fmincon has converged to an infeasible point

My objective function is as follows :

 [x,fval,exitflag,output]=fmincon(@(x)CalcLatency(x,SwitchArray),x0,[],[],[],[],lb,ub,@(x)integercon(x,ContNum,Capacity),options);

The parameter x is a matrix of integer values. The objective function tries to find minimum average distance between locations represented by the first row of x and locations represented by the rest of the matrix (i.e a cellular network with an assignment between transmitters and receivers). The nonlinear constraint assures that the number of links does not exceed a certain capacity.

I'm having difficulties finding the right initial guess and search tolerances. Therefore I get a message says "fmincon has converged to an infeasible point".

Please advise.

Answer 1

[John D'Errico]

Fmincon CANNOT be used to solve integer problems. Fmincon requires a continuous, differentiable objective function. Therefore, integer constraints make use of fmincon invalid.

[Aviram Zilberman]

thanks for the quick answer. Since I need only integer numbers I'm using the rounded values that the objective function produces, but the objective function is continuous and differentiable.

[John D'Errico]

But you are apparently trying to use integer constraints in some form. Is that not what this function does?

@(x)integercon(x,ContNum,Capacity)

After all, you have provided no information at all except what you typed here.