Additive Schwarz algorithms for neural network approximate solutions
Additive Schwarz algorithms are proposed as an iterative procedure for neural network approximate solutions of partial differential equations. Based on the convergence analysis of the additive Schwarz algorithms in a general Hilbert space setting, the convergence of the neural network approximate solutions is analyzed for the one-level and two-level iterative schemes. Numerical results of the proposed methods are presented for test examples.
READ FULL TEXT