Fitness landscape analysis for evolutionary non-photorealistic rendering

The best evolutionary approach can be a difficult problem. In this work we have investigated two evolutionary representations to evolve non-photorealistic renderings: a variable-length classic genetic algorithm representation, and a tree-based genetic algorithm representation. These representations exhibit very different convergence behaviour, and despite considerable exploration of parameters the classic genetic algorithm was not competitive with the tree-based approach for the problem studied in this work. The aim of the work presented in this paper was to investigate whether analysis of the fitness landscapes described by the different representations can explain the difference in performance. We used several current fitness landscape measures to analyse the fitness landscapes, and found that one of the measures suggests there is a correlation between search performance and the fitness landscape.

[1]  Andrew M. Sutton,et al.  Understanding elementary landscapes , 2008, GECCO '08.

[2]  Julian Francis Miller,et al.  Information Characteristics and the Structure of Landscapes , 2000, Evolutionary Computation.

[3]  John R. Koza,et al.  Genetic programming - on the programming of computers by means of natural selection , 1993, Complex adaptive systems.

[4]  Günther R. Raidl,et al.  Empirical Analysis of Locality, Heritability and Heuristic Bias in Evolutionary Algorithms: A Case Study for the Multidimensional Knapsack Problem , 2005, Evolutionary Computation.

[5]  Bernard Manderick,et al.  The Genetic Algorithm and the Structure of the Fitness Landscape , 1991, ICGA.

[6]  Lee Altenberg,et al.  The Schema Theorem and Price's Theorem , 1994, FOGA.

[7]  Inman Harvey,et al.  Proceedings of the Fourth European Conference on Artificial Life , 1997 .

[8]  P. Stadler.  Landscapes and their correlation functions , 1996 .

[9]  Marc Lipsitch,et al.  Adaptation on Rugged Landscapes Generated by Iterated Local Interactions of Neighboring Genes , 1991, ICGA.

[10]  Phil Husbands,et al.  Fitness Landscapes and Evolvability , 2002, Evolutionary Computation.

[11]  Terry Jones,et al.  Fitness Distance Correlation as a Measure of Problem Difficulty for Genetic Algorithms , 1995, ICGA.

[12]  Riccardo Poli,et al.  Information landscapes , 2005, GECCO '05.

[13]  J. Wesley Barnes,et al.  Some additional properties of elementary landscapes , 2009, Appl. Math. Lett..

[14]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[15]  Weinberger,et al.  Local properties of Kauffman's N-k model: A tunably rugged energy landscape. , 1991, Physical review. A, Atomic, molecular, and optical physics.

[16]  Bernd Freisleben,et al.  Fitness landscapes and memetic algorithm design , 1999 .

[17]  Peter Merz,et al.  Memetic algorithms for combinatorial optimization problems : fitness landscapes and effective search strategies , 2006 .

[18]  E. Weinberger,et al.  Correlated and uncorrelated fitness landscapes and how to tell the difference , 1990, Biological Cybernetics.

[19]  J. Wesley Barnes,et al.  The theory of elementary landscapes , 2003, Appl. Math. Lett..

[20]  Wim Hordijk,et al.  A Measure of Landscapes , 1996, Evolutionary Computation.

[21]  Ernesto Costa,et al.  Multidimensional Knapsack Problem: A Fitness Landscape Analysis , 2008, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[22]  Terry Jones,et al.  One Operator, One Landscape , 1995 .

[23]  Riccardo Poli,et al.  Information landscapes and the analysis of search algorithms , 2005, GECCO '05.

[24]  Colin R. Reeves.  Fitness Landscapes and Evolutionary Algorithms , 1999, Artificial Evolution.

[25]  Riccardo Poli,et al.  Information landscapes and problem hardness , 2005, GECCO '05.