A Collocation Based Block Multistep Scheme without Predictors for the Numerical Solution Parabolic Partial Differential Equations
Authors: Ehigie Julius1
Affiliations: 1. Department of Mathematics, Faculty of Science University of Lagos, Nigeria
Abstract
Introduction: Many life problems often result into differential equations models when formulated mathematically, particularly problems that depends on time and rates which give rise to Partial Differential Equations (PDE).
Aims: In this paper, we advance the solution of some Parabolic Partial Differential Equations (PDE) using a block backward differentiation formula implemented in block matrix form without predictors.
Materials and Methods: The block backward differentiation formula is developed using the collocation method such that multiple time steps are evaluated simultaneously.
Results: A five-point block backward differentiation formula is developed. The stability analysis of the methods reveals that the method is stable.
Conclusion: The implementation on some parabolic PDEs shows that the method yields better accuracy than the celebrated Crank-Nicholson’s method.
Aims: In this paper, we advance the solution of some Parabolic Partial Differential Equations (PDE) using a block backward differentiation formula implemented in block matrix form without predictors.
Materials and Methods: The block backward differentiation formula is developed using the collocation method such that multiple time steps are evaluated simultaneously.
Results: A five-point block backward differentiation formula is developed. The stability analysis of the methods reveals that the method is stable.
Conclusion: The implementation on some parabolic PDEs shows that the method yields better accuracy than the celebrated Crank-Nicholson’s method.
Keywords
Collocation
Backward Differentiation Formula
Stability
and Crank-Nicholson