### Evolutionary Computation News

Interesting news on genetic algorithms and a quick analysis of their academic/design impact upon those (and many other) areas. See here.

As you may see, there are persons which consider genetic algorithms too slow and others think they are too cute. :)

Some excerpts from the news:

"Yet there are drawbacks to the method. Although genetic algorithms have been applied with fantastic success in some cases, they fall short of being universal problem-solvers. And although evolutionary eons can be compressed into hours, finding the precise settings that will give a good solution can take months. How do you figure out which solutions are best? How often do you mutate the offspring? Computer scientists are reduced to trial-and-error knob-twiddling to get the right conditions."

"They’re too cute. Genetic algorithms don’t get bonus credibility just because that’s what nature did,"

"They are quite slow, and they require quite a bit of fiddling,"

"Genetic algorithms’ main contribution today are that they opened the door to biologically-inspired methods. The beauty and success of genetic algorithms motivated other computer scientists to look to biology for inspiration,"

By the way, I consider that genetic algorithms (and evolutionary algorithms in general) are very nice to solve problems, mainly when taking into account the lack of problem informations, such as no derivative available and/or so many constraints to handle.

Labels: Artificial Evolution, Evolution, Evolutionary Algorithm, Evolutionary Computation, Genetic Algorithm, Simulated Evolution

## 2 Comments:

Hi, Marcelo. I've been using genetic algorithms to do simulation optimization, and they are great tools for this kind of application mainly for one reason: they do not make any assumptions with regards to the structure of the problem. All they need is the assignment of a solution quality, no matter how you got the measure. This is fantastic for the solution of complex engineering problems, since the trend nowadays is turning rapidly to the extensive use of simulation to assess the performance of huge mechanical structures or systems of any kind. I have a feeling that our rather abstract mathematics isn't sufficient to capture all idiosyncrasies that our real life problems exhibit. Even more I believe that the solution of problems is a matter of how you correctly explore and exploit information about it, than a matter of how you manipulate analytical mathematical relations based on possibly flawed assumptions.

Hi, Profeta!

I quite agree with you upon the usefulness of genetic algorithms.

That featureless assumption has provided interesting and, from time to time, strange results - even though useful ones - before the eyes of those trained on the traditional grounds of optimization, such as gradient-based methods and/or brute force numerical computation methods.

You wrote:

"All they need is the assignment of a solution quality, no matter how you got the measure."I carefully agree. :)

For the GA success, it is important to model a well sounded fitness/objective function, since reasoned results are strong tied to this modelling.

"This is fantastic for the solution of complex engineering problems, since the trend nowadays is turning rapidly to the extensive use of simulation to assess the performance of huge mechanical structures or systems of any kind."That statement should be taken with some caution, because, not so rarely, there are methods which find solutions through a much cheaper and easier way than implementing an evolutionary algorithm. Sure, if the user is well skilled on evolutionary methods and feels that nice results may be got through them, then it is advisable to use them, being aware that other methods could be equally interesting, mainly from the computational cost viewpoint.

Our mathematical models are limited assumptions about the inner working of nature phenomena, but this does not qualify them as futile efforts toward understanding few aspects of our physical world.

"Even more I believe that the solution of problems is a matter of how you correctly explore and exploit information about it [...]."Even this may be flawed, however it does not mean that it is useless. On the contrary. :)

Best Regards!

Marcelo

Post a Comment

## Links to this post:

Create a Link

<< Home