Power series neural network solution for ordinary differential equations with initial conditions

Differential equations are very common in most academic fields. Modern digital control systems require fast on line and sometimes time varying solution schemes for differential equations. This paper presents new nonlinear adaptive numeric solutions for ordinary differential equations (ODE) with initial conditions. The main feature is to implement nonlinear polynomial expansions in a neural network-like adaptive framework. The transfer functions of the employed neural network follow a power series. The proposed technique does not use sigmoid or tanch non-linear transfer functions commonly adopted in conventional neural networks at the output. Instead, linear transfer functions are employed which leads to explicit power series formulae for the ODE solution. This allows extrapolation and interpolation which increase the dynamic numeric range for the solutions. The improved and accurate solutions for the proposed power series neural network (PSNN) are illustrated through simulated examples. It is shown that the performance of the proposed PSNN ODE solution outperforms existing conventional methods.

[1]  R. F. Hoskins,et al.  Ordinary Differential Equations of Order > 1 , 1973 .

[2]  W. Chew,et al.  Coupled Integral Equations for Microwave Induced Elastic Wave in Elastic Media , 2010, IEEE Transactions on Antennas and Propagation.

[3]  Summarizing the Method of Solution of the Majority of Differential Equation by Using Tensor Product B-Spline Wavelet Scaling Functions in Network Security and Software Security Area, the Imitating Von Neumann Stability Analysis and a Typical Example , 2009, 2009 International Conference on Intelligent Human-Machine Systems and Cybernetics.

[4]  George Lindfield,et al.  Numerical Methods Using MATLAB , 1998 .

[5]  E. K. Ifantis Analytic solutions for nonlinear differential equations , 1987 .

[6]  P. Hartman Ordinary Differential Equations , 1965 .

[7]  Stanley I. Grossman,et al.  Elementary Differential Equations , 1997 .

[8]  Zhiyong Yu,et al.  A Pontryagin's Maximum Principle for Non-Zero Sum Differential Games of BSDEs with Applications , 2010, IEEE Transactions on Automatic Control.

[9]  W. F. Trench Elementary Differential Equations , 1969 .

[10]  Zeng Zhe-Zhao,et al.  The Algorithm of Neural Networks on the Initial Value Problems in Ordinary Differential Equations , 2007, IEEE Conference on Industrial Electronics and Applications.

[12]  Andrew J. Meade,et al.  The numerical solution of linear ordinary differential equations by feedforward neural networks , 1994 .

[13]  P. Mythili,et al.  ECG noise removal using GA tuned sign-data least mean square algorithm , 2012, 2012 IEEE International Conference on Advanced Communication Control and Computing Technologies (ICACCCT).

[14]  Jerry M. Mendel,et al.  Structured trainable networks for matrix algebra , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[15]  Hyuk Lee,et al.  Neural algorithm for solving differential equations , 1990 .

[16]  Qing-Jiang Chen,et al.  Numerical solution of differential equations by using Haar wavelets , 2007, 2007 International Conference on Wavelet Analysis and Pattern Recognition.

[17]  A. H. Khater,et al.  Potential symmetry and invariant solutions of Fokker–Planck equation modelling magnetic field diffusion in magnetohydrodynamics including the Hall current , 2004 .

[18]  Potential symmetry and invariant solutions of Fokker-Planck equation in cylindrical coordinates related to magnetic field diffusion in magnetohydrodynamics including the Hall current , 2005, gr-qc/0510099.

[19]  Mona E. Zaghloul,et al.  VLSI implementation of locally connected neural network for solving partial differential equations , 1996 .