Research Fellow in Optimisation and Control

Department of Electrical and Electronic Engineering
School of Electrical, Mechanical and Infrastructure Engineering
University of Melbourne

Salary: PhD entry Level A.6 $87,415 -$93,830 p.a.) plus 9.5% superannuation

The research fellow will join a team of academic staff and postgraduate students working on problems pertaining to real-time decision making in dynamic systems. The team maintains a longstanding partnership with Defence Science and Technology Group in this area.

In this research, the investigation will focus on the coordination of a group of autonomous vehicles to achieve a common objective under uncertain and time varying operational considerations. The flexibility to be able to handle different scenarios and associated mission constraints will be an integral aspect of the research.

The aim of the research is to explore and develop novel algorithms that are implementable in real-time, can predict over a long horizon and make optimal decisions under given performance metrics while satisfying hard constraints. This will involve a combination of predictive control and machine learning (data-driven) techniques to optimally respond to rapidly changing conditions, and to disrupt the planning algorithms of adversarial actions.

The research fellow will have an outstanding background in Engineering or Applied Mathematics (or equivalent), and experience with the implementation of numerical methods and engineering applications of optimisation techniques in real-time control of dynamical systems with exposure to mathematical foundations of learning, graph theory, and combinatorial optimisation.

The research fellow will be located in the Department of Electrical and Electronic Engineering within the Melbourne School of Engineering, and collaborate with researchers and engineers internally and externally. In addition to preparing technical reports, research publications, and computer simulations, the research fellow may also have the opportunity to undertake teaching and student supervision is areas directly related to their research.

Candidates are encouraged to apply via this link.

Application Closing Date: 31 December 2017.
Start Date: As soon as possible from March 2018.

Open PhD Positions

Applications are invited for two PhD scholarships in the areas of (1) real-time decision making for autonomous systems in uncertain environments, (2) secure networked control systems.

The work will be based within the Control and Signal Processing (CSP) Lab, MIDAS Lab, the Department of Electrical and Electronic Engineering, the University of Melbourne. The student will be supervised by Dr Iman Shames (see for more information) and co-supervised by another MIDAS Lab faculty member.

The ideal candidate should have a degree in electrical and electronic engineering, computer science, applied mathematics, or mechanical engineering with solid background in applied mathematics and particularly in numerical methods, control theory, and/or optimisation.

The PhD scholarship provides 30K AUD a year tax-free stipend for up-to four years (subject to passing the PhD candidature confirmation after 12 months) as well as tuition fees for the duration of studies. Additionally, up to 15K AUD will be made available for travel funding to visit other research institutes and attending international conferences.

MIDAS is a large and complementary group of academic researchers and postgraduates collaborating across multiple departments within the Melbourne School of Engineering in areas including automation, control, analytics, machine learning, and optimisation.

Expressions of interest are invited from candidates with (or who expect to gain) a first-class honours degree or an equivalent degree in engineering, computer science, physics, mathematics, or a related discipline. Interested candidates should contact Dr. Shames ( directly and include (1) a brief statement of interest in this position with clear indication of preference for the 2 areas mentioned above (maximum 2 pages), (2) their detailed CV, (3) bachelors and masters/honours transcripts, and (4) names and contact details of at least two referees.

Application Closing Date: 1 December 2017.
Start Date: As soon as possible from March 2018.


What questions to ask in a journal club?

I have looked around and thought a bit about what type of questions one (especially students) ask in journal clubs and I came up with the following list. It is not the most complete list and I am happy to hear suggestions about other things to have in mind. Also, it is prepared for students in the areas where engineering and applied mathematics meet, but can be applicable to other areas as well.

  • What is the most important result of the discussed paper?
  • Is the result incremental with well-understood foundations in the area? Is it new to the area but well-understood in another field of engineering/applied maths? Is it world shatteringly new? Or a waste of time?
  • Was the paper clearly written? How was the flow of arguments? Were the  variables defined properly? How was it structured?
  • What is the most interesting aspect of the paper? It does not need to be the same as the most important bit above.
  • What is the most fundamental mathematical concept mentioned and used?
  • What can you say about computational aspects of the paper? For example, how does the proposed method scale? Or how does it perform in real time?
  • How useful is the result? How realistic are the assumptions? How can you relax the assumptions?
  • What would you do differently? How would you improve the paper?

The common-sense scientific method, where are you?

You can go on and on about what the scientific method entails. If you are after some mild controversy you can just check out the wikipedia talk page. Anyhow, I do not want to discuss its philosophy here, what I really want to briefly rant about is the fact that it in its most common sense form is being forgotten, even in some of the “leading research institution” (what the hell a leading research institution is can be the subject of another lengthy discussion, but I digress). Let’s first clarify what is meant when I use the term “common-sense scientific method”. It is the process of making an educated guess based on some prior knowledge about a system or phenomenon, computing the consequences and implications of the aforementioned guess, checking the veracity of the guess through observations and experiment, and then repeating the process if the guess turns out to be rubbish (Check out this legendary Feynman’s lecture on it if you have time). 

Now, I do not know what happened, and when it happened for that matter. I have some culprits in mind for the rise of this apparent disregard for the common-sense scientific method, particularly the hacker/agile bull-shit entrepreneur culture, but I haven’t been able to really pin it on them. Is their world view the reason for this disregard or they are its product. We never will know the answer. I even tried to blame the baby boomers for it, sure they are definitely at fault; at some points they just decided to maximise their own pay-offs with no respect for facts. But I still really cannot solely blame them. They just took advantage of a mood in the society where everyone stopped appreciating doing anything systematically based on hard facts. I don’t know what can be done except for begging my students (who are going to be engineers soon) to maybe, just maybe, consider doing things in a bit old-fashioned way.