Subscripted assignment dimension mismatch in linprog optimization

Question

Stefano Di Vito 2018-3-29

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/391483-subscripted-assignment-dimension-mismatch-in-linprog-optimization

回答： Alan Weiss 2018-4-2

Hello everybody! I get the following error while performing my linear program (in order to calculate the objective function value on 100 different points.) Here's the code:

T=size(X,1);
N=size(X,2);
M=2+N;
tau=zeros(N,P);
Xtau=zeros(P,T);
H=zeros(P,T+M);
for j=1:P % P is the number of portfolios
  tau(:,j)=matpesi(:,j); % weights vector 
  Xtau(j,:)=(X*tau(:,j))'; %obj function variables Beta-coeff. row-vector
H(j,:)=[1 -1 zeros(1,N) -1/T*(Xtau(j,:))]; % each row of the matrix h contains the coefficients of objective function
Aeq=[ones(N,1) -ones(N,1) -eye(N) -1/T*(X')]; 
beq=zeros(N,1);
C=-eye(T,T); 
for s=1:T-1
  C(s,s+1)=1;
end
A=[zeros(T,M) C; -eye(M,T+M) ];
b=[-1/T*ones(T,1); zeros(M,1)];
end
x=zeros(T+M,P);
for c=1:P
  f=H(c,:);
  options = optimoptions('linprog','Algorithm','interior-point','Display','final','MaxIterations',10e5);
  [x(:,c),fval(:,c),exitflag]=linprog(f,A,b,Aeq,beq)
  fval(fval<1.0000e-6)=0.0;
end
end

Once i run the code i get the error written in the title. I'm facing problems at indexing the "f" coefficient vector. To be clearer, i should perform the LP for each of the 100 rows of matrix H( each row should be a "f"coeff. vector for the function linprog.) This means i should perform the test on each of the 100 portfolios. Literally, i should obtain a vector of fvals, as each one of its elements should tell me if the correspondant portfolio is efficient or not under second order stochastic dominance assumptions. Anyone could help me? Thank you in advance