On some methods of parametrization of complex decay curves

Three methods of parametrization of complex decay curves are compared from the point of view of the correspondence of estimates obtained to the requirement of ''non-shift'' and efficiency. Pseudoexperimental curves were calculated on the ideal curve by the Monte-Carlo method and were processed at the fixed period values and the known real values of group outputs. A new program on the basis of an accelerated method of the least directed discrepancy is proposed. The estimates of parameters on this program correspond completely to the ''non-shift'' requirement and have a minimum selective dispersion, i.e. are the most effective ones