Abstract
AbstractWe present some new trends in fitness landscape analysis in evolutionary computation and meta-heuristic study. The fitness landscape concept is brought from that of evolutionary biology that attempts to present and visualize the relationship between genotype and its success. Evolutionary computation algorithm utilizes this concept to describe the success of an optimized variable. It takes a bridge between the evolutionary computation algorithm and its optimizing problem. We present three study subjects to analyse and model fitness landscape for evolutionary computation and meta-heuristic. They are fitness landscape approximation using functional regression in original space and projected lower-dimensional space, frequency information of fitness landscape using Fourier transform, and estimation of a convergence point from moving vectors in parameter space. The approximation and analysis of fitness landscape can be applied to enhance the search of evolutionary computation, to solve the multimodal optimization problem, and to accelerate the single-objective and multi-objective evolutionary search, etc. Some open topics, future opportunities, and works of new trends on fitness landscape are analysed and discussed as well.KeywordsEvolutionary computationMeta-heuristicFitness landscapeApproximationFourier analysisEstimation
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
