Comparative Analysis of Some Algorithms for Solving Problems of Multidimensional Constrained Optimization

Abstract

The paper presents the results of solving problem of a multidimensional constrained optimization by four methods: classical gradient descent, accurate analytical solution of the system, genetic algorithm, and own method “pulling up the middle”. The need for setting this problem arises when solving systems of linear algebraic equations, building regression models, training of artificial neural networks, etc. For the test optimization problem each of the above methods was consistently applied, which allowed to perform a comparative analysis and identify the advantages and disadvantages of the used algorithms.