A fast algorithm for multi-exponential analysis of time-resolved frequency-domain data

In this paper, we propose a new algorithm for analysis of variable-frequency phase shift dependencies in terms of multi-exponential model reformulated for frequency domain. The developed routine ensures fast speed of processing and acceptable accuracy of parameter estimation. A substantial advantage of the proposed algorithm is its independence on prior selection of initial guesses for fit parameters. This allows using the developed data analysis algorithm for generating start values for iterative fit procedures thus making their performance more reliable and faster. Results of simulation experiments are presented to demonstrate the possibilities of the proposed approach.