MATLAB Answers

Nonlinear Constraints in Multijob optimization

17 views (last 30 days)
Panagiotis Koronaios
Panagiotis Koronaios on 7 Apr 2021
Hello,
I am running a multijob optimization with some design variables (3 or even more like 15 let's say x1,x2,...,x15) and I want to minimize my objective function let's say f. With every set of (x1,x2,...,x15) a specific array can be computed (let's say Y) and I want my constraint to be the elements of Y be always descending (Y1<Y2<...<Yn) . I thought that the solution to this was to introduce a variable k that when it is equal to zero then Y would be descending and when it k equals to 1 then Y would not be descending but it did not work. I got results that did not take into consideration this constraint. How can I prevent matlab from giving me such a result?
Another relevant question is how to say to matlab that ''dont give me results that f (objective function) is biger than 0.2 for example...
Thank you

Answers (1)

Alan Weiss
Alan Weiss on 8 Apr 2021
Your constraints can be formulated as nonlinear constraints. Make sure that you read and follow the note on that page.
Alan Weiss
MATLAB mathematical toolbox documentation
  1 Comment
Panagiotis Koronaios
Panagiotis Koronaios on 8 Apr 2021
function [C Ceq] = nonlinear_constraints(x)
[....]
flag=0;
for i=1:numel(Y)-1
if Y(i+1)>Y(i)
flag=flag+1;
end
end
C = flg-5;
Ceq = [];
The truth is that my design variables and the restriction is not directly connected as you can see above. I want to prevent matlab from giving my results that C>=5 but this constraint is ignored. Do you know how to fix this?

Sign in to comment.

Community Treasure Hunt

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

Start Hunting!