RE: University of South Africa students' concerns regarding the APM3711 (Numerical Methods II) Oct/Nov 2024 Final Examination

Date and Session: Thursday, 7th November 2024, 08:00-11:00 SAST

We are writing to address issues regarding APM3711 exam that was held this morning. We were informed that the exam content would closely resemble our assignments, giving us a fair basis for preparation. However, the actual exam content was markedly different, covering topics neither aligned with the assignments nor clearly outlined in the module outcomes. The module outcome for 2024 Tutorial letter states 
3.2 Outcomes By the end of this module, students should 2.2.1 be competent in using Taylor’s method of order 2 and higher to approximate the solution of an initial-value problem. 2.2.2 be able to use Euler and modified Euler methods on initial-value problems. 2.2.3 demonstrate fluency in Runge-Kutta methods and their error control. 2.2.4 must know how multistep methods work and their relative merits. 2.2.5 when given a function, they should be able to find a Pade approximation. 2.2.6 be able to use the Chebyshev polynomials to find the Chebyshev series and estimate the maximum error of the Chebyshev series over the interval [−1,1]. 2.2.7 be able to economise the given power series.
.2.8 be able to apply Gerschgorin’s circle theorems in finding eigenvalues and hence the corresponding eigenvectors. 2.2.9 be competent in calculating the dominant eigenvalue and the smallest absolute value using the power method. 2.2.10 be able to use the shooting method to solve a boundary-value problem both theoretically and numerically. 2.2.11 be competent in solving characteristic-value problem using finite difference method. 2.2.12 be able to model a steady-state heat by Laplace’s equation and approximate it by the 5-point difference formula and hence obtain the solution numerically. 2.2.13 be able to apply the iteration formula for S.O.R. on Laplace equation and hence find numerical solution. 2.2.14 haveathoroughgraspofthealternating-direction-implicit method (ADI) for solving Laplace/Poisson equation and do this numerically.

The exam 2024 contained: 
-Linear finite difference method
-Nonlinear finite difference method
-Possion Equation finite difference algorithm 
-Superimposing graphs 

As a result, this disconnect has left many of us feeling misled and frustrated, as we dedicated significant time preparing in the way we were guided to. We believe it is essential for exams to fairly reflect both the syllabus and the guidance provided throughout the course. Testing on material outside the syllabus not only diminishes our efforts but also affects our performance unfairly. Despite all this, many students emailed the lecturer regarding the formatting of exams and what exactly to expect, which in return the lecturer either blankly ignored or responded very  vaguely giving us no form of assistance or guidance as to what to expect in the exam.

We ask for your understanding and consideration in addressing this matter. We hope a solution can be found to ensure fair assessment for all students in this course.

Thank you for your attention to this important issue.