请编写一个Dyna-Q算法的matlab代码

以下是一个简单的Dyna-Q算法的Matlab代码实现：

% Dyna-Q算法
% Q-learning更新规则，模型学习和模拟更新

% 初始化
num_states = 10; % 状态数
num_actions = 4; % 动作数
Q = zeros(num_states, num_actions); % Q表
model = zeros(num_states, num_actions); % 模型表
s = 1; % 初始状态
alpha = 0.1; % 学习速率
gamma = 0.9; % 折扣因子
epsilon = 0.1; % 探索率
num_episodes = 100; % 迭代次数
num_steps = 100; % 每个迭代步骤数
model_iterations = 50; % 模型学习的迭代次数

% 迭代
for i = 1:num_episodes
    % 初始化状态
    s = 1;
    for j = 1:num_steps
        % 选择动作
        if rand < epsilon
            a = randi(num_actions);
        else
            [~,a] = max(Q(s,:));
        end
        % 执行动作
        s_new = max(min(s + (a-2)*2, num_states), 1); % 状态转移
        r = -1; % 固定奖励
        % Q-learning更新Q表
        Q(s,a) = Q(s,a) + alpha*(r + gamma*max(Q(s_new,:)) - Q(s,a));
        % 模型学习
        model(s,a) = s_new;
        for k = 1:model_iterations
            s_k = randi(num_states);
            a_k = randi(num_actions);
            s_new_k = model(s_k,a_k);
            r_k = -1; % 固定奖励
            Q(s_k,a_k) = Q(s_k,a_k) + alpha*(r_k + gamma*max(Q(s_new_k,:)) - Q(s_k,a_k));
        end
        % 更新状态
        s = s_new;
    end
end

% 测试
s = 1;
while s ~= num_states
    [~,a] = max(Q(s,:));
    s_new = max(min(s + (a-2)*2, num_states), 1);
    disp(['State: ', num2str(s), ', Action: ', num2str(a), ', New state: ', num2str(s_new)]);
    s = s_new;
end