Generating random 3D vectors

Hi, I'm trying to create 100 random 3D vectors with entries [-1,1]. I'm new to matlab so it's very confusing to me at the moment.

I would very much appreciate your help.

Thank you.

Answer 1

[John D'Errico]

rand(100,3)*2-1

Read the help. It shows how to do exactly this type of thing.

doc rand

I'm not sure why you are confused though. Rand generates random numbers between 0 and 1. Multiply by 2, and they are between 0 and 2. Subtract 1, and what do you have?

The above assume that you intended elements that lie in the interval [-1,1]. That is what that notation shows of course. If you really intended only the integer elements -1 and 1 as possibilities, then you should say so clearly. But that too is easy, and is a trivial extension of the above.

round(rand(100,3))*2-1

Then you want to generate 100 sets of 3 such numbers. A vector is just a list of three numbers. 100 of them is an array, so store them in a100x3 array.

Answer 2

[James Tursa]

E.g., see this discussion:

http://www.mathworks.com/matlabcentral/answers/306783-a-matlab-code-of-getting-a-value-uniformly-at-random-from-a-the-set-1-1

If you mean a 3-element vector (i.e., a vector in 3-space), then just:

v = 2*round(rand(3,100))-1;  % for values from the discrete integer set -1,+1

or

v = 2*rand(3,100)-1;  % for floating point values between -1 to +1

If you mean 3-element vectors with endpoints uniformly distributed on a unit sphere, then you would need to do something else. E.g. using the randn function:

v = randn(3,100);
v = bsxfun(@rdivide,v,sqrt(sum(v.*v)));

Answer 3

[Charles Jing]

I guess is Jason asking for this?

x=rand(100,100,100)*2-1;

[John D'Errico]

Surely not so. That generates an array of size 100x100x100. They are not sets of random numbers in a 3-dimensional space, but a 3-dimensional array of numbers. Had you said

x = rand(100,3)*2-1;

or

x = rand(3,100)*2-1;

then your answer would have been correct. In these two cases, you can now view the rows (or columns as appropriate) as sets of random numbers in a 3-dimensional space, so R^3.

Remember, the request was to generate 100 random 3-d vectors.

[Charles Jing]

Agree John, thanks!