Matlab creates same input values every time? Why?

Hi everybody,

I'm creating processing times as follows:

function [ processing_times ] = PT( number_of_activity )

number_of_activity = 12;
min= 100;
max = 2500;
processing_times = randi([min max],1,number_of_activity)';

end

When I run this code, I obtain 12x1 matrix as follows:

1481
1372
760
696
1184
646
2031
2467
172
1386
309
2025

After that, I quit the Matlab and open it again.

I run the same code and I obtain the same 12x1 matrix. This is possible? I use

randi([min max],1,number_of_activity)'

code. This code includes randomness. How can I obtain the exactly same value when every time I close and reopen the Matlab. Shouldn't the Matlab give me different values every time I open and close it?

I use Matlab 2015a and I ran this code in 2023a but I encountered the same problem again.

Answer 1

[Torsten]

Shouldn't the Matlab give me different values every time I open and close it?

No. But you get different values if you call the function several times in one run.

Selecting

rng("shuffle")

at the start of your code should also give you different random numbers for different runs:

https://uk.mathworks.com/help/matlab/ref/rng.html

But usually this is not desirable because the results from one run should be reproducable.

[Abdullah Türk]

How and where should I use rng("shuffle") in my code? Sorry, I don't understand Torsten.

[Torsten]

How and where should I use rng("shuffle") in my code?

As first line in your main script.

[Abdullah Türk]

Torsten, I positioned rng('shuffle') at first line in my code as follows and the problem has solved. Thank you.

clc;
clear;

rng('shuffle')

% numWorkers = 2;
% pool = parpool('local', numWorkers, 'IdleTimeout', Inf);

activities = [12 16 20];
pactivities = [5 7 9];
workers = [200 250 300];
skills = [5 10 15];

I learned a new information about Matlab. I've been trying for hours for problem and rng('shuffle') solved it. Thanks again :)

Answer 2

[Karl Cronburg]

You can get a random seed based on the current operating system time with the rng function:

rng("shuffle");

See the doc page here: https://www.mathworks.com/help/matlab/ref/rng.html

[Abdullah Türk]

Karl Cronburg, when every time I open Matlab every time, "how does it give me same processeing times" I said myself. But, it has a logical explanation, thanks for your contribution for the problem.

Answer 3

[Walter Roberson]

Every time you load matlab, it does the equivalent of rng('default') which initializes the seed to 0.

It is expected that the random number order is the same starting from the beginning of each MATLAB session.

See rng for more information on controlling the random number generator.

Caution: although you can use rng('shuffle') to initialize the seed based on the current time, doing so is not considered secure. Do not use it for producing seeds for cryptographic communications, or for producing seeds for drawing lottery numbers. Creating good seeds for secure purposes is hard

Answer 4

[Steven Lord]

Others have told you how to get the numbers to not repeat after startup. But to answer your "Why?" question, this is expected behavior. See this documentation page that explains why this happens.

[Abdullah Türk]

"Why Do Random Numbers Repeat After Startup?"

Steven Lord, thank you. I read the the article. There are very useful information in it. Thanks.