Multilevel Parallel Computations for Solving Multistage Multicriteria Optimization Problems

In the present paper, a novel approach for solving the computationally costly multicriteria optimization problems is considered. Within the framework of the developed approach, the obtaining of the efficient decisions is ensured by means of several different methods for the scalarization of the efficiency criteria. The proposed approach provides an opportunity to alter the scalarization methods and the parameters of these ones in the course of computations that results in the necessity of multiple solving the time-consuming global optimization problems. Overcoming the computational complexity is provided by reusing the computed search information and efficient parallel computing on high-performance computing systems. The performed numerical experiments confirmed the developed approach to allow reducing the amount and time of computations for solving the time-consuming multicriteria optimization problems.