Title :
Solving inverse problems in nonlinear PDEs by recurrent neural networks
Author :
Uchiyama, Tadasu ; Sonehara, Noboru
Author_Institution :
NTT Corp., Kanagawa, Japan
Abstract :
A neural network approach for solving inverse problems in nonlinear partial differential equations (PDEs) is proposed, and a computer simulation based on this approach is described. The network is designed based on the differential difference equation (DDE) approximating the PDE. The network is trained so that its output and the known boundary values of connection weights and thresholds represent the approximated coefficients of the PDE governing the system. Simulation shows that the adjustable connections converge to the approximate values of the original coefficients identifying the system
Keywords :
inverse problems; nonlinear differential equations; partial differential equations; recurrent neural nets; adjustable connections; approximate values; approximated coefficients; boundary values; connection weights; differential difference equation; inverse problems; nonlinear PDEs; partial differential equations; recurrent neural networks; Boundary conditions; Computer simulation; Difference equations; Differential equations; Intelligent networks; Inverse problems; Neural networks; Nonlinear equations; Partial differential equations; Recurrent neural networks;
Conference_Titel :
Neural Networks, 1993., IEEE International Conference on
Conference_Location :
San Francisco, CA
Print_ISBN :
0-7803-0999-5
DOI :
10.1109/ICNN.1993.298524
