How to solve this problem using PSO?

2 vues (au cours des 30 derniers jours)
Jeong_evolution
Jeong_evolution le 13 Sep 2017
Commenté : Alan Weiss le 14 Sep 2017
Hi all,
I have some problem using pso algoithm.
fcn = @(x) x(1) + x(2) + x(3)
lb = [-40 -40 -40];
ub = [40 40 40];
nvars = 3;
x = particleswarm(fun,nvars,lb,ub);
In this code, I want to 8.8 <= x(1)+x(2)+x(3) <= 79.2 (constraint)
But I don't know use this expression.
Plz help ..

Réponse acceptée

Alan Weiss
Alan Weiss le 13 Sep 2017
Unfortunately, particleswarm does not currently support linear constraints. You really should use linprog for this kind of problem.
You can also try a penalty function, say fcn = 1e6 when sum(x) < 8.8 or sum(x) > 79.2.
But linprog is the best solver for linear problems with linear constraints.
Alan Weiss
MATLAB mathematical toolbox documentation
  2 commentaires
Jeong_evolution
Jeong_evolution le 14 Sep 2017
Excuse sir, Should I know Why can not pso do it?
Alan Weiss
Alan Weiss le 14 Sep 2017
There are two solvers in Global Optimization Toolbox that support only bound constraints: particleswarm and simulannealbnd. Similarly, there are solvers in Optimization Toolbox that do not support constraints or support only bound constraints: fminunc, lsqcurvefit, fsolve, and others. Choose a solver that works for you.
And even more, for smooth problems, use an Optimization Toolbox solver, not a Global Optimization Toolbox solver.
Alan Weiss
MATLAB mathematical toolbox documentation

Connectez-vous pour commenter.

Plus de réponses (0)

Catégories

En savoir plus sur 비선형 최적화 dans Help Center et File Exchange

Community Treasure Hunt

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

Start Hunting!