Exiting: One or more of the residuals, duality gap, or total relative error has grown 100000 times greater than its minimum value so far:

Dear All I am running my code but I faced this problem Exiting: One or more of the residuals, duality gap, or total relative error has grown 100000 times greater than its minimum value so far: the primal appears to be infeasible (and the dual unbounded). (The dual residual < TolFun=1.00e-008.) Could you please suggest what shall I do? It is so urgent

Thanks again for any suggestions Regards

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/29922-why-your-question-is-not-urgent-or-an-emergency

Answer 1

[Titus Edelhofer]

Hi,

I guess (!) you are using linprog and get this error message? In this case the combination of bounds, equality constraints and inequality constraints are not permitting a solution. Try the following: either run linprog with bounds and inequality constraints only (setting Aeq=[], beq=[]), then with bounds and equality constraints only (setting A=[], b=[]). If one of them works, the other fails, start to take only parts of the "bad" constraints (i.e., some rows of Aeq and beq or A and b) to find out) which are the offending rows ...

Titus

Answer 2

[Bestun]

Dear Titus

Thanks a lot for you answer. Yes I am using linprog in my code. this is the command line

[variables, solution] = linprog(obj, [], [], equalityMatrix, equalityRHS, lowerB, upperB);

I used the first suggestion that you wrote in you message but I am still having the same problem.

the lines below are the section that I got the message error from.

function [variables, solution] = solve(B, N, padN, fD, fL, tied, objP);
bigNumber = 1000; warning off MATLAB:divideByZero;
[padB padFL padTied] = deal(sparse(size(B, 1), size(N, 2)), sparse(1, size(N, 2)), sparse(size(tied, 1), size(N, 2)));
equalityMatrix = [B padB; padN  N; fL padFL; tied padTied];
equalityRHS = sparse(size(equalityMatrix, 1), 1);
equalityRHS(size(B, 1) + size(N, 1) + 1, 1) = 1;
obj = [fD; objP']; % objP' contains plastic multiplier terms
lowerB = [-bigNumber * ones(size(B, 2), 1); sparse(size(N, 2), 1)];
upperB = [bigNumber * ones(size(B, 2), 1); bigNumber * ones(size(N, 2), 1)];
[variables, solution] = linprog(obj, [], [], equalityMatrix, equalityRHS, lowerB, upperB);

Just I am asking is there is any change could be done in the linprog line to solve this error

I look forward to hearing from you

Regards

Bestun

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup