Alternative to EVAL string expressions in optimization

Hi,
I have a huge matrix, each element is a string like "16*q^2+4*q", i call it new_sol_array, now i want to define multiple optimconstr object with each row one set of constraints. This is what i did:
constr= optimconstr(M,N);
for i=1:rows
for j =1:column
constr(i,j)=eval(new_sol_array(i,j)+'==0');
end
end
As you would see, this is extremely slow. Is there a better alternative to do the same thing? i have read some of the answers about eval on this website yet did not find anything relevant enough to my problem.
Thank you so much!!!

 Accepted Answer

Instead of eval() everything, use file operations to emit a file along the lines of
constr = [
EQN, EQN,... EQN
EQN, EQN,... EQN
]
along with a function header. Then "clear" the name of the file. Then invoke the file as a function passing in q and whatever else is required
This reduces the burden by not needing to eval() each one individually.

4 Comments

Thank you so much for your comment Walter! I'm not sure i understand your answer correctly so i'm just gonna repeat what you said and please correct me if i'm wrong.
1. Write a text file maybe with content like
constr = [
16*q^2+4*q==0, 25*q*t+20*t==0,... EQN
EQN, EQN,... EQN
]
2. Read this text file, substitute in variables (this step i'm actually not sure how to do, could you be more specific?)
3. Evaluate this one-hell-of-a-long line of command only once by eval.
Is that so? Appreciate it!
Just to update. I tried
temp = new_sol_array+'==0';
temp = join(join(temp,2),';',1);
temp = 'constr=['+temp+']';
eval(temp);
It took me less than 3 min to assign 10000 by 10 constraints. Not perfect but definitely acceptable.
  1. yes
  2. no
  3. no
Did you try writing a function file and calling it, as Walter Roberson described?
Thank you Stephen, I did not do that simply because I don't get the idea. What kind of function file are you talking about? I have only a string array, why would a function takes variable names as input help?

Sign in to comment.

More Answers (0)

Categories

Products

Asked:

on 30 May 2022

Commented:

on 31 May 2022

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!