Stability and Effective Algorithms for Solving Multiobjective Discrete Optimization Problems with Incomplete Information

摘要

The stability problem of vector discrete optimization problems with different principles of optimality with respect to perturbations of all input data of the problem is investigated based on the obtained results on the stability kernel property and the subset of those feasible solutions which steadily do not belong to the optimum set. We present the review of the latest results concerning the estimations of stability radius of solutions of Boolean multicriteria problems with nonlinear criteria. For the problem with the known optimum value of the goal function there is constructed the algorithm with the best known guaranteed estimation. The described scheme applied group technologies and dynamic lower estimations for the optimum value of the objective functional which can be used in various versions of problems with incomplete information.