Возможности и перспективы применения методов искусственного интеллекта для решения краевых задач математической физики в инженерной практике
The history of evolution of methods for solving boundary value problems of solid mechanics is traced, the comparative analysis of methods from the point of view of reliability of the obtained solutions is carried out. An attempt is made to develop the method of fictitious canonical domains by applying the technology of genetic algorithms. On the example of the boundary value problem of elasticity theory it is shown that the use of genetic optimization algorithm instead of gradient allows to significantly reduce the error of solving boundary value problems performed by the method of fictitious canonical domains.
Keywords: boundary value problem, elasticity theory, analytical method, numerical method, error, artificial intelligence, genetic algorithm