项目作者： Ivaskuu
项目描述 ：
This is a processing java project that simulates the genetic evolution of a string of text.
高级语言： Processing
项目地址: git://github.com/Ivaskuu/evolution_of_strings.git
Work in progress
Evolution of strings
This is a java (processing) project that simulates the genetic evolution of a string of text.
It’s my first Machine Learning - Artificial Intelligence project, so please be gentle with me :-).
Files
The project is composed of 3 files :
- main.pde : It creates a population based on the target, the mutation rate and the number of elements, and train it.
- population.pde : It manages all the elements, that form the population, cioè a method to train the elements, and method to reproduce them.
- dna.pde : It contains the genes of an element, a method to calculate it’s fitness, a method to mutate and a method to create an element from 2 parents.
main.pde
Methods
- void setup() : Called once at the start of the program. It creates a new population, then trains it in a while loop.
population.pde
Methods
- Population(String stringTarget, double mutationRate, int populationNum) : The constructer of the class, it creates a new population with the specified parameters.
- void calculatePopFitness() : It calculates the fitness of all elements, and prints the median fitness, best’s genes and generation number.
- void reproduce() : It creates a new generation of population composed of popNum elements, based on the previous best elements DNA.
- void prepareTheMatingPool() : Puts all the elements fitness in a cumulative char array.
- int binarySearch(double value) : Returns the position of the range of the elements. The parameter value is a random double (from 0 to the max cumulative fitness).
- void mutate() : Calls the method mutate() for each element.
dna.pde
Methods
- void wip() : WIP :-).
- void wip() : WIP :-).
Mutations type and probabilities
- One gene : One gene of the DNA is modified with a random char. Probability : 40/105.
- Multiple genes : A random number of genes (from 0 to max) of the DNA is modified with random chars. Probability : 30/105.
- Split & rearrange : A random point in the array is choosen, where the DNA gets splitted nad the first part goes last. Probability : 20/105.
- Reverse genes : The DNA is reversed. Probability : 10/105.
- All genes : All the genes of the DNA are random chars. Probability : 5/105.
