Not freeing 100% of the resources.
This project is a basic genetic algorithm that guesses a string passed as argument.
make
Usage: ./genetic target population_size mutation_rate
The goal is to guess the targetparameter based on potential association with a lot of data.
A Population object is instanced. It contains an array of Entity objects sized after the population_size parameter. Entity objects contain a _dna string typed.
A potential value is given to each Entity object. This potential value is calculated based on the match with the target. (+1 potential for a correct and well placed character). The potential is squared to make the potential somewhat exponential.
All those Entity are placed in a matching pool based on their potential. For example an Entity object with a potential of 12 will be placed twelve times in the matching pool which leads it to have more chance to be picked as a parent in the next function.
The Crossover function randomly picks two parents in the matching pool. It randomly determines a half point and take the first part of the first parent till the half point and the second part of the second parents after the half point.
For example :
| Parent | Dna |
|---|---|
| First parent | abcdef |
| Second parent | ghijkl |
| Random cut point | 4 |
| Child parent | abcdkl |
Once the child is born it has a chance to mutate based on the mutation_rate parameter. The mutation is applied to all chromosomes of its Dna. Basically each characters has a chance to be randomly replaced. It helps creating variety in the population. (the mutation rate is usually about 1%. Too much variety introduced in a new generation is not efficient).
When enough children are born a new population is generated and the whole process is recursively called until a potential equivalent to 100% is found.