Originally posted by: VahidDehkordi

I am trying to replace intlinprog matlab function with cplexmilp but I don't know what's wrong! here is a simple code and it does not find a feasible answer in cplex, however, intlinprog can solve it.

clc
clear
f = [-1 -2 -3]';
Aeq = [-1 1 1; 1 -3 1];
beq = [20 30]';
lb = [0 0 0]';
ub = [20 90 30]';
A = [0 0 1; 1 0 0];
b = [30 200]';
intcon = [1 2 3];
% matlab solver
%  [x, fval, exitflag] = intlinprog(f, intcon, A, b, Aeq, beq, lb, ub);
% Cplex Solver 
     sostype=[];
     sosind=[] ;
     soswt=[];
     ctype = [];
      for i=1:length(A),
     ctype = [ctype;'I'];
      end
  ctype = ctype'; 
   [x,fval,exitflag,output] = cplexmilp(f,Aeq,beq,A,b,sostype,sosind,soswt,lb,ub,ctype)

Results: cplexstatus: 103 cplexstatusstring: 'integer infeasible' iterations: 0 algorithm: 12 time: 0 message: 'No feasible point was found.'

I think I made a very simple mistake but I can not find it myself. Can somebody give me a comment about what's wrong with my code!

You mixed up the order of arguments for cplexmilp(). The first matrix and vector are for inequality constraints, the second matrix and vector are for inequality constraints. So you should write

cplexmilp(f,A,b,Aeq,beq,sostype,sosind,soswt,lb,ub,ctype)

Originally posted by: VahidDehkordi

Thank you so much. you are right although when I am writing the function in matlab it says in this order ( in pic) but changing the order did work.