Variables input generation for Minimization in High dimensional space

Question

I'm currently doing optimization in a high dimensional space. I was thinking of giving FindMinimum (or NMinimize not certain yet) a shot, however what bugs me is how I can state the command, given that its input a vector of size ~200.

Most examples for FindMinimum have functions of at most 2 variables.

Now I've seen this answer on StackOverflow where the user generates all his variables using

vars = Array[Symbol["x$" <> ToString[#]] &, n];

However reading the comments showed that there were certain concerns with this method:

I would not recommend doing it in a program, but it is safe to do in an interactive session.

So is this standard, or is there a more clever solution to get to it ?

@shrx It is not about assigning. Creation of symbols is solved, comment's to linked answer are OP's concern. — Kuba, Dec 11 '15 at 17:07

score 1 · Answer 1 · edited Dec 11 '15 at 16:46

1

The point being made in the quoted answer was to not use the name$number convention, because Module usues it, and who knows what conflicts might happen internally if you were to generate such variables. Using x$$$s in place of x$ would be fine.

edited Dec 11 '15 at 16:46

m_goldberg

107,779
16
103
257

answered Dec 11 '15 at 09:07

Kuba

136,707
13
279
740

1

Using Unique would also be acceptable. – Oleksandr R. Dec 12 '15 at 01:42

Variables input generation for Minimization in High dimensional space

1 Answers1