Programme systems to estimate the Pareto-approximation quality in the problem of multi-criteria optimization. A review.
V. V. Belous, С. В. Грошев, Anatoly Karpenko, I. A. Shibitov · Science and Education of the Bauman MSTU · 2014
The paper considers a relatively new and rapidly growing class of multi-criteria optimization (MCO) methods. This class is based on the preliminary creation of a finite-dimensional approximation of the Pareto set and, subsequently, the Pareto-approximation of MCO problem. Since many different methods are known, there is a problem to compare these methods with each other. Solving real world MCO problems one often faces more than two criterion functions. This fact makes it difficult to perform a visual analysis of the proper Pareto-approximation. Therefore, A lot of quality indicators of Pareto-approximation have been developed. In this work a statement of multi-criteria optimization problem was presented along with a brief review of the specified class of methods and existent quality indicators. The paper itself is devoted to the review of existent programme systems, designed for building Pareto-approximation, which implement various algorithms of quality estimation.