Genetic algorithms for positron lifetime data

Abstract

Recently, genetic algorithms have been applied for ultrafast optical spectrometry in systems with several convoluted lifetimes. We apply these algorithms and compare the results with POSFIT (by Kirkegaard and Eldrup) and LT programme (by Kansy). The analysis was applied to three types of samples: molybdenum monocrystals, Czochralski-grown silicon with oxygen precipitates, Si with under-surface cavities obtained by He + H ion co- implantation. In all three tests, the genetic algorithm performs very well, in particular for short lifetimes. Further developments to model the resolution function in genetic algorithms are needed.