twpbvpl

Algorithms for the Solution of Two-Point Boundary Value Problems: The Fortran 77 code TWPBVP was originally developed by Jeff Cash and Margaret Wright and is a global method to compute the numerical solution of two point boundary value problems (either linear or non-linear) with separated boundary conditions. In the code TWPBVP, MIRK schemes of orders 4, 6 and 8 are solved in a deferred correction framework in an attempt to give a solution accurate to a prescribed local tolerance at a discrete set of mesh points. TWPBVP has been found to be especially effective at solving non-stiff and mildly-stiff ODEs efficiently. For problems of a greater stiffness, fully implicit Runge Kutta schemes have been found to be more suitable (although requiring more work per step) than MIRK schemes. A deferred correction code TWPBVPL based on Lobatto IIIA schemes of orders 4, 6 and 8 has been developed. For stiff problems it is recommended that this code should be tried first. ...