What is genetic algorithm, and give two examples

ga(genetic algorithm)

Genetic algorithms. GA is an efficient exploration algorithm based on the genetic evolution mechanism of natural populations, which was first proposed by the American scholar Holland in 1975.

It abandons the traditional search method, simulates the biological evolution process in nature, and uses artificial evolution to conduct a random search of the target space. It treats the possible solution in the problem domain as an individual or chromosome of the population, and encodes each individual into a symbolic string form, simulates the biological evolution process of Darwin's genetic selection and natural elimination, repeatedly performs genetics-based operations (heredity, crossover and variation) on the population, evaluates each individual according to the predetermined target fitness function, continuously obtains a better population according to the evolutionary rules of survival of the fittest and survival of the fittest, and searches for the optimal individuals in the optimal population by global parallel search. Find the optimal solution that satisfies the requirements.

The genetic algorithm created by Holland is a probabilistic search algorithm that uses some coding technique to act on strings of numbers called chromosomes, and its basic idea is to simulate the evolutionary processes that are made up of these. The tarsal algorithm reassembles the well-adapted strings through an organized but random exchange of information, and in each generation, uses the well-adapted bits and segments of the previous generation string structure to generate a new group of strings; As an added addition, occasionally try to replace the original parts with new bits and segments in the string structure.

Genetic algorithm is a kind of randomization algorithm, but it is not a simple random walk, it can effectively use the existing information processing to search for those strings that have the potential to improve the quality of the solution, similar to natural evolution, genetic algorithm by acting on the genes on the chromosome, looking for good chromosomes to solve the problem. Similar to nature, genetic algorithms are ignorant of solving problems per se, and all they need is to evaluate each chromosome produced by the algorithm and counter chromosomes based on fitness values, so that chromosomes with good applicability have more chances to reproduce than chromosomes with poor adaptability.

Gene: The unit that makes up a chromosome, which can be represented as a binary bit, an integer, or a character, etc.

Chromosome or individual: Represents a possible solution to the problem to be solved, consisting of several genes, which is the basic object of the GA operation.

Fitness or moderation: represents the superiority or weakness of an individual's corresponding solution, usually represented by a fitness function.

Selection: One of the basic operations of GA is to select individuals who can be paternal parents in the group according to a certain generalization according to the fitness of the individual, and the selection is based on the high probability that the individual with large fitness will be selected. The selection operation embodies the evolutionary rules of survival of the fittest and survival of the fittest.

Crossover: One of the basic operations of GA, that is, the paternal individual randomly exchanges genes to form a new individual according to a certain probability.

Variation: One of the basic operations of GA, that is, to randomly change the gene value of an individual with a certain probability.

First of all, there is a very amazing phenomenon: the evolution of human beings and animals is developing in a good direction, although some have developed in a bad direction, but the overall development is definitely in a good direction. This may not seem strange, but we know that the genetic mix of human beings is random and not subject to God's constraints.

The results of this stochastic process are consistent!! That's where our genetic algorithm got its inspiration! For example, if I ask for the maximum value of y=x1+x2, two variables, I don't use traditional mathematical methods, just use the kindergarten method, bring in all possible values, and then find out the largest!

However, sometimes the value is continuous, and that's okay! Make it discret, just like turning an analog signal into a digital signal! Another question is, what if there are too many values?

That's the essence of genetic algorithms!

First of all, I don't need to take all the possible values, I only take dozens or hundreds (set by myself), and then process them, how to deal with them? Let's go back to the beginning of human evolution, although there is no help from God, but we know that nature follows the survival of the fittest, follows the law of cross-mutation, although it cannot be digitized, but this is a trend! That's what we mathematize!

Which of the dozens of values do I have left? Which ones to throw away? Which ones to deal with?

This has to be our own choice, must be to choose the most appropriate value to stay, after a series of processing, a new group is generated, and then processed, their own agreement to deal with the first few times, take the maximum value.

Don't worry about whether it is the maximum value obtained, because it has been mathematically proven that this method is convergent and the probability is 1, so although you can do it with confidence, the specific method should refer to the relevant books, which is not difficult.

The greatest use of genetic algorithms is to solve problems that mathematical theories can't! For example, path planning, scheduling problems ......

Genetic algorithm is a global optimization probability algorithm, and the main advantages are 1The genetic algorithm does not have too many mathematical requirements for the optimization problem to be solved, and due to its evolutionary characteristics, the internal properties of the problem are not required in the search process, and any form of objective function and constraint, whether linear or nonlinear, discrete or continuous, can be handled.

2.The eericulturality of evolutionary operators enables genetic algorithms to perform global searches of probabilistic meanings very effectively.

3.Genetic algorithms can provide great flexibility for a variety of special problems to mix and construct domain-independent heuristics, so as to ensure the effectiveness of the algorithm.

There are many applications, such as numerical optimization, combinatorial optimization, machine learning, intelligent control, artificial life, image processing, pattern recognition and other fields.

Rights. The simplest application is the function optimization problem, which is to find the extreme value of a more complex function. If you want relevant MATLAB** or C**, I can send it to you and leave an email.

There are many applications, and they can be used in almost all problems to find optimal solutions.

Genetic algorithms are mainly used to solve optimization problems.

Generally speaking, it can solve the maximum and minimum values of functions, and can also be combined with some other methods to solve (non-)linear regression, classification problems, and so on.

However, genetic algorithm has two disadvantages, one is the long time, and the other is that the selection of the initial value will affect the convergence effect.

Modern medical research has shown that DNA is the most important genetic material for living life. Heredity, on the other hand, refers to the transmission of genes so that the offspring can acquire the characteristics of their parents. Genetics is a discipline that studies heredity, which is the symptom of the spine destruction, and in addition to genetic factors, there is also the environment, and the interaction between the environment and heredity.

It is also a factor that determines biological characteristics.

Genetic algorithms. as an operating object; Second, the genetic algorithm directly uses fitness as the search information, without other auxiliary information such as derivatives. Thirdly, the genetic algorithm uses multiple points to search for information, which has implicit parallelism. In the end, instead of using non-deterministic rules, Sakura sold the probabilistic search technique.

After investing in a breakfast franchise.