Enhancing Divergent Search through Extinction Events

A challenge in evolutionary computation is to create representations as evolvable as those in natural evolution. This paper hypothesizes that extinction events, i.e. mass extinctions, can significantly increase evolvability, but only when combined with a divergent search algorithm, i.e. a search driven towards diversity (instead of optimality). Extinctions amplify diversity-generation by creating unpredictable evolutionary bottlenecks. Persisting through multiple such bottlenecks is more likely for lineages that diversify across many niches, resulting in indirect selection pressure for the capacity to evolve. This hypothesis is tested through experiments in two evolutionary robotics domains. The results show that combining extinction events with divergent search increases evolvability, while combining them with convergent search offers no similar benefit. The conclusion is that extinction events may provide a simple and effective mechanism to enhance performance of divergent search algorithms.

[1]  Ricard V. Solé,et al.  Macroevolutionary algorithms: a new optimization method on fitness landscapes , 1999, IEEE Trans. Evol. Comput..

[2]  V. Ramakrishnan,et al.  Measurement of the top-quark mass with dilepton events selected using neuroevolution at CDF. , 2008, Physical review letters.

[3]  X. Yao.  Evolving Artificial Neural Networks , 1999 .

[4]  M. Félix,et al.  Phenotypic neighborhood and micro-evolvability. , 2004, Trends in genetics : TIG.

[5]  Günter P. Wagner,et al.  Complex Adaptations and the Evolution of Evolvability , 2005 .

[6]  J. Brookfield.  Evolution: The evolvability enigma , 2001, Current Biology.

[7]  Stéphane Doncieux,et al.  Encouraging Behavioral Diversity in Evolutionary Robotics: An Empirical Study , 2012, Evolutionary Computation.

[8]  T. Metzinger.  The evolution of evolvability Ruth Garret Millikan Varieties of Meaning: The 2002 Jean Nicod Lectures , 2005, Trends in Cognitive Sciences.

[9]  Kenneth O. Stanley,et al.  Abandoning Objectives: Evolution Through the Search for Novelty Alone , 2011, Evolutionary Computation.

[10]  Kenneth O. Stanley,et al.  Evolvability Is Inevitable: Increasing Evolvability without the Pressure to Adapt , 2013, PloS one.

[11]  Garrison W. Greenwood,et al.  Evolutionary computation with extinction: experiments and analysis , 2000, Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512).

[12]  Kenneth O. Stanley,et al.  On the Benefits of Divergent Search for Evolved Representations , 2012 .

[13]  Jean-Baptiste Mouret.  Novelty-Based Multiobjectivization , 2011 .

[14]  Risto Miikkulainen,et al.  Evolving Neural Networks through Augmenting Topologies , 2002, Evolutionary Computation.

[15]  Antoine Cully,et al.  Robots that can adapt like animals , 2014, Nature.

[16]  Jean-Baptiste Mouret,et al.  Illuminating search spaces by mapping elites , 2015, ArXiv.

[17]  Inman Harvey,et al.  The artificial evolution of adaptive behaviour , 1994 .

[18]  Kenneth O. Stanley,et al.  Improving evolvability through novelty search and self-adaptation , 2011, 2011 IEEE Congress of Evolutionary Computation (CEC).

[19]  Antoine Cully,et al.  Robots that can adapt like natural animals , 2014, ArXiv.

[20]  M. Pigliucci.  Is evolvability evolvable? , 2008, Nature Reviews Genetics.

[21]  Petros Faloutsos,et al.  Complex networks of simple neurons for bipedal locomotion , 2009, 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[22]  Risto Miikkulainen,et al.  Active Guidance for a Finless Rocket Using Neuroevolution , 2003, GECCO.

[23]  Risto Miikkulainen,et al.  Solving Non-Markovian Control Tasks with Neuro-Evolution , 1999, IJCAI.

[24]  Hod Lipson,et al.  The evolutionary origins of modularity , 2012, Proceedings of the Royal Society B: Biological Sciences.

[25]  Risto Miikkulainen,et al.  Competitive Coevolution through Evolutionary Complexification , 2011, J. Artif. Intell. Res..

[26]  Andrew P. Martin.  Increasing Genomic Complexity by Gene Duplication and the Origin of Vertebrates , 1999, The American Naturalist.

[27]  T. Krink,et al.  Self-organized criticality and mass extinction in evolutionary algorithms , 2001, Proceedings of the 2001 Congress on Evolutionary Computation (IEEE Cat. No.01TH8546).

[28]  Stéphane Doncieux,et al.  Exploring new horizons in evolutionary design of robots , 2009 .

[29]  Kenneth O. Stanley,et al.  Evolving a diversity of virtual creatures through novelty search and local competition , 2011, GECCO '11.

[30]  Leslie G. Valiant,et al.  Evolvability , 2009, JACM.

[31]  L. Darrell Whitley,et al.  Delta Coding: An Iterative Search Strategy for Genetic Algorithms , 1991, ICGA.

[32]  Russell K. Standish,et al.  Open-Ended Artificial Evolution , 2002, Int. J. Comput. Intell. Appl..

[33]  Risto Miikkulainen,et al.  Incremental Evolution of Complex General Behavior , 1997, Adapt. Behav..

[34]  Marcus W Feldman,et al.  SPATIAL ENVIRONMENTAL VARIATION CAN SELECT FOR EVOLVABILITY , 2011, Evolution; international journal of organic evolution.