class NetGA(uint PopSize, alias fitness, alias generator, alias selector = topPar!(to!uint(PopSize * 0.1)), alias crossover = randomCopy, alias mutator = randomWeight, alias comp = "a > b"): SimpleGA!(Network, PopSize, comp, fitness, generator, selector, crossover, mutator);
- Parameters
PopSize The size of the population fitness User defined fitness function. Must return double selector Selection method used to pick parents of next generation. crossover Used to crossover individuals to create the new generation. mutator Used to alter the population. comp Used to determine whether a larger or smaller fitness is better.
Genetic algorithm for genomes in the form of artificial neural nets
this();
Default constructor. No statistics will be printed, and mutation will be set at 1%, crossover rate at 80%
this(float mutRate, float crossoverRate, uint statFreq);
Convienience constructor, equivilant to default constructing and setting mutation rate and statistic frequency
bool autoGenerateGraph(bool generate);
const bool autoGenerateGraph();
Whether to generate a graph of the net after 'evolution' completes
const void generateGraph(const(Network) net, string filename = "output.dot", string description = "");
Generate a a Graphviz dot file named filename with additional description 'description' using node