Author(s) | Collection number | Pages | Download abstract | Download full text |
---|---|---|---|---|
Kudriashova A. V. | № 2 (69) | 21-28 |
![]() |
![]() |
Evolutionary multicriteria optimization is concerned with finding optimal solutions that satisfy a set of objective functions. The main concepts include dominant and non-dominant (pareto-optimal) solutions. A solution x is considered dominant when there is another admissible solution y that is not worse than x for all objective functions.
Modern methods in this area are divided into traditional and evolutionary. The first group includes the weighted objective method and the constraint method. Evolutionary algorithms are represented by various models, including VEGA (Genetic Algorithm with Vector Evaluation), SPEA (Strength Pareto Evolutionary Algorithm), and other methods and algorithms. Evolutionary algorithms are characterized by flexibility in applying to complex problems. They imitate natural selection, providing a search for a set of solutions simultaneously. This makes it possible to work effectively with continuous and discrete problems.
Solving real-world multi-objective optimization problems using multi-objective optimization algorithms becomes difficult when the number of objectives is large, because the types of algorithms commonly used to solve these problems are based on the concept of non-dominance, which stops working as the number of objectives increases. This problem is known as the curse of dimensionality. At the same time, the presence of many objectives, which is typical for practical optimization problems, makes it very difficult to choose a solution. Various approaches are used in the literature to reduce the number of objectives required for optimization.
Keywords. multicriteria optimization, evolutionary algorithms, pareto-optimal solutions, curse of dimensionality, weighted objective method.
doi: 10.32403/1998-6912-2024-2-69-11-20