In a multiobjective problem of integer linear programming, parametrization of optimality principle is introduced by dividing a set of objectives into a family of disjoint subsets. The introduction of this principle makes it possible to connect two classical optimality sets, namely, extreme and Pareto. The admissible independent perturbations in such a problem are formed by a set of additive matrices, with arbitrary H¨older’s norms specified in the solution and criterion spaces. The lower and upper bounds for the radius of stability are obtained. The main result is complemented with several important corollaries.