Scribd
Upload
2 views

message (1)

This document outlines a genetic algorithm implemented in MATLAB to optimize a fitness function over a specified number of iterations. It defines parameters such as crossover and mutation probabilities, initializes a population of chromosomes, and iteratively computes fitness values, selects parents, performs crossover, and applies mutations. The results of each iteration, including the standard deviation of fitness, are displayed to track the algorithm's progress.

Uploaded by

Miguel
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
2 views

message (1)

This document outlines a genetic algorithm implemented in MATLAB to optimize a fitness function over a specified number of iterations. It defines parameters such as crossover and mutation probabilities, initializes a population of chromosomes, and iteratively computes fitness values, selects parents, performs crossover, and applies mutations. The results of each iteration, including the standard deviation of fitness, are displayed to track the algorithm's progress.

Uploaded by

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

% Define parameters

pc = 0.3; % Crossover probability


pm = 0.2; % Mutation probability
num_iters = 50; % Number of iterations
pop_size = 10; % Population size

% Initial population (each row is a chromosome)


population = [265 395 139;
294 274 273;
41 42 479;
313 414 71;
51 210 211;
315 347 458;
257 367 372;
312 404 196;
213 294 328;
12 328 86];

% Random values for selection and crossover


initial_random_values = rand(pop_size, 1);
parent_selection_values = rand(pop_size, 1);

% Mutation data (indices in 1D format)


mutated_tiles = [1, 25, 28, 20, 23, 22];
mutated_values = [196, 328, 86, 353, 16, 138];

% Fitness function (Max-Z function)


max_z_function = @(x1, x2, x3) (10*x1 + 9*x2 + 8*x3) ...
- 5 * (max(0, 4*x1 + 3*x2 + 2*x3 - 1300))^2 ...
- 5 * (max(0, 3*x1 + 2*x2 + 2*x3 - 1000))^2;

for iter = 1:num_iters


fprintf('Iteration %d\n', iter);

% Compute fitness values for population


num_values = size(population, 1);
Z_values = zeros(num_values, 1);
for i = 1:num_values
x1 = population(i, 1);
x2 = population(i, 2);
x3 = population(i, 3);
Z_values(i) = max_z_function(x1, x2, x3);
end

% Shift fitness values to ensure non-negative probabilities


Z_min = min(Z_values);
Z_shifted = Z_values + abs(Z_min);
total_Z = sum(Z_shifted);
probabilities = Z_shifted / total_Z;
cumulative_probabilities = cumsum(probabilities);

% Selection Process (Roulette Wheel Selection)


new_population = zeros(size(population));
for i = 1:length(initial_random_values)
rand_val = initial_random_values(i);
selected_idx = find(cumulative_probabilities >= rand_val, 1, 'first');
new_population(i, :) = population(selected_idx, :);
end
% Parent Selection
parent_indices = find(parent_selection_values <= pc);
selected_parents = new_population(parent_indices, :);

% Crossover Process
num_parents = size(selected_parents, 1);
if num_parents >= 2
new_offspring = zeros(num_parents, size(selected_parents, 2));
for i = 1:num_parents
parent1 = selected_parents(i, :);
parent2 = selected_parents(mod(i, num_parents) + 1, :);
offspring1 = [parent1(1), parent2(2), parent2(3)];
offspring2 = [parent2(1), parent1(2), parent1(3)];
new_offspring(i, :) = offspring1;
if i < num_parents
new_offspring(i + 1, :) = offspring2;
else
new_offspring(1, :) = offspring2;
end
end
for i = 1:num_parents
if i <= length(parent_indices)
new_population(parent_indices(i), :) = new_offspring(i, :);
end
end
end

% Mutation Process
[num_rows, num_cols] = size(new_population);
for i = 1:length(mutated_tiles)
tile_idx = mutated_tiles(i);
row = floor((tile_idx - 1) / num_cols) + 1;
col = mod(tile_idx - 1, num_cols) + 1;
if row <= num_rows && col <= num_cols
new_population(row, col) = mutated_values(i);
end
end

% Compute fitness for mutated population


Z_values = zeros(num_values, 1);
for i = 1:num_values
x1 = new_population(i, 1);
x2 = new_population(i, 2);
x3 = new_population(i, 3);
Z_values(i) = max_z_function(x1, x2, x3);
end

% Compute standard deviation


mean_Z = mean(Z_values);
variance = sum((Z_values - mean_Z).^2) / num_values;
fitness_std_dev = sqrt(variance);

% Display iteration results


fprintf('Standard Deviation of Fitness: %.4f\n', fitness_std_dev);

% Update population for next iteration


population = new_population;
end

You might also like