Genetic algorithms in plasma diagnostic analysis

Abstract

A novel sophisticated technique for data reduction utilising Bayesian statistics and a genetic algorithm has been developed by the authors. The technique gives superior signal recovery in poor signal-to-noise ratio conditions compared to simple linear least-squares methods. In this paper, the method is discussed along with a detailed case study: the Thomson scattering diagnostic of a magnetically confined plasma.