Applications of the Evolutionary Computation

Manuel de la Herrán Gascón


Types of Applications

Any problem puede deal as a problem of some of these types:

[ Back to Index ]

Equivalences between types of Applications

Exist equivalences between some and other types. A forecast is a classification in the that intervene the time, and a forecast also it is puede to interpret as a search within of the space of possible states future or as a problem of minimización of the time employee in this search.

[ Back to Index ]

Algorithm Evolutionary as a method of optimization

To interpret a algorithm evolutionary as a method of optimization is it more intuitive. The function of evaluation returns values high for the solutions good, with it that each time we will will have better solutions, and in that consist precisely to optimize.

[ Back to Index ]

Algorithm Evolutionary as a method of search

For to consider a algorithm evolutionary as a method of search, simply we have that to make that the function of evaluation returns values but high how much more near estemos of that that it is it is seeking. We can to identify each chain of the algorithm with a position in a space imaginary, that it will be the space of search. If the chains had two elements, we would would have a space of two dimensions. With three elements, we would would have a space three-dimensional, and thus sucesivamente. Normally we will will have many more dimensions, but enough with to think in two or three dimensions and extrapolar the concept of "movement for the space" for to interpret correctly the idea. For example, a mutation of a gene would displace to the chain to a position different in one of the shafts.

Evidently, it is it must power to recognize until that point we are next to that that we seek. Not is possible to find something if not we know that is it that we are seeking. There is a case in the that, to the less in appearance, this rule has a exception. In some systems of recognition of standards as some nets neuronales not supervised, exist the possibility of to learn without no training and without supervision some. ¿What is it that learn then the net? The response it is in the own definition of the that makes the net: to recognize standards. The net makes precisely that: to recognize series, sequences of data repeated, they will be of the type that they will be; regularities present in the data, categorizing them in classes not determined to priority.

[ Back to Index ]

Examples of Applications

[ Back to Index ]

¡You are guest to hold, criticize and collaborate with your ideas!
Send your commentaries, questions, suggestions, or criticize to E-mail
[ Home Page Castellano | Home Page English ]