How to produce Gaussian random variable between two vectors ?

Hello all, I am having two vectors:

sr = [1,2,2,2,3,3,3,4,5]; % various possible source
ta = [2,3,6,8,6,4,7,6,6]; % various possible targets

I want to generate Gaussian random variable (rv) between corresponding source and target so that we get 9 different Gaussian rv i.e., first rv between 1 and 2, second rv between 2 and 3, third rv between 2 and 6 and so on.

This is what I had tried so far:

clc;
clear all;
close all;
sr = [1,2,2,2,3,3,3,4,5]; % various possible source
ta = [2,3,6,8,6,4,7,6,6]; % various possible targets
for i = 1:length(sr)
    for j = 1:length(ta)
        nrm_rv1 = random('norm', 0, sqrt(2)); %generate Gaussian (Normal) RV 
    end                          
end

But I am not getting how to arrange nrm_rv1 so as to get 9 different Gaussian rv.

Any help in this regard will be highly appreciated.

Answer 1

[Torsten]

sr = [1,2,2,2,3,3,3,4,5]; % various possible source
for i = 1:numel(sr)
  pd{i} = makedist('Normal',0, sqrt(2)); %generate Gaussian (Normal) RV 
end
pd

pd = 1×9 cell array
  Columns 1 through 6

    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}

  Columns 7 through 9

    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}    {1×1 prob.NormalDistribution}

If you mean something else when you write "I want to generate Gaussian random variable (rv) between corresponding source and target", please explain it in more detail.

[charu shree]

Thank u sir for your answer...but how to acces values in cell array

[Torsten]

I don't know what you want to do with the 9 normal distributions.

Here are some gimmicks:

sr = [1,2,2,2,3,3,3,4,5]; % various possible source
for i = 1:numel(sr)
    pd{i} = makedist('Normal',0, sqrt(2)); %generate Gaussian (Normal) RV 
end
random(pd{1})

ans = -0.5965

cdf(pd{4},0.5)

ans = 0.6382

pdf(pd{6},4)

ans = 0.0052

Did you study this ?

https://de.mathworks.com/help/simevents/queue-service-and-route-modeling.html

[charu shree]

@Torsten Thank u so much sir...

Answer 2

[Luca Ferro]

I would do it like this (assuming i understood correctly):

sr = [1,2,2,2,3,3,3,4,5]; % various possible source

ta = [2,3,6,8,6,4,7,6,6]; % various possible targets

for ii = 1:size(sr,2) %assuming that sr and ta have the same dimension

lowThs=min(sr(ii),ta(ii));

highThs=max(sr(ii),ta(ii));

nrm_rv1(ii) = lowThs+abs(lowThs-highThs)*rand(1,1);

end

i'll break it down:

• lowThs is the lowest number of the 2 in the array (i assumed no masking is done to assure that the sr value is higher than the ta value)
• highThs is the highest, same reasoning
• rand(1,1) generates a random number between 0 and 1
• abs(lowThs-highThs) this gets the span of the range between the 2 numbers low/highThs
• abs(lowThs-highThs)*rand(1,1); so by doing this we get a random number from 0 to the span
• lowThs+abs(lowThs-highThs)*rand(1,1); by adding the lowThs we get a random number that spans from lowThs to lowThs+the random number in the spane, meaning from lowThs to highThs
• nrm_rv1(ii) this is just to store it in an array. First element is random between the first couple, second between the second and so on

Note that this code works even if ta>sr and ta==sr so that you have freedom on the source arrays. The only constraint is that they must have the same dimension. If you would like to have freedom in that as well just ask, i'll gladly fix the code.