Use of cubic splines and the second central finite differences in numerical solution of PDEs (Q4943264)

The new method for the approximation of second derivatives using both cubic spline and central finite difference is proposed. The second derivative \(M_i \) of the cubic spline at the point \(x_i \) and the second central finite difference \(\frac{\delta ^2 u(x_i)}{h^2}\) of the function \(u(x_i)\) approximate \(u''(x_i)\). By using this property, the number \(C_i \) is introduced to be the average of \(M_i \) and \(\frac{\delta ^2 u(x_i)}{h^2 }\). It is concluded that \(C_i \) gives a better approximation to \(u''(x_i)\). Finally, the \(C_i \) is used in the numerical solution of second-order PDEs. Numerical examples are given and results obtained using the proposed method, the spline method, and the finite difference method are compared.