车间布局优化的遗传算法 Matlab 代码，车间长10米，宽5米，设备16台，每台设备长1.2米，宽0.8，要求布局为两行，且不重叠

以下是一个简单的遗传算法的 Matlab 代码，用于车间布局优化：

% 定义问题的参数
room_width = 10; % 车间宽度
room_length = 5; % 车间长度
num_machines = 16; % 设备数量
machine_width = 0.8; % 设备宽度
machine_length = 1.2; % 设备长度
num_rows = 2; % 行数
row_width = room_width / num_rows; % 每行宽度

% 定义遗传算法的参数
pop_size = 100; % 种群大小
num_generations = 100; % 迭代次数
mutation_rate = 0.01; % 变异率

% 初始化种群
pop = zeros(pop_size, num_machines * 2);
for i = 1:pop_size
    pop(i,:) = randperm(num_machines * 2);
end

% 迭代计算
for generation = 1:num_generations
    % 计算适应度
    fitness = zeros(1, pop_size);
    for i = 1:pop_size
        layout = zeros(num_machines, 2);
        for j = 1:num_machines
            machine_id = pop(i, j);
            row_id = ceil(j / (num_machines / num_rows));
            layout(machine_id,:) = [row_width * (row_id - 1), mod(j - 1, num_machines / num_rows) * machine_length];
        end
        % 判断是否有重叠
        for j = 1:num_machines
            for k = 1:j-1
                if (abs(layout(j,1) - layout(k,1)) < machine_width) && (abs(layout(j,2) - layout(k,2)) < machine_length)
                    fitness(i) = fitness(i) - 1;
                end
            end
        end
        % 判断是否越界
        if max(layout(:,1) + machine_width) > room_width || max(layout(:,2) + machine_length) > room_length
            fitness(i) = fitness(i) - 1;
        end
        fitness(i) = fitness(i) + sum(layout(:,1));
    end
    % 选择
    [sorted_fitness, idx] = sort(fitness, 'descend');
    pop = pop(idx,:);
    % 交叉
    for i = 1:pop_size/2
        parent1 = pop(i,:);
        parent2 = pop(i+1,:);
        crossover_point = randi([1,num_machines*2-1]);
        child1 = [parent1(1:crossover_point), parent2(crossover_point+1:end)];
        child2 = [parent2(1:crossover_point), parent1(crossover_point+1:end)];
        pop(i,:) = child1;
        pop(i+1,:) = child2;
    end
    % 变异
    for i = 1:pop_size
        if rand < mutation_rate
            mutation_point1 = randi([1,num_machines*2]);
            mutation_point2 = randi([1,num_machines*2]);
            temp = pop(i,mutation_point1);
            pop(i,mutation_point1) = pop(i,mutation_point2);
            pop(i,mutation_point2) = temp;
        end
    end
end

% 输出最优解
best_layout = zeros(num_machines, 2);
for i = 1:num_machines
    machine_id = pop(1, i);
    row_id = ceil(i / (num_machines / num_rows));
    best_layout(machine_id,:) = [row_width * (row_id - 1), mod(i - 1, num_machines / num_rows) * machine_length];
end
disp(best_layout)

该代码包括了定义问题参数、初始化种群、计算适应度、选择、交叉和变异等遗传算法的基本步骤。其中，计算适应度的部分考虑了设备之间的重叠和边界条件，并根据设备位置对适应度进行评分。最终输出最优解的设备位置。