![[The University of Leicester]](http://www.le.ac.uk/corporateid/departmentresource/000066/unilogo.gif) | Department of Mathematics |
 |
Next: MA3061 Ordinary Differential Equations
Up: ModuleGuide03-04
Previous: MA3001 Relativity and Electromagnetism
MA3011 Applied Numerical Mathematics
Credits: 20 |
Convenor: Professor B. Leimkuhler |
Semester: 1 |
Prerequisites: |
essential: MA1001(=MC126), MA1002(=MC127), MA1151(=MC146),
MA1152(=MC147), MA2001(=MC224), MA2101(=MC248), MA2102(=MC241),
MA2071(=MC228) |
desirable: |
Assessment: |
Regular coursework and computer assignments, computer
project (written report, oral presentation): 50% |
1.5 hour exam: 50% |
Lectures: |
18 |
Problem Classes: |
none |
Tutorials: |
none |
Private Study: |
104 |
Labs: |
18 |
Seminars: |
none |
Project: |
none |
Other: |
none |
Surgeries: |
none |
Total: |
150 |
Subject Knowledge
Aims
- -
- To be able to identify certain types of elliptic, parabolic
and hyperbolic partial differential equation;
- -
- To be able to classify various types of ordinary differential
equations such as stiff systems and Hamiltonian systems;
- -
- To know physical contexts in which these differential equations
may arise and the types of boundary conditions needed for
well-posedness.
- -
- To be able to construct finite difference schemes for
differential equations, and, where appropriate, analyze their
stability;
- -
- To know and understand appropriate linear algebra methods
needed in the implementation of these schemes;
- -
- To be able to implement the schemes in Matlab.
Subject Skills
Aims
To equip the student with some of the advanced numerical methods and
programming
skills required in engineering and research laboratories.
Learning Outcomes
Programming skills; the ability to solve differential equations
and large linear algebra problems numerically. The course will include
the
design of a substantial computer project, a written project report,
and the oral presentation of such a project.
Assessment
The coursework will consist of regularly assigned exercise sheets,
including computer assignments. There will be weekly instructor-assisted
lab sessions. A substantial individual computer project will be
required and the student will be required to present the project orally
to the class and the instructor.
The June examination will have 4 questions, it being (theoretically)
possible to obtain full marks by answering 3 of them.
Explanation of Pre-requisites
Students must be familiar with elementary material on ordinary
differential
equations (MA1002) and multivariate calculus (MA1001, MA2001). A basic
working
knowledge of linear algebra, and an introductory course in scientific
computing (MA2071) will be required, as will a working knowledge of
classical real
analysis (MA2101). Students must be able to program in MATLAB (MA2071).
Course Description
This course will introduce students numerical approaches to scientific
problems, including applications in engineering and physics. Various
nonlinear differential equations will be introduced as simplified models
for real-world systems, discretizations will be introduced, numerical
methods will be developed for each application, and computer
implementation (primarily in Matlab) will be considered. The idea is
to help students learn how to tackle an application using numerical
methods and to relate the computational results obtained to the model
setting.
Examples of problems to be studied include elliptic systems governing
the load on an elastic membrane, the parabolic heat equation for
temperature distribution in a bar, and nonlinear hyperbolic wave
equations. Methods to be studied include iterative methods such as
conjugate gradients for solving elliptic differential equations, stiff
ordinary differential equation solvers for solving the heat equation,
and symplectic discretizations for conservative models.
The emphasis will be on practical aspects of each model and
implementation of numerical methods, but some theory will be also be
developed.
Syllabus
The course will introduce (a) boundary value problems for elliptic
partial
differential equations in two dimensions, (b) initial/boundary value
problems for parabolic partial differential equations in one and two
dimensions, (c) hyperbolic nonlinear partial differential equations,
especially
the nonlinear wave equation.
It will introduce finite difference schemes appropriate to these
problems, including standard centred finite differences
for elliptic equations, Crank-Nicholson and related methods. Stability
analysis by evolution
of Fouries modes will be presented, together with the
Courant-Friedrichs-Lewy
(CFL) condition for stability.
Finite difference schemes replace partial differential equations by
large
systems of linear equations. Numerical linear algebra methods to be
presented will include SOR (for simple problems) and Krylov subspace methods.
Ordinary differential equation solvers including elementary schemes such
as Trapezoidal Rule, Euler's Method, and explicit Runge-Kutta methods
will also be presented and analysed. A comparison of stiff and
non-stiff ODEs and appropriate numerical methods will be given. The
Leapfrog integrator will be introduced for solving conservative
semi-discretized nonlinear wave equations.
A major component of the course will be the development of a substantial
computer project to tackle an application. Students will present this
work both in written form and in a 20-minute oral presentation.
Reading list
Recommended:
D.F. Mayers and K.W. Morton,
Numerical solution of partial differential equations: an
introduction,
Cambridge University Press (1994).
Recommended:
G. Golub and J. Ortega,
Scientific Computing and Differential Equations : An
Introduction to Numerical Methods,
Academic Press (1991).
Resources
Marked problem sheets, computer assignments.
Module Evaluation
Module questionnaires, module review, year review.
Next: MA3061 Ordinary Differential Equations
Up: ModuleGuide03-04
Previous: MA3001 Relativity and Electromagnetism
Author: C. D. Coman, tel: +44 (0)116 252 3902
Last updated: 2004-02-21
MCS Web Maintainer
This document has been approved by the Head of Department.
© University of Leicester.