The Second Harmonic Generation (SHG), a process that turns out to be a good test case in the physics lab, can also be considered as a fairly simple theoretical test function for global optimization. Despite its symmetry properties, that will be derived here analytically, it seems to capture the complexity of the Fourier transform between the decision space to the evaluation space, and by that to challenge optimization routines. And indeed, counter-intuitively to some extent, locating its global maximum seems to be not an easy task for Evolutionary Algorithms (EAs).
Although this research originates from the real-world applications domain, it aims to introduce a theoretical test case to Evolution Strategies (ES), being a possible theoretical gateway to the real-world physics regime of quantum control problems. After presenting some theoretical results, this paper introduces the study of the scalability of the decision space subject to optimization by specific variants of Derandomized Evolution Strategies. We show that the Evolution Strategy in use requires a quasi-quadratic increase of function evaluations for locating the global maximum as the dimensionality increases.
[1]
Marcos Dantus,et al.
Systematic control of nonlinear optical processes using optimally shaped femtosecond pulses.
,
2005,
Chemphyschem : a European journal of chemical physics and physical chemistry.
[2]
D. Fogel.
Evolutionary algorithms in theory and practice
,
1997,
Complex..
[3]
K. W. Cattermole.
The Fourier Transform and its Applications
,
1965
.
[4]
Nikolaus Hansen,et al.
Step-Size Adaption Based on Non-Local Use of Selection Information
,
1994,
PPSN.
[5]
Thomas Bäck,et al.
Evolutionary algorithms in theory and practice - evolution strategies, evolutionary programming, genetic algorithms
,
1996
.
[6]
Nikolaus Hansen,et al.
Completely Derandomized Self-Adaptation in Evolution Strategies
,
2001,
Evolutionary Computation.
[7]
Ofer M. Shir,et al.
Evolutionary Algorithms in the Optimization of Dynamic Molecular Alignment
,
2006,
2006 IEEE International Conference on Evolutionary Computation.