matlab 写能运行的代码 用matlab遗传算法做一个逃出盒子的案例，要有一个20*10的网格，有起点终点。

好的，以下是一个简单的逃出盒子的示例代码，使用遗传算法求解：

% 逃出盒子问题的遗传算法求解示例代码

% 初始化
num_rows = 20;
num_cols = 10;

start_pos = [1, 1];
end_pos = [20, 10];

num_genes = num_rows * num_cols;
num_population = 100;
num_generations = 500;

mutation_rate = 0.01;

% 生成随机种群
population = randi([1, 4], num_population, num_genes);

% 遗传算法迭代
for i = 1:num_generations
    % 计算适应度
    fitness = zeros(1, num_population);
    for j = 1:num_population
        path = decode_path(population(j, :), num_rows, num_cols);
        fitness(j) = evaluate_fitness(path, start_pos, end_pos);
    end
    
    % 选择
    [sorted_fitness, indices] = sort(fitness, 'descend');
    elite = population(indices(1:10), :);
    selected = roulette_wheel_selection(population, sorted_fitness);
    
    % 交叉
    offspring = crossover(selected, num_population - size(elite, 1));
    
    % 变异
    mutated = mutation(offspring, mutation_rate);
    
    % 合并
    population = [elite; mutated];
end

% 输出结果
best_path = decode_path(population(1, :), num_rows, num_cols);
disp(['Best path: ', mat2str(best_path)]);

% 将一维基因序列解码成路径
function path = decode_path(genes, num_rows, num_cols)
    path = zeros(size(genes));
    current_pos = [1, 1];
    for i = 1:length(genes)
        path(i) = current_pos(1) + (current_pos(2) - 1) * num_rows;
        switch genes(i)
            case 1  % 向上
                if current_pos(1) > 1
                    current_pos(1) = current_pos(1) - 1;
                end
            case 2  % 向下
                if current_pos(1) < num_rows
                    current_pos(1) = current_pos(1) + 1;
                end
            case 3  % 向左
                if current_pos(2) > 1
                    current_pos(2) = current_pos(2) - 1;
                end
            case 4  % 向右
                if current_pos(2) < num_cols
                    current_pos(2) = current_pos(2) + 1;
                end
        end
    end
end

% 计算适应度
function fitness = evaluate_fitness(path, start_pos, end_pos)
    fitness = 0;
    current_pos = start_pos;
    for i = 1:length(path)
        if isequal(current_pos, end_pos)
            break;
        end
        next_pos = [mod(path(i) - 1, size(path, 1)) + 1, floor((path(i) - 1) / size(path, 1)) + 1];
        if isequal(next_pos, current_pos)
            continue;
        end
        fitness = fitness + 1;
        current_pos = next_pos;
    end
    if isequal(current_pos, end_pos)
        fitness = fitness + 1000;
    end
end

% 轮盘赌选择
function selected = roulette_wheel_selection(population, fitness)
    num_population = size(population, 1);
    total_fitness = sum(fitness);
    probability = fitness / total_fitness;
    cumulative_probability = cumsum(probability);
    selected_indices = zeros(1, num_population);
    for i = 1:num_population
        r = rand();
        for j = 1:num_population
            if r < cumulative_probability(j)
                selected_indices(i) = j;
                break;
            end
        end
    end
    selected = population(selected_indices, :);
end

% 交叉
function offspring = crossover(population, num_offspring)
    num_population = size(population, 1);
    num_genes = size(population, 2);
    offspring = zeros(num_offspring, num_genes);
    for i = 1:num_offspring
        parent1 = population(mod(i - 1, num_population) + 1, :);
        parent2 = population(mod(i, num_population) + 1, :);
        crossover_point = randi([1, num_genes - 1]);
        offspring(i, :) = [parent1(1:crossover_point), parent2(crossover_point+1:end)];
    end
end

% 变异
function mutated = mutation(offspring, mutation_rate)
    num_genes = size(offspring, 2);
    mutated = offspring;
    for i = 1:size(offspring, 1)
        for j = 1:num_genes
            if rand() < mutation_rate
                mutated(i, j) = randi([1, 4]);
            end
        end
    end
end

在这个示例代码中，我们使用了一个一维的基因序列来表示路径，其中每个元素都是一个数字，表示向上、向下、向左、向右四个方向中的一个。使用 `decode_path` 函数将基因序列解码成路径，使用 `evaluate_fitness` 函数计算路径的适应度，使用 `roulette_wheel_selection` 函数进行选择，使用 `crossover` 函数进行交叉，使用 `mutation` 函数进行变异。最终，我们输出最优路径的基因序列并将其解码成二维的路径。