Scribd
Upload
4 views

GA Optimization

Uploaded by

andrew Smith
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
4 views

GA Optimization

Uploaded by

andrew Smith
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

% Load your data

load('successssss.mat'); % Load your dataset

% Define parameters
populationSize = 50; % Number of individuals in the population
numGenerations = 100; % Number of generations
numWeights = 11; % Number of weights to optimize (change as needed)
mutationRate = 0.1; % Probability of mutation

% Initialize population
population = rand(populationSize, numWeights) * 2 - 1; % Random weights between -1 and 1

% Fitness function (to minimize)


function fitness = evaluateFitness(weights, input, target)
net = feedforwardnet(numWeights);
net = configure(net, input, target);
net.IW{1} = weights'; % Set weights
output = net(input);
fitness = mean((output - target).^2); % Mean Squared Error
end

% Main GA loop
for generation = 1:numGenerations
fitnessScores = zeros(populationSize, 1);

% Evaluate fitness for each individual


for i = 1:populationSize
fitnessScores(i) = evaluateFitness(population(i, :), datanice, targetOutputs);
end

% Selection (tournament selection)


selectedIndices = tournamentSelection(fitnessScores, populationSize);
selectedPopulation = population(selectedIndices, :);

% Crossover
offspring = crossover(selectedPopulation);

% Mutation
offspring = mutate(offspring, mutationRate);

% Create new population


population = [selectedPopulation; offspring];

% Display best fitness for the generation


bestFitness = min(fitnessScores);
disp(['Generation ' num2str(generation) ', Best Fitness: ' num2str(bestFitness)]);
end

% Function for tournament selection


function selectedIndices = tournamentSelection(fitnessScores, populationSize)
tournamentSize = 5;
selectedIndices = zeros(populationSize, 1);

for i = 1:populationSize
competitors = randi(populationSize, tournamentSize, 1);
[~, bestIdx] = min(fitnessScores(competitors));
selectedIndices(i) = competitors(bestIdx);
end
end

% Function for crossover


function offspring = crossover(selectedPopulation)
[numSelected, numWeights] = size(selectedPopulation);
offspring = zeros(numSelected, numWeights);

for i = 1:2:numSelected-1
parent1 = selectedPopulation(i, :);
parent2 = selectedPopulation(i+1, :);

% One-point crossover
crossoverPoint = randi(numWeights - 1);
offspring(i, :) = [parent1(1:crossoverPoint), parent2(crossoverPoint + 1:end)];
offspring(i+1, :) = [parent2(1:crossoverPoint), parent1(crossoverPoint + 1:end)];
end
end

% Function for mutation


function mutatedPopulation = mutate(population, mutationRate)
mutatedPopulation = population;
[numIndividuals, numWeights] = size(population);

for i = 1:numIndividuals
for j = 1:numWeights
if rand < mutationRate
mutatedPopulation(i, j) = rand * 2 - 1; % New random weight between -1 and 1
end
end
end
end

You might also like