_batch,eFAST(),[[lb[i],ub[i]] for i in 1:4],samples=15000,batch=true)
source