ITU

Intrinsic Mortality Governs Evolvability

Research output: Conference Article in Proceeding or Book/Report chapterArticle in proceedingsResearchpeer-review

Standard

Intrinsic Mortality Governs Evolvability. / Veenstra, Frank; de Prado Salas, Pablo González; Bongard, Josh; Stoy, Kasper; Risi, Sebastian.

Artificial Life Conference Proceedings: ALIFE 2018. MIT Press, 2018. p. 242-249.

Research output: Conference Article in Proceeding or Book/Report chapterArticle in proceedingsResearchpeer-review

Harvard

Veenstra, F, de Prado Salas, PG, Bongard, J, Stoy, K & Risi, S 2018, Intrinsic Mortality Governs Evolvability. in Artificial Life Conference Proceedings: ALIFE 2018. MIT Press, pp. 242-249. https://doi.org/10.1162/isal_a_00050

APA

Veenstra, F., de Prado Salas, P. G., Bongard, J., Stoy, K., & Risi, S. (2018). Intrinsic Mortality Governs Evolvability. In Artificial Life Conference Proceedings: ALIFE 2018 (pp. 242-249). MIT Press. https://doi.org/10.1162/isal_a_00050

Vancouver

Veenstra F, de Prado Salas PG, Bongard J, Stoy K, Risi S. Intrinsic Mortality Governs Evolvability. In Artificial Life Conference Proceedings: ALIFE 2018. MIT Press. 2018. p. 242-249 https://doi.org/10.1162/isal_a_00050

Author

Veenstra, Frank ; de Prado Salas, Pablo González ; Bongard, Josh ; Stoy, Kasper ; Risi, Sebastian. / Intrinsic Mortality Governs Evolvability. Artificial Life Conference Proceedings: ALIFE 2018. MIT Press, 2018. pp. 242-249

Bibtex

@inproceedings{cdfc6ef3ca564dd9b3b6f2b94054eaf8,
title = "Intrinsic Mortality Governs Evolvability",
abstract = "An ongoing discussion in biology concerns whether intrinsicmortality, or senescence, is programmed or not. The death(i.e. removal) of an individual solution is an inherent featurein evolutionary algorithms that can potentially explain howintrinsic mortality can be beneficial in natural systems. Thispaper investigates the relationship between mutation rate andmortality rate with a steady state genetic algorithm that has aspecific intrinsic mortality rate. Experiments were performedon a predefined deceptive fitness landscape, the hierarchicalif-and-only-if function (H-IFF). To test whether the relationshipbetween mutation and mortality rate holds for more complexsystems, an agent-based spatial grid model based on theH-IFF function was also investigated. This paper shows thatthere is a direct correlation between the evolvability of a populationand an indiscriminate intrinsic mortality rate to mutationrate ratio. Increased intrinsic mortality or increased mutationrate can cause a random drift that can allow a populationto find a global optimum. Thus, mortality in evolutionaryalgorithms does not only explain evolvability, but might alsoimprove existing algorithms for deceptive/rugged landscapes.Since an intrinsic mortality rate increases the evolvability ofour spatial model, we bolster the claim that intrinsic mortalitycan be beneficial for the evolvability of a population.",
author = "Frank Veenstra and {de Prado Salas}, {Pablo Gonz{\'a}lez} and Josh Bongard and Kasper Stoy and Sebastian Risi",
year = "2018",
doi = "10.1162/isal_a_00050",
language = "English",
pages = "242--249",
booktitle = "Artificial Life Conference Proceedings",
publisher = "MIT Press",
address = "United States",

}

RIS

TY - GEN

T1 - Intrinsic Mortality Governs Evolvability

AU - Veenstra, Frank

AU - de Prado Salas, Pablo González

AU - Bongard, Josh

AU - Stoy, Kasper

AU - Risi, Sebastian

PY - 2018

Y1 - 2018

N2 - An ongoing discussion in biology concerns whether intrinsicmortality, or senescence, is programmed or not. The death(i.e. removal) of an individual solution is an inherent featurein evolutionary algorithms that can potentially explain howintrinsic mortality can be beneficial in natural systems. Thispaper investigates the relationship between mutation rate andmortality rate with a steady state genetic algorithm that has aspecific intrinsic mortality rate. Experiments were performedon a predefined deceptive fitness landscape, the hierarchicalif-and-only-if function (H-IFF). To test whether the relationshipbetween mutation and mortality rate holds for more complexsystems, an agent-based spatial grid model based on theH-IFF function was also investigated. This paper shows thatthere is a direct correlation between the evolvability of a populationand an indiscriminate intrinsic mortality rate to mutationrate ratio. Increased intrinsic mortality or increased mutationrate can cause a random drift that can allow a populationto find a global optimum. Thus, mortality in evolutionaryalgorithms does not only explain evolvability, but might alsoimprove existing algorithms for deceptive/rugged landscapes.Since an intrinsic mortality rate increases the evolvability ofour spatial model, we bolster the claim that intrinsic mortalitycan be beneficial for the evolvability of a population.

AB - An ongoing discussion in biology concerns whether intrinsicmortality, or senescence, is programmed or not. The death(i.e. removal) of an individual solution is an inherent featurein evolutionary algorithms that can potentially explain howintrinsic mortality can be beneficial in natural systems. Thispaper investigates the relationship between mutation rate andmortality rate with a steady state genetic algorithm that has aspecific intrinsic mortality rate. Experiments were performedon a predefined deceptive fitness landscape, the hierarchicalif-and-only-if function (H-IFF). To test whether the relationshipbetween mutation and mortality rate holds for more complexsystems, an agent-based spatial grid model based on theH-IFF function was also investigated. This paper shows thatthere is a direct correlation between the evolvability of a populationand an indiscriminate intrinsic mortality rate to mutationrate ratio. Increased intrinsic mortality or increased mutationrate can cause a random drift that can allow a populationto find a global optimum. Thus, mortality in evolutionaryalgorithms does not only explain evolvability, but might alsoimprove existing algorithms for deceptive/rugged landscapes.Since an intrinsic mortality rate increases the evolvability ofour spatial model, we bolster the claim that intrinsic mortalitycan be beneficial for the evolvability of a population.

U2 - 10.1162/isal_a_00050

DO - 10.1162/isal_a_00050

M3 - Article in proceedings

SP - 242

EP - 249

BT - Artificial Life Conference Proceedings

PB - MIT Press

ER -

ID: 83638175