Effective fitness
In natural evolution and artificial evolution (e.g. artificial life and evolutionary computation) the fitness (or performance or objective measure) of a schema is rescaled to give its effective fitness which takes into account crossover and mutation.
Effective fitness is used in Evolutionary Computation to understand population dynamics.[1] While a biological fitness function only looks at reproductive success, an effective fitness function tries to encompass things that are needed to be fulfilled for survival on population level.[2] In homogeneous populations, reproductive fitness and effective fitness are equal.[1] When a population moves away from homogeneity a higher effective fitness is reached for the recessive genotype. This advantage will decrease while the population moves toward an equilibrium.[1] The deviation from this equilibrium displays how close the population is to achieving a steady state.[1] When this equilibrium is reached, the maximum effective fitness of the population is achieved.[3]
Problem solving with evolutionary computation is realized with a cost function.[4] If cost functions are applied to swarm optimization they are called a fitness function. Strategies like reinforcement learning[5] and NEAT neuroevolution[6] are creating a fitness landscape which describes the reproductive success of cellular automata.[7][8]
The effective fitness function models the number of fit offspring[1] and is used in calculations that include evolutionary processes, such as mutation and crossover, important on the population level.[9]
The effective fitness model is superior to its predecessor, the standard reproductive fitness model. It advances in the qualitatively and quantitatively understanding of evolutionary concepts like bloat, self-adaptation, and evolutionary robustness.[3] While reproductive fitness only looks at pure selection, effective fitness describes the flow of a population and natural selection by taking genetic operators into account.[1][3]
A normal fitness function fits to a problem,[10] while an effective fitness function is an assumption if the objective was reached.[11] The difference is important for designing fitness functions with algorithms like novelty search in which the objective of the agents is unknown.[12][13] In the case of bacteria effective fitness could include production of toxins and rate of mutation of different plasmids, which are mostly stochastically determined[14]
Applications
When evolutionary equations of the studied population dynamics are available, one can algorithmically compute the effective fitness of a given population. Though the perfect effective fitness model is yet to be found, it is already known to be a good framework to the better understanding of the moving of the genotype-phenotype map, population dynamics, and the flow on fitness landscapes.[1][3]
Models using a combination of Darwinian fitness functions and effective functions are better at predicting population trends. Effective models could be used to determine therapeutic outcomes of disease treatment.[15] Other models could determine effective protein engineering and works towards finding novel or heightened biochemistry.[16]
References
- ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 ""Effective" fitness landscapes for evolutionary systems". Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406). 1999. pp. 703–714. doi:10.1109/CEC.1999.782002. ISBN 0-7803-5536-9.
- ↑ "Effects of stochasticity and division of labor in toxin production on two-strain bacterial competition in Escherichia coli". PLOS Biology 15 (5): e2001457. May 2017. doi:10.1371/journal.pbio.2001457. PMID 28459803.
- ↑ 3.0 3.1 3.2 3.3 "Effective Fitness as an Alternative Paradigm for Evolutionary Computation I: General Formalism". Genetic Programming and Evolvable Machines 1 (4): 363–378. 2000. doi:10.1023/A:1010017207202.
- ↑ "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 - GECCO 09. ACM Press. 2009. doi:10.1145/1570256.1570378.
- ↑ "Optimization with auxiliary criteria using evolutionary algorithms and reinforcement learning". Proceedings of 18th International Conference on Soft Computing MENDEL 2012. 2012. 2012. pp. 58–63.
- ↑ "The Effect of Fitness Function Design on Performance in Evolutionary Robotics". Proceedings of the 2015 on Genetic and Evolutionary Computation Conference - GECCO 15. ACM Press. 2015. doi:10.1145/2739480.2754676.
- ↑ "Landscapes and Effective Fitness". Comments on Theoretical Biology (Informa UK Limited) 8 (4–5): 389–431. 2003. doi:10.1080/08948550302439. http://ul.qucosa.de/id/qucosa%3A31924.
- ↑ Bagnoli F (1998). "Cellular automata". arXiv:cond-mat/9810012.
- ↑ "φ-evo: A program to evolve phenotypic models of biological networks". PLOS Computational Biology 14 (6): e1006244. June 2018. doi:10.1371/journal.pcbi.1006244. PMID 29889886. Bibcode: 2018PLSCB..14E6244H.
- ↑ Creating a fitness function that is the right fit for the problem at hand. 2017.
- ↑ "Fitness function for finding out robust solutions on time-varying functions". Proceedings of the 8th annual conference on Genetic and evolutionary computation GECCO 06. ACM Press. 2006. doi:10.1145/1143997.1144186.
- ↑ "Abandoning objectives: evolution through the search for novelty alone". Evolutionary Computation (MIT Press - Journals) 19 (2): 189–223. 2011. doi:10.1162/evco_a_00025. PMID 20868264.
- ↑ Woolley BF, Stanley KO (2012). "Exploring promising stepping stones by combining novelty search with interactive evolution". arXiv:1207.6682 [cs.NE].
- ↑ "Abandoning objectives: evolution through the search for novelty alone". Evolutionary Computation 19 (2): 189–223. 2010-09-24. doi:10.1162/EVCO_a_00025. PMID 20868264.
- ↑ "Phenotypic heterogeneity in modeling cancer evolution". PLOS ONE 12 (10): e0187000. 2017-10-30. doi:10.1371/journal.pone.0187000. PMID 29084232. Bibcode: 2017PLoSO..1287000M.
- ↑ "On simplified global nonlinear function for fitness landscape: a case study of inverse protein folding". PLOS ONE 9 (8): e104403. 2014-08-11. doi:10.1371/journal.pone.0104403. PMID 25110986. Bibcode: 2014PLoSO...9j4403X.
External links
 |  |
