[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Genetic Algorithms in Swarm
From: |
Ferraris |
Subject: |
Genetic Algorithms in Swarm |
Date: |
Sun, 14 May 2000 16:21:24 +0200 |
Hi all
I've read messages from Sid Bhattacharyya and Paul E. Johnson about
swarm agents with genetic algorithms. Using my poor english, for which I beg
your pardon, I'll try to describe an experimentation I've done during working
on GM. I have written a small model in which the GA has to find the maximum of
the function:
y = 1 + cos(x) / (1 + 0.01 * x ^ 2)
In my model each individual encode a natural number that is translated to a
real value in an assigned range as follow:
x = a + n * 1/((2 ^ l - 2)/b)
where a and b are the minimum and maximum values of the inspected range of the
domain, n is the number obtained reading each individual, i.e. each sequence of
"0" and "1", as binary number and l is the length of each individual. Then the
function is computed and the result is given to the individual as fitness
value.
Using the "useDeltaFitness" algorithm coded in GM, I have got good
results: the GA was more precise of a traditional one. Using GM the faature
enabling simpy consists in setting to one the useDeltaFitness parameter in the
probeMap of the ModelSwarm. More about that is contained in the how-to file.
This model is not included in the distribution file of GM, but if anyone likes
have the model I'll pleased to send it to him.
The distributed SwarmModel of GM is based on the individual approach of
genetic algorithm, i. e. each individual of the GA is like an agent, but the
kernel of GM is able to act as well following a mind-approach, i. e. the
individuals of the GA are like ideas into the brain of a single agent and each
agent have an own population. Because the kernel is able to operate on several
population, in fact each agent can have an own GA.This approach is followed in
CW for classifier systems. For more information please mail me.
About developing GM and CW I'll send as soon as possible a new version
of them with few additional features. I know the problems arising from using
software under modification and I promise to take care in preserve
compatibility with old versions.
Actually I'm also trying to realise a unique package to deal both with
GAs and CSs that will allow cooperation between individuals from different
population, operating so in a parallel way. This type of approach has been well
described by Alba & Troya in the article "A Survey of Parallel Distribuetd
Genetic Algorithms" in Complexity vol. 4 pages 31-52.
Thanks for each observation and suggestion
Gigi
==================================
Swarm-Support is for discussion of the technical details of the day
to day usage of Swarm. For list administration needs (esp.
[un]subscribing), please send a message to <address@hidden>
with "help" in the body of the message.
- Genetic Algorithms in Swarm,
Ferraris <=