Talk:Effective fitness

Draft for updating the article

Prose text

Problem solving with evolutionary computation is realized with a cost function.[7] If cost functions are applied to swarm optimization they are called a fitness function. Strategies like reinforcement learning[10] and NEAT neuroevolution[3] are creating a fitness landscape which describes the reproductive success of cellular automata.[1][5]

A normal fitness function fits to a problem[9], while an effective fitness function is an assumption if the objective was reached.[8] The difference is important for designing fitness function with algorithm like novelty search in which the objective of the agents is unkown.[2] [6]

Notes

• effective fitness is located in evolutionary computation [1]
• Fitness = reproductive success [1]
• Fitness landscapes is a set of fitness values connected configuration space [1]
• is effective fitness a measurement for the entire fitness landscape? [1]
• swarm e.g. Flocks of birds
• "Effective fitness landscape" "Effective fitness function" “Effective fitness models”
• evolutionary computation measures with a fitness function if the objective was reached [2] Problem is, if the : Objective functions is wrong
• Fitness Function Design with NEAT [3] -> Evolutionary robotics
• effective fitness landscape -> symmetry breaking [4]
• effective fitness is a plot in a diagram to describe the fitness of cellular automata [5] it stays in contrast to static fitness
• assumption: objective function is unknown, optimizing not for predefined goals but do a "novelty search" [6]
• evolutionary computation designs effective fitness functions [7]
• expected fitness function = Effective Fitness Function [8]
• fitness function must fit to a problem [9]
• Reinforcement Learning can search for the best (=effective) fitness function [10]
• is "effective fitness function" = "cost function"?

additional notes

• open:[4]

Literature

• [1] Stadler, Peter F., and Christopher R. Stephens. "Landscapes and effective fitness." Comments® on Theoretical Biology 8.4-5 (2003): 389-431.
• [2] Lehman, Joel, and Kenneth O. Stanley. "Abandoning objectives: Evolution through the search for novelty alone." Evolutionary computation 19.2 (2011): 189-223.
• [3] Divband Soorati, Mohammad, and Heiko Hamann. "The effect of fitness function design on performance in evolutionary robotics: The influence of a priori knowledge." Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation. 2015.
• [4] Stephens, Christopher R. "Effect of mutation and recombination on the genotype-phenotype map." arXiv preprint nlin/0006051 (2000).
• [5] Bagnoli, Franco. "Cellular automata." arXiv preprint cond-mat/9810012 (1998).
• [6] Woolley, Brian G., and Kenneth O. Stanley. "Exploring promising stepping stones by combining novelty search with interactive evolution." arXiv preprint arXiv:1207.6682 (2012).
• [7] Schaffer, J. David, Heike Sichtig, and Craig Laramee. "A series of failed and partially successful fitness functions for evolving spiking neural networks." Proceedings of the 11th Annual Conference Companion on Genetic and Evolutionary Computation Conference: Late Breaking Papers. 2009.
• [8] Handa, Hisashi. "Fitness function for finding out robust solutions on time-varying functions." Proceedings of the 8th annual conference on Genetic and evolutionary computation. 2006.
• [9] Fernandez, Aaron Carl T. "Creating a fitness function that is the right fit for the problem at hand." (2017).
• [10] Afanasyeva, Arina, and Maxim Buzdalov. "Optimization with auxiliary criteria using evolutionary algorithms and reinforcement learning." Proceedings of 18th International Conference on Soft Computing MENDEL 2012. Vol. 2012. 2012.

— Preceding unsigned comment added by ManuelRodriguez (talk • contribs) 09:26, 9 March 2020 (UTC) --ManuelRodriguez (talk) 09:06, 26 March 2020 (UTC) updated --ManuelRodriguez (talk) 07:06, 22 April 2020 (UTC)