|
Evolution is a powerful force that has shaped all life forms on
earth. By replicating the fundamental mechanisms of evolution in
a computer - variation, mutation, recombination and survival of
the fittest - scientists have not only been able to not only study
biological evolution on a fast time scale, but have also found a
revolutionary way of solving a wide range of problems of extreme
complexity.
One popular example of evolutionary computation,
which constitutes one of Icosystem's technology pillars, is the
genetic algorithm. Beginning with an array of potential solutions
to a problem, you evaluate their efficacy and performance and eliminate
the undesirable ones. You then allow the most successful solutions
to pass to the next generation, mutate them slightly and let them
"mate" to produce "offspring". Because the offspring
of a pair of good solutions inherit features from both of their
successful parents, there is a good chance that they will end up
being better than their parents. By repeating this process iteratively
for many generations, better and better solutions emerge. Icosystem
has used this approach to solve a number of optimization and control
problems in industry.
|
