Please help me to using genetic algorithm
1 vue (au cours des 30 derniers jours)
Afficher commentaires plus anciens
I write this code but I want to solve this problem with 'ga' not with 'intlinprog' solver!
Can anyone guide me?
costprob = optimproblem;
% Indices
k = 15;
j = 2;
f = 10;
l = 5;
r0 = 6;
r = 6;
% Parameters
cr0 = 0 + 1*rand(1,r0);
dr0f = 0 + 1*rand(r0,f);
csl = 0 + 1*rand(1,l);
DE = 200 + 100*rand(1,1);
csur0f = 2000 + 1000*rand(r0,f);
ctl = 1000 + 1000*rand(1,l);
cvl = 10 + 10*rand(1,l);
cpjk = 0 + 1*rand(j,k);
corj = 0 + 1*rand(r,j);
pr0f = 0 + 1*rand(r0,f);
vjk = 0 + 1*rand(j,k);
cvjrk = 0 + 1*rand(j,r,k);
M = 10000000000000;
% Variables
xl = optimvar('xl',1,l,'LowerBound',0);
yr0f = optimvar('yr0f',r0,f,'Type','integer','LowerBound',0,'UpperBound',1);
xx1r0f = optimvar('xx1r0f',r0,f,'LowerBound',0);
xx2r0f = optimvar('xx2r0f',r0,f,'LowerBound',0);
yjk1 = optimvar('yjk1',j,k,'Type','integer','LowerBound',0,'UpperBound',1);
yl2 = optimvar('yl2',1,l,'Type','integer','LowerBound',0,'UpperBound',1);
zjkr = optimvar('zjkr',j,k,r,'LowerBound',0);
wrj = optimvar('wrj',r,j,'LowerBound',0);
% Objective function
objfun1 = sum(sum(dr0f.*xx1r0f,2).*cr0',1);
objfun2 = sum(sum(corj.*wrj,2),1);
objfun3 = sum(sum(pr0f.*xx1r0f,2),1);
objfun4 = sum(sum(cpjk.*yjk1,2),1);
objfun5 = sum(csl.*xl,2);
costprob.Objective = objfun1 + objfun2 + objfun3 + objfun4 + objfun5;
% Constraints
cons1 = sum(xl,2) >= DE;
cons2 = sum(xl,2)*ones(j,1,r) == sum(zjkr,2);
cons3 = xx1r0f <= csur0f.*yr0f;
cons4 = xl <= ctl.*yl2;
cons5 = xl >= cvl.*yl2;
cons6 = sum(yjk1,2) == ones(j,1);
cons7 = squeeze(sum(zjkr,3)) <= M*yjk1;
cons8 = (1-dr0f).*xx1r0f == xx2r0f;
costprob.Constraints.cons1 = cons1;
costprob.Constraints.cons2 = cons2;
costprob.Constraints.cons3 = cons3;
costprob.Constraints.cons4 = cons4;
costprob.Constraints.cons5 = cons5;
costprob.Constraints.cons6 = cons6;
costprob.Constraints.cons7 = cons7;
costprob.Constraints.cons8 = cons8;
2 commentaires
Réponse acceptée
Matt J
le 21 Oct 2019
Modifié(e) : Matt J
le 21 Oct 2019
You can use prob2struct to obtain most of your problem parameters in solver form,
problem=prob2struct(costprob);
problem=rmfield(problem,'solver');
problem.nvars=numel(problem.lb);
problem.fitnessfcn=@(x) dot(problem.f,x);
x=ga(problem);
However your problem, as currently formulated, has both integer and equality constraints, which ga cannot handle. See here, for guidelines on how to rewrite the problem without equality constraints:
5 commentaires
Plus de réponses (0)
Voir également
Catégories
En savoir plus sur Genetic Algorithm 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!