Genetic algorithm--a rough understanding

Examples of genetic algorithms:
http://blog.csdn.net/qq_27755195/article/details/56597467 Encoding

The common coding methods are:
Binary encoding, floating-point encoding, natural number coding, gray coding, real coding.
Personal understanding: Coding refers to the mapping method of phenotype to genotype, according to the characteristics of solving problems, the solution of a problem is expressed as a string method. So in fact, according to the problem of choosing a more appropriate string to express their own problems, each gene can accurately express the meaning, and the gene composition of the chromosome can express the form of the solution. genetic operator

The genetic operator contains 2 aspects of crossover and mutation. Cross

The common intersection method of binary coding is: single-point crossover, two-point crossover, uniform crossover , etc. Cross-use should be designed according to their own coding strategy, and it is necessary to pay attention to the intersection of the process can effectively limit the future generations in the feasible domain. and different coding methods also have a unique cross-cutting method.
The implementation of the crossover process consists of 3 steps:

Take the two-point crossover as an example:
mutation

Mutation is to jump out of the local solution, there are quite a lot of novel ideas, literature has not come and read. Descendant Selection Method

Also called the hypothesis choice, one kind of survival choice strategy, the common choice strategy has the roulette, the tournament, the elite retention strategy, the sorting choice and so on. Genetic Programming

It can be used for the discovery of feature combination in data mining, feature mining, but the concrete effect needs to be verified.