## Genetic algorithm components

Genetic algorithms have the following three components:

**Genetic encoding**(**and decoding**): This is the conversion of a solution candidate and its components into the binary format (an array of bits or a string of`0`

and`1`

characters)**Genetic operations**: This is the application of a set of operators to extract the best (most genetically fit) candidates (chromosomes)**Genetic fitness functions**: This is the evaluation of the fittest candidate using an objective function

Encodings and the fitness function are problem dependent. Genetic operators are not.

### Encoding

Let's consider the optimization problem in machine learning that consists of maximizing the log likelihood or minimizing the loss function. The goal is to compute the parameters or weights, *w={w _{i}}*, that minimize or maximize a function

*f(w)*. In the case of a nonlinear model, variables may depend on other variables, which make the optimization problem particularly challenging.

#### Value encoding

The genetic algorithm manipulates...