Comparing the Evolvability of Generative Encoding Schemes

Abstract

The evolvability of a system is the ability to generate heritable, novel and non-lethal phenotypes, from random genetic mutations. However, most evolutionary computation studies estimate evolvability either as, (i) the proportion of mutations beneficial to an individual's performance, irrespective of the phenotypic diversity of the mutants, or (ii) the range and diverseness of mutated phenotypes, without taking into account the viability of the genetic change. This paper reports a novel approach to measure the evolvability provided by an encoding, by characterizing both the quality of the mutations and the quantity of phenotypic variation. We evolved controllers for hexapod robot locomotion using a parameterized direct encoding, and the generative encoding of artificial neural networks (similar to HyperNEAT) and single-unit pattern generators (SUPGs). Our results reveal that the performance of an encoding is not always a good assessment of evolvability. Although both the generative encodings evaluated had individuals with high performance gaits, there were apparent differences in their measured evolvability. A direct and predictive relationship is indicated between our measure of evolvability, and the number of generations required by amputated individuals to recover an effective gait.

Reality-assisted evolution of soft robots through large-scale physical experimentation: a review

“…CPPN's have been combined with a various optimization algorithms that evolve their underlying network to design soft robots in a simulation environment [9], with some having been transferred to the real world (Fig 2g). Growing evidence has supported the idea that generative encoding schemes using CPPN's or similar outperform other approaches [94,120].…”

Evolvability signatures of generative encodings: Beyond standard performance benchmarks

“…Overall, we investigate five encodings for the classical ER problem of legged robot locomotion Lewis et al (1992); Gruau (1994); ; Clune et al (2009a); Bongard et al (2006); Clune et al (2009b; Yosinski et al (2011); Koos et al (2013); Lee et al (2013): (1) open-loop central pattern generator (CPG) evolved with a direct encoding, 1 A preliminary study on our approach to visualize evolvability is published in a conference paper Tarapore and Mouret (2014).…”

“…A preliminary study on our approach to visualize evolvability is published in a conference paperTarapore and Mouret (2014).…”