DNA, the real fabric of our lives!How do you approach problems you don't know how to solve? Do you
analyze it down to its roots until you get a logical solution? Maybe you just jump in and try something to see
how well it works, then maybe modify it until the problem is solved? Genetic Algorithms use the latter
try-it-and-modify-it approach to solve complicated problems by creating many solutions to their particular
problem and measuring those solutions' efficacy.
The "genetic" part of the name comes from in their paradigm of breeding solutions. In short, GA's use a survival-of-the-fittest technique for choosing which given solution (called "genomes") is best for solving the problem. This means that genomes are selected from a population with a certain probability, based on their fitness, and are mated with another genome to make a whole new "child". After many generations of this, the population eventually comes up with a viable solution to the problem.
Sounds cool, huh? Turns out they work really well, too!
I've done a considerable amount of work in this area an undergraduate/hobbyist. If you want more to learn about GAs, then this is the place to be!
|
Articles On Genetic Algorithms
 |
Galapagos
 |
|
|
|
At one point I felt very motivated
to write a lot about Genetic Algorithms. There's a ton of insight here on the various nuances in these
things, all of which have come from my own programming research. The writing isn't perfect, but I am still
quite proud of this body of work.Page last updated: February 04, 2008