MTH3051: Introduction to Computational Mathematics

11 x applied class participation/problem task worth 1% each for a maximum of 8%. The task is to be sent in before the applied class(does not have to be correct) and you have to go and participate in the applied class to be counted.
12 x weekly quizzes worth 2% each(20% in total, only the best 10 are counted)
4 x assignments worth 3% each
Final exam worth 60%

This was both my favourite subject and my most stressful one. The main topic covered were errors, solving linear and nonlinear systems using algorithms, data fitting, and using types of data fitting to help do various mathematical things. Basically, a lot of learning how computers do things that we know how to do manually.
Janosch was ok. He explained things very visually, often using various drawings to illustrate his point. There weren't always a lot of examples though, and it was harder to try to apply the content to actual questions as a result. I found that generally going through the notes helped. The content was surprisingly theoretical but the theory was interesting. There were a lot of proofs involved, but we were only expected to be able to replicate a few of the shorter ones for the exam.
In terms of the assessment, the applied class participation was basically a free 8%, which I was very grateful for. The assignments weren't too hard if you've ever done coding before(and I think from next year you need to be able to anyway). The marking scheme was basically "if the code works you get full marks, otherwise 0 for the question". It was good if you did everything right but annoying if you had one slight mistake in the code. The quizzes were very annoying to me. They were done at the end of the week, before applied classes, which meant you had to try and have a grasp of the content with little opportunity to ask questions or practice questions. I thought having 50% of the insemester mark being based on your understanding of work before you had a chance to look at it properly was a little unfair. What was most annoying was that there was negative marking involved. Every quiz had at least one "tick the box" style question in which you had to tick all the correct answers, but if you ticked an incorrect one, you'd lose marks. For example, if there was 2 correct answers and you ticked one correct and one that was incorrect, you would get a 0, despite having some understanding. Was not a great system. The negative marking also applied to the exam, which sucked but at least by then we were used to it.
Biggest recommendation is to ask a tonne of questions whenever you get space, don't just settle if you don't understand something. And also, make sure you test out your code before submitting.

Janosch Rieger

There was a sample exam that was also our week 12 quiz so was included in our assessment.

3.5 out of 5

Yes with screen capture(all over zoom)

No textbooks are required, but I believe you could buy a paper copy of the notes if you needed to.

Three 1 hour lectures and one 1.5 hour applied class per week

2021 semester 2(all online)

81 HD

First and foremost, READ THE OTHER REVIEW TO GET INSIGHTFUL COMMENTS.

This unit is OKAY at best. The thing that pissed me off was that the assignments weren't related to the actual unit/exam questions. For example, one of the assignments was literally about converting psuedocode into MATLAB or some other programming language. I ended up doing the majority of my code in Python because Python is cool and MATLAB is gross.

Also, Leo (although a super nice guy) is really, really, REALLY slow with the lectures (e.g. he spent like 20 minutes writing down 1 matrix/figuring out how to find and use the eraser for his iPad). He's a super smart guy, and really down to earth, but he can be slow with his explanations at times.. and that's why I didn't bother going to his lectures.

Jennifer Flegg on the other hand is nothing less than awesome. She really knows her stuff and is amazing at explaining tough concepts. I didn't end up going to any of her lectures either, but from what I saw, she was awesome. Also, if you have a 1 on 1 consultation with her, you won't leave feeling like you've wasted your time.

Now, for the exam. The MCQ are the exact same as every other year, but they might have some different numbers. As for the SA, there is some similarities (e.g. define round off error), but the majority is totally different! I was under the assumption that they were the same, but they aren't, so make sure you understand the concepts in the unit!

All in all, the unit is easy. The majority of people walk into the exam room only needing 30% to pass. The majority of people also walk into the exam room with 2 days of study under their belt. As for me, I was in both categories. I went to like 3 lectures out of the 30 odd lectures (and didn't catch up, even when I was studying for the exam).. did all the assignments (and did well because they aren't hard), and then did 2 days of study and still managed to get a decent mark. Do this unit if you cbf with uni but still want to get a good mark.

Yes, one, with solutions

1 out of 5

Yes, with screen capture

The recommended textbook is Getting Started with MATLAB 7 by Rudra Pratap, Oxford Uni Press. Don't bother reading it, I didn't.

2015, Semester 1

Ye not 2 bad given I crammed 12 weeks in 2 days

There are 5 assignments, each worth 6%. Each assignment has one big question and we are meant to solve it using MATLAB. The assignment are not hard and in fact quite fun, the programming involved was really basic and 3 out of the 5 assignments are not related the course material whatsoever. Some of them can be tricky but you can always discuss the assignments with your tutor so even if you had no clue how to start, your tutor would definitely guide you through it. The marking, on the other hand, was quite harsh. Almost everyone gets 100% for the first assignment as long as you get the answer right. (It is hard not to get it right and Leo spends like half a lecture discussing about that first assignment.) But from second one, the marks drop quite a bit for most students, apparently the marking scheme has changed and every steps needed to be made clear in the assignment in order to get a good mark. But then, if effort were put in and everything were explained step by step, getting a good mark was not too hard. Don't forget you can always ask you tutor for help.

First of all, this unit is mostly about numerical analysis and not much computing/programming. That said, MATLAB programming can be picked up through the semester but nothing more than from a first year eng computing unit.

Topics include:

• Introduction to Matlab
• Truncation and Round-off Errors
• Solutions of Non-Linear Equations in One Variable
• Solutions of Linear System of Equations
• Solutions of Non-Linear System of Equations
• Polynomial Interpolation, Cubic Splines, Fourier Series
• Richardson Extrapolation
• Definite Integrals and Romberg Integration
• Numerical Differentiation. Finite Difference and Numerical Errors
• Numerical Solutions of Ordinary Differential Equations
• Optimization Methods, Golden Search and Genetic Algorithm
• Computer Generated Random Numbers

As most (real-life) mathematical problems cannot be solved analytically/exactly, we have to somehow approximate the solutions. In this unit, you will learn how to solve a wide range of problems using various numerical methods, their derivations and error analyses. Taylor series is used A LOT in this unit, in fact, 2/3 of the numerical methods are based on Taylor Series expansion, so make sure you are comfortable with that. All the topics sound pretty easy but there is no other way than using numerical approximations when solving equations like or evaluating . Surely you can use a calculator, but how does your calculator do the job? That's what you will be learning in this unit. Not everyone enjoys this unit. (It really can be told by the attendance of lectures.) The main reason being many people think computational maths is not real maths. (Just like when I first encountered Euler's method in VCE Specialist Maths, I didn't appreciate it much either.) But in this unit, it's not like all these numerical methods are just thrown at you without explanations, in fact, everything is derived from scratch. But then, there are lots of different topics, and the topics are not really interrelated and I didn't really tried make sense of it, as a result, I kind of regretted choosing it and neglected it throughout the semester (I didn't study for it, didn't do a single exercise sheet question, although I did spend a decent amount of time working on the assignments as they are actually enjoyable.) However, it was my first exam and all my other units were 2 weeks after this, so I spent quite a lot of time on it during the exam period. Once I sat down and read and tried and understood everything slowly, everything just made so much sense and it was actually really enjoyable (In comparison to MTH3011, which makes sense as Michael Page makes everything clear but not really 'enjoyable').

Tutorials are not compulsory for this unit. I heard that the way the tutors run the tutes vary a lot from tutor to tutor. Some tutors go through questions from the problem set in the tutes and teach a fair bit, but not mine. I had Jennifer Flegg as my tutor. As aforementioned, she was the lecturer of the second half of the unit and she definitely knows her stuff, and she is super good with MATLAB. In my tutes, instead of going through the problem sets as a class, each week up to around week 9 or 10, we simply just work on the assignments and ask for help. This is a good and bad thing for me. As there are 5 assignments and there is always one due every second week, I'm always working on the assignments and never really spend anytime on the exercise sheet. Starting from week 9 or so, each week we also spend some time talking about the exercise sheet questions and also past exam questions. Once again, although the tutes are not compulsory, go to the tutes and ask questions about the assignments.

The exam is worth 70% and is consisted of 20 MC questions and the rest are long questions. Most questions ask you to derive a certain numerical method, which of course are all covered in lectures so you will be fine if you memorize most the derivation and also error analysis for the 20 or 30 numerical methods. Also, the exam is quite similar to past exam; most MC are exactly the same, and same for the long questions. For a third year mathematics unit, the exam is really easy. If you spent time studying for it, even start cramming from SWOTVAC like me, the exam wouldn't be much of a problem. Apparently some people finished this exam in like 1.5 or 2 hours comfortably.
Overall, a fantastic and actually useful unit. If you are an Applied Mathematics major, you will most likely be doing this unit anyway, but even if you are a Pure Mathematics major (I would imaging Pure Maths major enjoy this unit more than MTH3011) or even not a Mathematics major, definitely consider doing this unit as it is actually quite fun and useful, not overly difficult and you will be picking up a programming language.

Leo took the first 6 weeks of the course. Many students had Leo in first year and I think he was one of the lectures of MTH1035, but I didn't go to a single lecture for MTH1035 so who knows. Anyway, I'm not a big fan of him myself. He definitely is a good mathematician (he teaches Honours General Relativity), but for some reason I just feel lost in his lecture despite the topics are really simple. Jennifer came in in week 7. She happened to be my tutor for this unit. In my opinion she is a much better lecturer than Leo. She is organized, makes everything clear and I actually understanding what she is talking about in lectures without any confusion.

Yes, one, with solutions

4.5 out of 5

Yes, with screen capture

The recommended textbook is Getting Started with MATLAB 7 by Rudra Pratap, Oxford Uni Press. As MABLAB is only covered in the first 1 or 2 lectures and the rest of the course is about numerical analysis, plus the assignments require minimum programming skills, the textbook is of little use.

2014, Semester 1

TBA