Abstract. In this study, the cubic B-spline approximation equation has been derived by using the cubic B-spline discretization scheme to solve two-point boundary value problems. In addition to that, system of cubic B-spline approximation equations is generated from this spline approximation equation in order to get the numerical solutions. To do this, the Accelerated Over Relaxation (AOR) iterative method has been used to solve the generated linear system. For the purpose of comparison, the GS iterative method is designated as a control method to compare between SOR and AOR iterative methods. There are two examples of proposed problems that have been considered to examine the efficiency of these proposed iterative methods via three parameters such as their number of iterations, computational time and maximum absolute error. The numerical results are obtained from these iterative methods, it can be concluded that the AOR iterative method is slightly efficient as compared with SOR iterative method.
IntroductionBoundary value problems plays a important roles to apply and solve many application of science and engineering phenomenan. Researchers in science, physics and engineering get more advantages from the numerical solutions which obtained from the two-point boundary value problems. Actually there are various methods are used to get the solutions in boundary value problems such as families of Galerkin methods namely Sinc-Galerkin method [1] and hybrid Galerkin method [2], Adomain decomposition method [3] and shooting method [4]. The other researchers also have been used the families of B-spline [5,6,7] to solve the same problems. B-spline scheme was considered in this paper to apply with the aim of discretizing the proposed problem.To obtain the cubic B-spline approximation equation of the proposed problem, firstly, the proposed problem needs to be discretized via family of spline or more specifically by imposing cubic B-spline discretization scheme. Then the approximation equation can be derived and used to contruct a linear system. Next, in this paper, the linear system can be solved via the iterative methods. This is because of motivation given by Young [8], Hackbush [9] and Saad [10] in which they have been proposed and discussed about the various iterative methods to solve any linear system.As mentioned in the second paragraph, the utmost objective of this paper is used the cubic Bspline approximation equation to solve two-point boundary value problem with attention to investigate the application of AOR iterative method. To analyze the performance of AOR, the implementation of the SOR iteration family method has been used as control methods.