VeNT said:
upload it so we can all read it and randomly quote it in replys to you.
I'll try and ipload it later, but if anyone is really interested here is the abstract and introduction:
Abstract:
A spatially-dispersed GA with co-evolutionary methodology was developed to
artificially evolve temporal-parameters for a spiking neural-model of the cricket auditory system
capable of performing phonotaxis. Male chromosomes containing genes that encode for the
temporal properties of calling songs were simultaneously evolved in the co-evolutionary model.
The application of A.I. modelling to the evolution of cricket species and their mating behaviour is
reviewed. The GA model produced discrete spatial groupings of individuals, which had distinct
genetic code within the male and female chromosomes. Networks with neural-parameters set
by the female chromosome’s genes showed a higher phonotactic performance when responding
to songs produced by males within that group than to songs produced by males from other
groups, supporting conspecific preference of calling song. However, this e ect varied greatly
between groups and trials. The algorithm’s behaviour is complex, dynamic and chaotic, with
highly dimensional data necessitating complex analysis. The resulting analysis does not provide
a clear or concise synopsis of the behaviour and has left some open questions that would require
further research.
1. Introduction
1.1 Project Outline
The primary aim of this project was to develop a ‘leaky integrate-and-fire’ spiking neural network
(NN) algorithm, as well as a Genetic Algorithm (GA) capable of evolving neuroethologically
inspired networks for simulated cricket phonotaxis- the behaviour of female crickets to acoustically
locate a proximal calling male for breeding purposes. The evolved NNs are designed to
control a biorobotic female cricket model that has previously been developed [37, 35], but the
hardware version could not be implemented within the scope of this project.
Biorobotics often studies and models insects as the basis of research, mainly because they
are relatively simple in both physiology and in neurology, and yet display complex and adaptive
behaviour which is not yet fully understood. They are comparatively easier to model mechanically
in a robot and have simpler sensory systems than other organims (e.g. insect eyes are less
developed than mammalian eyes), and they posses limited neurological capacity. This makes
them a realistic model for biorobotic researchers.
A GA is a heuristic used to find approximate solutions to dicult-to-solve problems through
the application of principles of evolutionary biology to computer science. GAs use biologicallyderived
techniques such as inheritance, mutation, natural selection and recombination (crossover),
and are a particular class of evolutionary algorithms [49]. GAs and evolutionary methods are
vital for the production of complex neural-circuitry and robotic controllers because of their inherent
complexity. The vastly complex and dynamic relationships between neural circuits and
the environment they inhabit make them extremely difficult to design by hand [10, 30]. GAs
are an excellent method for optimisation problems in complex and noisy search-spaces where
the search-space is poorly understood, and are commonly used to fine-tune systems with a large
number of parameters [49, 16, 27, 28]. Furthermore, GAs can be used to study the processes of
evolution itself: population dynamics, the effects of sexual selection, importance of mutation,
coevolution [15, 2], adaption, and convergence and divergence of species [12, 28, 19]. Kortmann
and Hallam [25] have suggested that a controller created by evolutionary means could indeed
account for evolutionary adaptation of behaviour and divergence observed in a natural system,
and particularly with robotic models of cricket phonotaxis.
The GA used in this project differs from the canonical form [16, 27] in that the chromosomes
have spatial representations and mating strategies [15, 8, 39, 34]. These were used to investigate
the dynamics of evolution such as co-evolution [15, 2, 18, 48], speciation and niching [6, 7, 39],
the emergence of differing calling song patterns, the coevolution of male song production and
female response behaviours, and interspecies interaction [1]. Known evolutionary cricket data
[31, 1, 26, 43, 20] is used to support these results.
)
)
