There are many constraint handling methods, for example if you have an equality constraint you can reduce the number of variables by the number of equality constraints you have. Consider a minimization problem with objective function being f(x1,x2,...,xn), and suppose you have an equality constraint given by a1x1+a2x2+...anxn=b, from this xn=b-a1x1-a2x2+...an-1xn-1. From this your objective function will be f(x1,x2,...,xn-1) and no more equality constraint.
you can also use penalty method, or feasibility check at each run of the algorithm (here you have to make sure your initial solutions are feasible)