To: All Msg #277, Jul1593 10:07AM Subject: Evolution simulator available Fellow t.o reader
From: Alan Geist
To: All Msg #277, Jul-15-93 10:07AM
Subject: Evolution simulator available
Organization: Supercomputer Systems Division (SSD), Intel
From: email@example.com (Alan Geist)
Fellow t.o readers,
I have uploaded a program for PCs that while simple, illustrates
some of the basic concepts of evolution. The ftp site is
wuarchive.wustl.edu, and the path and name are pub/goe.tar.Z.
Not being a biologist, parts of it are certainly going to be
innacurate. Hopefully, as many parts are at least somewhat
The reason I am posting notice here is that I am interested in
feedback from those who are far more knowing in the subjects of
evolution and biology than I am.
So if you have a PC and some time, give it a try, and let me
know what you think. My email address appears in the 'about'
file, which is part of the game. The about file is also
included below so you can see what it's all about before
bothering to download it.
Thanks in advance,
"Why do today what you can put off again tomorrow?"
The Game of Evolution
Copyright 1991, 1992, 1993 Alan Geist
Permission for reproduction or distribution is granted to all provided
that the files 'goe.hlp' and 'goe.abo' always accompany the compiled
Disclaimer: The opinions expressed in this file, the 'goe.hlp' file, and
the game are accidental, and thus are not those of either the author, or
any employer, past, present, or future. Furthermore, this game is not
meant to either support or discredit any scientific, political, or
social theory whatsoever.
The Game of Evolution (GOE) is a highly modified version of the
familiar Game of Life (GOL). The screen is made up of a grid of
"sites", each of which may be occupied by a living, but non-mobile
"cell". In GOE, like GOL, the cells survive and reproduce according
to certain rules.
The first difference between GOE and GOL is that in GOE the rules of
survival and reproduction are specified in probabilities instead of
absolutes. The probability is specified for each number of neighbors
a site can have. For example, if a 50% probability is specified for
a birth into an empty site with three neighbors, then about 50% of
the empty sites with three neighbors on the screen will have new
cells in them in a given generation. Each site is evaluated randomly
according to the probability, and independant of the evaluation of
all other sites on the screen. The user may specify those
probabilities, or use the defaults.
In addition, GOE has different species groups, and different
environments. Species group 1 is said to represent all species whose
relative success rate is "good" in environment 1. (For the purposes
of this game and description, consider the word "species" to mean
species group). Likewise, species 2 is said to represent all species
whose relative success rate is good in environment 2, and so on up to
species 4. Each of species 1 through 4 has a "good" success rate in
the environment by the same name, a "fair" success rate in the two
adjacent environments, and a "poor" success rate in the environment
which is diagonally opposite on the screen. Species 5 is said to
have a "poor" success rate in any environment, and is thus considered
to be a non-viable species.
Reproduction of one species over another is decided by the neighbors
of the empty site. For any given empty site, each neighbor is given
the opportunity to reproduce into that empty site in a random order.
So in an empty site with a variety of species of neighbors, each
neighbor gets a turn to give birth into that site, until a birth
occurs, or until all neighbors have had their opportunity. The
probability that a particular neighbor will give birth depends on the
success rate for that species, given the number of neighbors of the
empty site, and the environment that the particular neighbor is in.
Also, for the purposes of birth only, the probability is "degraded"
such that after checking all neighbors of an empty site, the
probability that a birth occured is equal to the number specified for
that species (assuming all neighbors were of the same species). In
other words, if an empty site has a 50% chance of getting a birth in
it for three neighbors, the chance for each of the three individuals
to give birth there is degraded such that the chance for a birth
after all three have been checked is 50%. If the chance wasn't
degraded, the chance after all three have been checked would be
(To understand this, flip a coin. The chance that the coin came up
heads is 50%. Flip it again. The chance that one of the two flips
was heads is 75%. Flip it again. The chance that one of the three
flips was heads is 87.5%. Etcetera. In the birth example above, the
coin has been weighted, such that after three flips, the chance that
one of the three flips was heads is 50%.)
Birth chance degredation depends only on the total number of
neighbors, and the relative success rate of the species trying to
give birth. It does not depend on the species of the other
The last difference between GOE and GOL is that GOE has evolution.
On a random basis, mutations will occur during the birth of a new
cell. A mutation will result in the "child" cell being of a
different species than the "parent" cell. To aid in telling the
difference between different mutations to the same species group, the
color rotates through a series of nine different colors each time a
new mutation occurs into a given species group. Thus, a blue species
1 and a green species 1 are said to be different species, but with
the same relative success rate in environment one.
Both the rate of mutation, and the chance that a given mutation will
be viable can be specified by the user. Non-viable mutations
automatically result in species 5. A viable mutation results in
species 1 through 4, and is independant of the environment. That is,
a mutation is considered viable, even if it has a poor success rate
in the environment into which it was born.
To see how species concentrate in their best environments, initialize
the screen with a variety of species scattered across the entire
screen. Within 100 generations or so (with default parameters), you
will see this happen. I recommend this your first time, as it shows
some of the effects that happen aside from evolution.
To actually watch evolution in action (other than watching bungie
jumping of course), intialize the whole screen with just one
species. After an unpredictable number of generations (but usually
less than 1000), at least one of the environments will probably have
a species that evolved especially for it.
There is one extremely interesting phenomenon which I have yet to be
able to explain. It seems that different species with identical
characteristics seem to segregate from one another, simply by being
labeled differently. This occurs independantly of evolution, and is
readily replicable with the probability of mutation equal to zero.
Further, there is nothing I know of in the program to cause this
effect other than a species only gives birth to an exact copy of
itself (assuming no mutation).
To observe this phenomenon, enter identical reproduction and
sustaining characteristics for each of the three success levels. When
prompted for the chance of a birth being a mutant, enter zero. At
this point all species should have identical characteristics
regardless of where they are located. Now initialize the screen with
a random mix of species. Within 100 generations or so, they will
have segregated into colonies, and even appear to sort of war against
Comments, suggestions, and explanations for the above phenomenon are
encouraged via email. The author's email address, as of July, 1993,
goe.exe is a PC executable file, though admittedly, it has not been
tried on a wide variety of machines prior to distribution.
Executables for other machines are not available, so please don't
Enjoy your game. If you are viewing this from within the game, the
help file is automatically displayed next.
E-Mail Fredric L. Rice / The Skeptic Tank