QA323 : Regularized symmetric positive definite matrix factorizations for solving differential Equations with radial basis functions
Thesis > Central Library of Shahrood University > Mathematical Sciences > MSc > 2015
Authors:
Sedigheh Ahmadi [Author], Mehdi Ghovatmand [Supervisor], Hojjat Ahsani Tehrani[Advisor]
Abstarct: We introduced the radial basis function first and interpolation of function by using radial basis functions were explained. By interpolation of this function, derivation o of function were determined by using direct and indirect methods. Then the device in interpolation of radial function that, was created by using direct factorization methods of LU, LLT and LDLT was solved. Cholesky factorization was not a good way of solving the problem because of some reasons, so we used the increase of diagonal and Rayleigh method. In this way the Cholesky factorization was usable and LDLT has better results and the created graph has less vacillations. In another part, the device that was not usable by stochastic method because of not being NSPD , so another method. That is called conjugate gradient method has been used to solve the problem. At the end of solving non-linear differential equations depending on time, by changing to linear and solved by radial basis functions network method.
Keywords:
#radial basis functions #interpolation function #solving device factorization methods #solving differential equations Link
Keeping place: Central Library of Shahrood University
Visitor: