LAUSR

In this article, some new efficient and accurate algorithms are developed for solving linear and nonlinear odd-order two-point boundary value problems. The algorithm in linear case is based on the application of Petrov–Galerkin method. For implementing this algorithm, two certain families of generalized Jacobi polynomials are introduced and employed as trial and test functions. The trial functions satisfy the underlying boundary conditions of the differential equations, and the test functions satisfy the dual boundary conditions. The developed algorithm leads to linear systems with band matrices which can be efficiently inverted. These special systems are carefully investigated, especially their complexities. Another algorithm based on the application of the typical collocation method is presented for handling nonlinear odd-order two-point boundary value problems. The use of generalized Jacobi polynomials leads to simplified analysis and very efficient numerical algorithms. Numerical results are presented for the sake of testing the efficiency and applicability of the two proposed algorithms.