如何在MATLAB中不使用rand函数生成类似b=round(rand(1,20))的随机序列？

Hey there! If you need to generate random-like number sequences in MATLAB without using the built-in rand function, here are a few solid approaches you can try out—each with its own use case:

This is the simplest method, leveraging the microsecond or millisecond portion of the current system time to generate non-repeating (for short intervals) values. Perfect for quick tests where you don't need industrial-grade randomness.

function random_seq = time_based_rand(num_elements)
    % Grab the current time with microsecond precision
    current_time = datetime('now', 'Format', 'yyyy-MM-dd HH:mm:ss.SSSSSS');
    microseconds = str2double(split(char(current_time), '.'){2});
    
    % Generate a sequence by offsetting the seed and taking modulo
    % Adjust the modulo value to match your desired range (here 0 or 1, like round(rand))
    random_seq = mod(microseconds + (1:num_elements)*13, 2);
end

% Generate your 1x20 sequence
b = time_based_rand(20);

Note: This will repeat if you call it within the same microsecond, so it's not ideal for high-volume or rapid calls.

LCGs are the backbone of many built-in random number generators. They use a mathematical formula to produce a sequence with a long period (if you pick the right parameters). This is way more reliable than the time-based method for most applications.

function random_seq = lcg_generator(num_elements, min_val, max_val)
    % Proven LCG parameters (from Numerical Recipes)
    a = 1664525;
    c = 1013904223;
    m = 2^32;
    
    % Use system time to initialize the seed
    seed = floor(datetime('now', 'ConvertFrom', 'datenum') * 86400000);
    
    random_seq = zeros(1, num_elements);
    for i = 1:num_elements
        seed = mod(a * seed + c, m);
        % Normalize to 0-1, map to your desired range, then round
        random_seq(i) = round((seed / m) * (max_val - min_val) + min_val);
    end
end

% Generate a 1x20 sequence matching round(rand(1,20)) (0 or 1)
b = lcg_generator(20, 0, 1);

% If you wanted integers from 0 to 20 instead, just call:
% b = lcg_generator(20, 0, 20);

Pro Tip: The parameters a, c, and m are chosen to maximize the period of the sequence—stick with tried-and-true values unless you know what you're doing.

If you need sequences with strong statistical randomness, chaotic maps like the Logistic map are a great choice. When tuned to the right parameter, they produce non-repeating, unpredictable sequences.

function random_seq = logistic_chaos_rand(num_elements)
    % Chaos parameter (this value ensures full chaotic behavior)
    r = 3.5699456;
    
    % Initialize with a value derived from system time (no rand functions!)
    time_seed = mod(datetime('now', 'datenum') * 1e6, 1);
    x = time_seed;
    
    random_seq = zeros(1, num_elements);
    for i = 1:num_elements
        x = r * x * (1 - x);
        % Map the chaotic value to 0 or 1 (matching round(rand))
        random_seq(i) = round(x);
    end
end

% Generate your 1x20 sequence
b = logistic_chaos_rand(20);

Heads Up: Chaotic sequences are super sensitive to initial conditions—even tiny changes in the seed will produce completely different results.

内容的提问来源于stack exchange，提问作者Ketaki_S