Semenikhin

The authors consider solving the discrete problem of Pareto set and frontier optimization in the multi-criterion optimization problem. The aim of this work was to examine the efficiency of Adaptive Weighted Sum method (AWS) which was initially developed by J-H. Ryu, S. Kim and H. Wan. Comparing to the original research, the authors used a wider range of multi-criterion benchmark optimization problems. A range of limitations of AWS methodology and usage complexity was found. Modification of the method was proposed to resolve the problems and limitations.

This authors present a review of numerical methods of approximate Pareto set generation in the multi-objective optimization problem. The following methods are discussed: "naive" methods, switching objective functions methods, methods of objective functions aggregation, methods based on ranking of population agents, and other methods. All cases referred to methods involving the use of genetic or swarm algorithms, such as particle swarm optimization algorithm.

The paper is dedicated to an approximation of a set of attainability for type “trunk” multisectional robot-manipulator. SPSO method was developed and used for the task resolution. SPSO is based on well-known MOPSO method that is used for multi-objective optimization task. SPSO method as well as MOPSO method implements one of the version of particle swarm optimization method (PSO) that is often used for global optimization tasks. Sigma-algorithm is used for finding non-dominated solutions. SPSO effectiveness is demonstrated for several tasks of a set of attainability search for both with and without obstacles.

This paper describes implementation of a particle swarm optimization method designed for Pareto set for a multicriteria optimization problem. The paper describes the problem definition, algorithm description, diagram and research result of the method efficiency in solving two-criterion test problem.