The complete-basis-functions parameterization in ES and its application to laser pulse shaping
This paper presents a new parameterization method for the Evolution Strategies (ES) field, and its application to a challenging real-life high-dimensional Physics optimization problem, namely Femtosecond Laser Pulse Shaping. The so-called Complete-Basis-Functions Parameterization method (CBFP), to be introduced here for the first time, is developed for tackling efficiently the given laser optimization task, but nevertheless is a general method that can be used for learning any n-variables functions. The emphasis is on dimensionality reduction of the search space and the speeding-up of the convergence process respectively. This is achieved by learning the target function by using complete-basis functions as building blocks in an evolutionary search. The method is shown to boost the learning process of the given laser problem, and to yield highly satisfying results.
Shir, O. M, Bäck, Th, Siedschlag, Ch, & Vrakking, M. J. J. (2006). The complete-basis-functions parameterization in ES and its application to laser pulse shaping. In M Cattolico (Ed.), Genetic And Evolutionary Computation Conference : Proceedings of the 8th annual conference on Genetic and evolutionary computation, July 08 - 12, 2006 Seattle, WA, USA (pp. 1769–1776). ACM. doi:10.1145/1143997.1144284