Numerikus módszerek fizikusoknak (BMETE92MF00/T0 - 2019/20/1)

Kurzus típus: 
Elmélet
Nyelv: 
angol
Félév: 
2019/20/1
Órarendi információ: 

Sz 12:15-14:00 (H406)
Cs 12:15-14:00 (H406)

Course requirements

News 

  •  New!   I would like to ask students who are going to write one of the tests again to indicate this intent using this webform. The retake tests will be on 19 December (Thursday), 10:00, room H406. Those who are going to retake both tests first write test II from 10:00 and then test I from 14:30 (H27) in the afternoon. Students may view their second test at 9:30-10-30, 17 December (Tuesday), H24b. 
  • BME has a TAH Matlab licence. Thus, Matlab is available for all students and teachers. The installation guide can be found at this inner link (in Hungarian) or at the beginning of the Matlab news page (in English).
  • Matlab presentation for students about the TAH licence. 19th of September, Thursday 16:00-18:00 - ABC of the MATLAB Campus Wide License, room: K195. Registration at regisztracio.bme.hu. The poster of the presentation is available at the Matlab news page.
  • MSc thesis topics (in Hungarian)
  • Webpage of the Miklós Farkas Seminar on Applied Analysis - Thursdays from 10:15, the talks are in English on regular basis

Main material for the course

Auxilary material for the course

 

The schedule of the lectures and computer labs:

Week Lecture (We12, Th12) Computer lab (We14 and Th14)
1. 09/09 

Wednesday: The requirements of the course. The topics of the course. Model construction and its necessity. Properly posed problems. Conditioning of a problem and of a computation. Error sources of a model. Thursday:  University Sports Day - the lecture is cancelled

We: Special matrices. Vector norms. (Pr. 1-6, 8)


Th: Due to the University Sports Day, the computer lab on Thursday is cancelled.

2. 16/09

Wednesday: Vector and matrix norms, spectral radius, relation between norms and eigenvalues, convergence speed of sequences. Thursday: Floating point numbers and their properties Conrad Zuse - Computer history (video)

We: Matrix norms, norms and eigenvalues, order of convergence, condition numbers, conditioning. (Pr.: 10,11-13,15-17,19-20(a-b) - we skipped Pr. 7,9,14,18 ; these are left to practice at home).
Homework for week 2 (Wednesday


Th: Special matrices. Vector and matrix norms. (Pr.: 1-10 - we skipped Pr. 9, this is left to practice at home) 

3. 23/09

Wednesday:  Conditioning of SLAEs, condition numbers of matrices, Gaussian method and its investigation. Thursday: LU decomposition. Performance of the Gaussian method. Pivoting.

We: Conditioning, foating point numbers, conditioning of linear systems. (Pr.: 20(c-d)-22,24--27,31-33 - we skipped Pr. 23 (discussed in the lecture) and Pr. 28-30 (discussed later))
Homework for week 3 (Wednesday)


Th: Matrix norms, norms and eigenvalues, order of convergence. (Pr.: 11-13, 15-17, 19 - we skipped Pr. 14,18; these are left to practice at home)
Homework for week 3 (Thursday)

4. 30/09

Wednesday: Dean's day - there are no lessons at the Faculty of Natural Sciences. The lecture is cancelled. Thursday: General LU decomposition. $LDM^T$ decomposition, Cholesky decomposition. The left division command of Matlab. Iteration methods for SLAEs. Necessary and sufficient condition for the convergence. Error estimation with the Banach fixed point theorem. 

We:  Due to the Dean's Day, the computer lab on Wednesday is cancelled.


Th: Condition numbers, conditioning. Floating point numbers. (Pr.: 20-28.)
Homework for week 4 (Thursday)

 

5. 07/10 Wednesday: Classical iterative methods (Jacobi, Gauss-Seidel and their relaxed versions) and their convergence. Thursday: Gradient and conjugate gradient methods.

We.: Direct methods for linear systems. (Pr.: 28-30, 34-39)
Homework for week 5 (Wednesday)


Th.: Conditioning of linear systems.Gaussian method, LU-decomposition, partial pivoting. (Pr.: 31-36, we skipped Pr. 29 (a similar problem will be solved later) and 30 (easy))
Homework for week 5 (Thursday)

6. 14/10 Wednesday: Householder reflection, QR decomposition. Thursday: Givens rotation and QR decomposition with Givens rotations. Solution of over-determined systems. Conditioning of eigenvalue problems.

We.: Iterative methods for linear systems. (Pr.: 40-42, Pr. 43-44 is for practice at home, we skip Pr. 45)
Homework for week 6 (Wednesday)


Th.: Cholesky decomposition. Classical iterative methods. (Pr.: 37-41)
Homework for week 6 (Thursday)
 

 

7. 21/10 Wednesday: National holiday, 23 October. The lecture is cancelled. Thursday: Eigenvalue problems (power method, inverse iteration, Rayleigh coefficient iteration, rank deflation).

We.: National holiday, 23 October. The computer lab is cancelled. 


Th.: Gradient and conjugate gradien method. Householder reflection, Givens rotation, QR decomposition. (Pr.: 42-45, 48, 46(only Householder))
No homework this week

8. 28/10 Wednesday: Eigenvalue problems (QR iteration, reduction to Hessenberg form, shifting). Thursday:  Test I in the time slot of the lecture but in different location: room R108. The topic is from the beginning of the semester to the conjugate gradient method. See the previous tests below.

We.: QR decomposition with Householder reflections and Givens rotations, over-determined linear systems (Pr:: 46-51)
Homework for week 8 


Th.: QR decomposition with Householder reflections and Givens rotations, over-determined linear systems (Pr.: 46-47, 49-52)
Homework for week 8 

9
4/11
Wednesday: Solution of nonlinear equations: Newton's method, fixed point iterations. Thursday: Solution of nonlinear systems, unconstrained numerical optimization. Interpolation with polynomials. Lagrange interpolation.

We.: Solution of eigenvalue problems (Pr.: 52-56 (as homework))
Homework for week 9


Th.: Solution of eigenvalue problems (Pr.: 53-56 (as homework))
Homework for week 9

10. 11/11 Wednesday:  Interpolation error. Interpolation on Chebyshev nodes. Thursday: Newton interpolation. Hermite and spline interpolation.

We.: Solution of nonlinear equations (Pr.: 57-58, 60-62 (Pr. 59 is similar to Pr. 57, Pr 62 is homework))
Homework for week 10


Th.: Solution of nonlinear equations (Pr.: 57-58, 60-62 (Pr. 59 is similar to Pr. 57, Pr 62 is homework))
Homework for week 10

11. 18/11 Wednesday: Trigonometric interpolation. Thursday: Fast Fourier transform. We.: Polynomial and spline interpolation (Pr.: 63-68.)
Homework for week 11

Th.: Polynomial and spline interpolation (Pr.: 63-68.)
Homework for week 11

12. 25/11 Wednesday: Numerical differentiation. Numerical integration with Newton-Cotes formulas. Thursday: Motivation of Gaussian quadrature. We.: Trigonometric interpolation (pianosound.mat), numerical differentiation. (Pr. 69-71).
Homework for week 12 (last homework)
 

Th.: Trigonometric interpolation (pianosound.mat), numerical differentiation (Pr. 69-71).
Homework for week 12 (last homework) 

13. 2/12 Wednesday: Gaussian quadrature. Introduction to the numerical solutions of ordinary differential equations. Thursday: Runge-Kutta methods. Absolute stability, stiffness. We.: Numerical integration (Pr. 74-77)

Th.: Numerical integration (Pr. 74-77)

14. 9/12 Wednesday: Multistep methods. Boundary value problems. Thursday: Test II in the time slot of the lecture but in different location: room R108. The topic is from Householder reflections to numerical integration. See the previous tests below. We.: Solution of ordinary differential equations (Pr. 78-84)

Th.: Solution of ordinary differential equations (Pr. 78-84)

Previous midterm tests with solutions (mostly in Hungarian) 

Useful links