|
Module Availability |
Autumn Semester
|
|
|
Assessment Pattern |
Unit(s) of Assessment
|
Weighting Towards Module Mark( %)
|
A test and a piece of assignment (including some programming work)
|
25%
|
2.5 hour closed-book examination
|
75%
|
Qualifying Condition(s)
A weighted aggregate mark of 50% is required to pass the module.
|
|
|
|
Module Overview |
|
|
|
Prerequisites/Co-requisites |
None |
|
|
Module Aims |
To introduce the concept of structure preserving numerical integration and the example of symplectic integrators for mechanical systems; To explain their good numerical behaviour using backward error analysis. |
|
|
Learning Outcomes |
By the end of the course the students should know how to design and implement simple symplectic Runge-Kutta methods and splitting methods. Moreover they should be able to explain the advantages of symplectic discretizations vs non-symplectic integrators using backward error analysis. |
|
|
Module Content |
Introduction:
Introduce examples of Hamiltonian systems from physics and their conserved quantities; motivate symplectic integrators versus non-symplectic integrators for some simple examples, in particular with respect to numerical conservation of energy.
One-step methods for ODEs:
Introduce consistency and order of convergence for one-step methods; treat some low order Runge-Kutta methods as examples.
Hamiltonian systems and symplectic discretizations:
Define symplectic maps, show that Hamiltonian flows are symplectic. Introduce implicit midpoint rule as symplectic integrator. Introduce simple partitioned symplectic Runge-Kutta methods (symplectic Euler method) and symplectic discretizations constructed by splitting methods; apply to problems from physics.
Backward error analysis for symplectic integrators:
Introduce the concept of a modified equation; derive exponential error estimates and approximate energy conservation for exponentially long times. Recall Noether’s theorem and prove conservation of angular momentum of symplectic methods.
|
|
|
Methods of Teaching/Learning |
Teaching is by lectures and tutorials. Learning takes place through lectures, tutorials, exercises and coursework. 3 hours lectures/tutorials/examples classes per week for 10 weeks. |
|
|
Selected Texts/Journals |
http://www.maths.surrey.ac.uk/personal/st/C.Wulff/Modules/MSM.GIN/msm.gin.html
A:
B. Leimkuhler and S. Reich. Simulating Hamiltonian Dynamics.
Cambridge Monographs on Applied and Computational Mathematics, 14. ISBN – 0521772907, 2004.
B:
E. Hairer and C. Lubich and G. Wanner. Geometric numerical integration. Structure-preserving algorithms for ordinary differential equations. Springer Verlag, Berlin, 2002
C:
J. M. Sanz-Serna and M. P. Calvo. Numerical Hamiltonian problems. Chapman and Hall, London, 1994. |
|
|
Last Updated |
30 July 2007 |
|