
Determination initial approximations in solving the problem of numerical optimization of a large-sized space structure using linear extrapolation of optimal solutions
Author(s) -
В. В. Салмин,
Konstantin V. Peresypkin,
Alexey Chetverikov,
Ivan Tkachenko
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1745/1/012097
Subject(s) - extrapolation , mathematical optimization , mathematics , point (geometry) , set (abstract data type) , space (punctuation) , optimization problem , boundary (topology) , numerical analysis , boundary value problem , iterative and incremental development , computer science , mathematical analysis , geometry , operating system , software engineering , programming language
The parameters of a large-sized space structure are determined using numerical optimization using the gradient method. The complexity of the boundary of the constraints makes the optimal solution poorly conditioned relative to the initial point of the iterative process. The initial point is selected based on the optimal solutions found with an incomplete set of design variables. The initial point is found by linear extrapolation of these optimal solutions. In the problem under consideration, this approach made it possible to obtain a better locally optimal solution than was possible with the help of an intuitive choice of the starting point.