Probing the correlations in composite signals

The technique of degree of randomness is used to model the correlations in sequences containing various subsignals and noise. Kolmogorov stochasticity parameter enables to quantify the randomness in number sequences and hence appears as an efficient tool to distinguish the signals. Numerical experiments for a broad class of composite signals of regular and random properties enable to obtain the qualitative and quantitative criteria for the behavior of the descriptor depending on the input parameters typical to astrophysical signals.