Wesley R. Elsberry
Posts: 4991 Joined: May 2002
|
Then there is phoodoo's comment:
Quote | There is no power of cumulative selection, if you don’t decide beforehand what you are selecting for.
That’s why evolutionary algorithms are so laughable.
|
While these guys are closing their eyes, covering their ears, and shouting, "LA-LA-LA-I-CAN'T-HEAR-YOU", the rest of the world is getting on with business. Evolutionary computation is being deployed in research, engineering, and even finance. It is responsible for the shape of vanes in jet engines, determination of well sites for water remediation, PCB drill paths, and many, many more applications.
I've used a form of evolutionary computation that discovers mathematical models inherent in data, applied to a forecasting problem. This is something that really shows the desperation behind the critique "phoodoo" makes; not only is there no "target" known beforehand, the whole point of using the technique is model *discovery*. There is no limitation of where the data comes from or what might be in it.
Sure, any time that there is a closed-form analytical solution to your problem, you use it. But there are plenty of problems that don't have effective methods established for them, and evolutionary computation has been giving us an assist on those for decades now.
-------------- "You can't teach an old dogma new tricks." - Dorothy Parker
|