Genetic algorithms used to produce fit neural networks. More...

Public Member Functions
	GeneticAlgo (unsigned int populationSize_, float mutationRate_, float crossoverRate_, int numberOfElitismCopies_, std::vector< double >(*getPopulationFitness_)(const std::vector< net::NeuralNet > &neuralNet))
	Initializes a genetic algorithm with the necessary parameters. More...

net::NeuralNet	getBestNeuralNetwork (int numberOfGenerations, net::NeuralNet modelNetwork)
	Returns the most fit neural network in a population that undergoes a specified number of generations. More...

Detailed Description

Genetic algorithms used to produce fit neural networks.

Constructor & Destructor Documentation

GeneticAlgo::GeneticAlgo	(	unsigned int	populationSize_,
		float	mutationRate_,
		float	crossoverRate_,
		int	numberOfElitismCopies_,
		std::vector< double >(*)(const std::vector< net::NeuralNet > &neuralNet)	getPopulationFitness_
	)

Initializes a genetic algorithm with the necessary parameters.

Parameters

populationSize_	the number of neural networks in the population for each generation
mutationRate_	the chance between 0 and 1 that a weight in a neural network will be mutated
crossoverRate_	the chance between 0 and 1 that a couple will produce offspring having a fused verion of their parents' networks, instead of offspring that have the same neural networks as their parents; if set to 0, the population's composition will only change as a result of mutations.
numberOfElitismCopies_	the number of times that best neural network from the past generation shows up in the current generation
getPopulationFitness_	a function that returns a vector of doubles housing the fitness of every individual in the population, when given the population of neural networks

net::NeuralNet GeneticAlgo::getBestNeuralNetwork	(	int	numberOfGenerations,
		net::NeuralNet	modelNetwork
	)

Returns the most fit neural network in a population that undergoes a specified number of generations.

Parameters

numberOfGenerations	the number of generations that the simulation will run for; in each generation, fitnesses are assigned, mating occurs, and offspring are added
modelNetwork	used to determine the architecuture of the neural networks in the population,.

The documentation for this class was generated from the following files: