1. But

Generate a trajectory of a mono-dimensional multi-variate stochastic process (i.e. with several components and indexed to a single variable), stationary, with zero mean from its spectral density matrix (interspectral matrix). The trajectories of the stochastic process are temporal functions that make it possible to then perform a transitory dynamic calculation.

The trajectories obtained have an interspectral matrix that converges on average to the target interspectral matrix and are the realizations of an asymptotically Gaussian process (i.e. when the number of draws tends to infinity). The algorithm used is a simulation algorithm using trigonometric series with random phase and inverse fast Fourier transform.

The temporal functions produced are deposited in an interspectrum data structure. Each function is identified by its serial number. It can be extracted using the RECU_FONCTION operator [U4.32.03].