Journal of Applied Sciences1812-56541812-5662Asian Network for Scientific Information10.3923/jas.2007.2812.2817HayatiMohsenKaramiBehnam122007719In this study a new method based on neural network has been developed for solution of differential equations. A modified neural network is used to solve the Burger’s equation in one-dimensional quasilinear partial differential equation. This method is generally applicable to nth order partial differential equations on a finite domain with boundary conditions. The results obtained by this method, have been compared with the exact solution and found to be in good agreement with each other and because of superior properties of neural network i.e., parallel processing thereby less computational cost, this method has advantages over conventional methods.]]>Arfken, G. and H. Weber,1995Cichocki, A. and R. Unbehauen,1993Cruito, M., C. Monterola and C. Saloma,2001Hagan, M.T. and M.B. Menhaj,1994Haykin, S.,1999Joes, G. and I. Freeman,1986Kunz,. K.S. and R.J. Luebbers,1993Kutluay, S., A. Esen and I. Dag,2004Lorentz, G.G.,1976Luo, Z.,1991Monterda, C. and C. Saloma,1998Press, W., S. Flannery, S. Teakolesky and W. Vetterling,1986Saarineen, S., R. Bramley and G. Cybenko,1991Saif, S., H.M. Abbas, S.M. Nassar and A.A. Wahdan,2007Yam, J. and T. Chow,2000