CompuServe Thread

#Genetic Algorithms

5 messages in this thread
#20152From: George B. HogensonJul 27, 1992 7:36 PM
Greetings. I'm new to the Cmpsrv system as well as this forum. I am wondering whether anyone is working with John Holland's genetic algorithms on this forum. Many thanks George
#20159From: PhilD [ADESK]Jul 27, 1992 8:44 PM
Hi George. Welcom to the forum. There have been a couple questions asked lately about John Holland's Genetic algorithms. There isn't much activity with it, however, that I am aware of. Not at this time, at least. From what I understand about it, it could be a lot of fun.
#20176From: Mike MorrisonJul 28, 1992 9:43 AM
What are these "Genetic" Algorithms all about? Plant/Organic 3D object creation? -MM
#20189From: PhilD [ADESK]Jul 28, 1992 1:36 PM
<<What are these "Genetic" Algorithms all about? Plant/Organic 3D object creation?>> I don't really have any details on it. It appears to be a step above the basic CA, involving the ability to mutate, to carry individual attributes, and so forth. More genetic like than are cellular automata usually.
#20234From: Grahame NaylorJul 29, 1992 12:54 PM
The genetic part refers to the potential for evolution. The theory revolves around algorithms which mutate(!) and by being used for the selected task are rejected or selected for further usage the idea is that eventually the program is optimised for its task, it also allows for adapting to changing conditions. There has been quite a lot of waffle published about them but not too much in the way of hard facts. This is probably because the primary application is seen in finacial markets and nobody wants their secrets too widely published. Grahame Some wiser cove may point you to specific papers. There was an article in Byte last Fall (or thereabouts)