What is A-Life? Simply put, Artificial Life is a broad
field of work in which attempts are made to simulate or recreate one
or more natural processes. Most A-Life work is done on computers, but
in some cases it's done in the real world with chemicals or just on
paper. Some examples of natural systems that are often modelled are:
predator/prey balances, darwinistic evolution, neural
networks, behaviour and learning, ecological interactions, flocking,
self organization, emergent behaviour or complexity, reproduction,
and genetics, to name a few. A-Life's practical purpose can be two-fold, to
use technology to understand and improve biological sciences, and
to use biological understanding to improve technological science.
But beyond the practical, insights into life, nature, and the universe
can be gleamed even from the most simple model, experiment, or simulation.
Ed T. Toton III
- System Administrator / Programmer
- Long-time A-Life enthusiast
- Self-Organized Emergent Philosophical Process :)
Since the summer before I entered college, artificial life (or "A-Life") has
been one of my interests. In general, I find all of the "fringe"
computer topics interesting, including fractals, AI, digital
music, 3D graphics, etc. But A-Life holds a certain appeal. There's
something particular compelling about the concept of having life
exist within the computer, something that evolves and develops
it's own behaviour or morphology, something that goes beyond the
initial design created by the programmer.
Artificial life takes on many forms, and different people approach
it from different viewpoints and with different goals. For some
people, it is a scientific tool, to help understand the natural
processes in the natural world. Darwinistic evolution and
predator/prey balances are two examples of such processes that
can be simulated and observed in a controlled environment. In some
cases, it's used to study emergent behaviour, or emergent complexity
from a set of simple rules (a trait that fractals also exhibit).
To others, it's more of an artistic endeavor. Some A-Life projects
have been developed to be a means of genetically generating
pleasing patterns and designs (Biomorphs have been used for this).
L-Systems can be used to model the external morphology
and appearance of plants. And in some cases (such as some of my
programs), they're done for the sake of creating something that
takes on "a life of its own", and they can be interesting to watch.
And for others, a-life is for entertainment. Virtual pets and
'games' in which you take an active role in the development of
the organisms, or populations of organisms can fall into this
category (such as SimAnt, Creatures, Tamagotchi, Dogz/Catz, etc).
|
Despite the wide-range of practical applications (or even impractical
applications), it should not be overlooked that A-Life is indeed a
legitimate field of study, and serious scientific research is being
done in the field. One that is only recently receiving wider attention,
perticularly with the continuing increase in available computer power.
Many A-Life experiments fall into one category, and typically
only model one or two specific processes or traits of individual
organisms or populations of organisms. Often they can also be
described in terms of other properties of the simulation- For
instance, some simulations simulate the interactions between
multiple individuals that are created on a high-level (such as
BugFest), wheras others attempt to
create a world with its own rules on a lower level, in which the
organisms are constructed from components of that world (such as
Tom Ray's Tierra
or Helix-4).
While my personal interest doesn't specifically exclude any
particular form of A-Life, my efforts typically fall within
the entertainment/artistic range, since most of my projects
are more for my own enjoyment, and for sharing the wonder
that such things can inspire in people, though I do attempt to
make them valid experiments in emergent behaviour, genetic
evolution, or other natural processes. I've made a variety
of types of A-Life programs, only the better of which are
available here (meaning they're aesthetically acceptible,
relatively bug-free, fairly bullet-proof, and interesting on
some level).
Although I tend to release programs of this nature very
infrequently, A-Life is the one field of interest that I can
never seem to stray very far from. Most of what I've made so
far has been fairly simplistic in many ways (though some of the
technical hurdles to keep program performance high were not
insignificant), but I have ideas floating in my head all the time.
There will undoubtedly be more life springing forth from this
website sooner or later. :-)
-Bones
Click here for the old A-Life comments
|
My A-Life Experiments:
- AT Robots 2 -
Not really an A-Life project, but related and very interesting. This is a
game in which you write a program to control a robot, and pit it against
other robots in a simulated arena. I've had an online robot server running
for several years, where you can test your robots against the current champions.
- Unnamed Project -
This is an A-Life project written completely in perl, only calling out to
one ImageMagick program to convert the graphical portion of the output to
a browser-compatabile image format. It's an online simulation, and the
CGI script that generates the page you view predicts the weather patterns
independantly (capable of forecasting with 100% accuracy to an arbitrary distance
into the future). It's the classic "Living vs Undead" struggle, competing for
control of territory, running as a cellular automata model with a multi-sinusoidal
weather generator.
- BugFest-1 - A graphically aesthetic
predator/prey balance simulation using insects.
- Helix-4 - This was an attempt
at a Tierra-like A-Life project. Self replicating programs compete
for space in the memory pool. The program/interface is a little
clunky, the instruction set is a little fragile, and I'd do things
differently now, but the program worked well and even included a
genebanker and creature viewer.
- Helix-1 - My very first attempt at
Artificial Life. It doesn't really demonstrate anything, except
how darwinism can improve efficiency and nothing more when there's
no possible open-ended evolution. The interface is nice though. :)
A-Life Links:
|
