Ga starts with infeasible states

I I have a polynomial objective function with nine decision variables. The first five variables are integers (the first three are discrete variables that I use index of a vector instead of their values as described here), and the rest are continuous. I have box constraints on all variables that indicate the integers cannot go beyond, i.e., 5. But, at the very beginning of the optimization, the state.population in GA codes generates infeasible states that do not satisfy the specified upper and lower bounds!

Is there anything that I have missed here?

p.s: I do not pass the IntCon constraint and use my createFnc to generate initial points. I sample the integers from multinomial distributions and continuous variables from uniform distributions Now, in the first iterations, I do not get the previous error, but it returns that the optimizer could not find the feasible solution. The objective is a polynomial function, and besides the linear constraints, I have polynomial constraints. After termination, the returned decision variables also satisfied all constraints, indicating a feasible solution within the constraints!

For your reference here is my lienar constrains:

A = [1     0     0     0     0     0     0     0     0
     0     1     0     0     0     0     0     0     0
     0     0     1     0     0     0     0     0     0
     0     0     0     1     0     0     0     0     0
     0     0     0     0     1     0     0     0     0
     0     0     0     0     0     1     0     0     0
     0     0     0     0     0     0     1     0     0
     0     0     0     0     0     0     0     1     0
     0     0     0     0     0     0     0     0     1
     1     1     1     0     0     0     0     0     0
    -1     0     0     0     0     0     0     0     0
     0    -1     0     0     0     0     0     0     0
     0     0    -1     0     0     0     0     0     0
     0     0     0    -1     0     0     0     0     0
     0     0     0     0    -1     0     0     0     0
     0     0     0     0     0    -1     0     0     0
     0     0     0     0     0     0    -1     0     0
     0     0     0     0     0     0     0    -1     0
     0     0     0     0     0     0     0     0    -1
    -1    -1    -1     0     0     0     0     0     0];
b = [   2.0000
    2.0000
    2.0000
    3.0000
    3.0000
    1.5603
    1.5605
    4.1158
    4.1174
    4.0000
   -1.0000
   -1.0000
   -1.0000
   -1.0000
   -1.0000
    1.1918
    1.1913
    1.1022
    1.1021
   -4.0000];
nonlineCntrs:   poly1(x6,x4)>= x8;
                poly2(x6,x4)<=x8;
                poly3(x5,x3)>=x7;
                poly4(x5,x3)<=x7;
                

In all situations, I am sure that poly1(x6,x4)>=poly2(x6,x4) and poly3(x5,x3)>=poly4(x5,x3)!