Local evolutionary search enhancement by random memorizing

For the calibration of laser induced plasma spectrometers robust and efficient local search methods are required. Therefore, several local optimizers from nonlinear optimization, random search and evolutionary computation are compared. It is shown that evolutionary algorithms are superior with respect to reliability and efficiency. To enhance the local search of an evolutionary algorithm a new method of random memorizing is introduced. This method is applied to one of the most simple evolutionary algorithm, the (1+1)-Evolution Strategy. It leads to a substantial gain in efficiency for a reliable local search. Finally, laser induced plasma spectroscopy and the calibration of a real example are sketched.