Regarding MSc and PhD research areas

[lonatico]

Hi,
I'm about to start my MSc in Computer Science studying "Combinatorial Optimization", which was already the theme of my undergraduation thesis. I like the field and the research, but I'd rather prefer working with numerical methods and techniques. The problem is my current university has a very weak research on that, and it is located in the maths departament.

My two questions are:
1) Is it possible for a CS student to work with numerical methods in MSc and PhD level?

2) If I complete my MSc in CS (optimization), is there a chance of me being able to apply for a PhD (on other universities) to study numerical methods (usually located under maths)? I have heard some people talking about the importance of "breeding" in academic research;

I know it usually depends on the university, but the questions are more on a "is it likely" fashion.

 

[micromass]

Hi,
I'm about to start my MSc in Computer Science studying "Combinatorial Optimization", which was already the theme of my undergraduation thesis. I like the field and the research, but I'd rather prefer working with numerical methods and techniques. The problem is my current university has a very weak research on that, and it is located in the maths departament.

My two questions are:
1) Is it possible for a CS student to work with numerical methods in MSc and PhD level?

Yes, but depending on the topic and your knowledge, you might have some catching up to do.

2) If I complete my MSc in CS (optimization), is there a chance of me being able to apply for a PhD (on other universities) to study numerical methods (usually located under maths)? I have heard some people talking about the importance of "breeding" in academic research;

Yes, if you get good grades and good recommendations from professors, there will be no problem.

 

[lonatico]

Thanks for the reply :)

So, my knowledge is not that wide. I know the basics about numerical methods: integration (Trapezium and simpson), interpolation (Lagrange, nevile, bezier and splines), zeros of functions (bissection and Newton), ODE (runge kutta). However, I don't have a deeper knowledge on math fundation such as differential Equations (all I know is because of the RK method and of image processing, i.e., only a very superficial knowledge), or numerical analysis, which as far as I know is important for the area. Though the classes, implementing and reading about such field is very entertaining to me. I would like to work with more math than CS as to say.

Regarding my grades, throughout my CS degree I have been a B+~A average student. About the recomendation letter, does it have to be from people working in the numerical methods field, or can it be, say, from my current advisor (combinatorial optimization)?

I'm very aware that my knowledge has lots of gaps, so being realistic, would it be hard to get accepted in a MSc or Phd? And would it be a better choice to perhaps drop my current MSc and try another program on numerical methods instead (risking being half or one year delayed)?

 

[micromass]

Thanks for the reply :)

So, my knowledge is not that wide. I know the basics about numerical methods: integration (Trapezium and simpson), interpolation (Lagrange, nevile, bezier and splines), zeros of functions (bissection and Newton), ODE (runge kutta). However, I don't have a deeper knowledge on math fundation such as differential Equations (all I know is because of the RK method and of image processing, i.e., only a very superficial knowledge), or numerical analysis, which as far as I know is important for the area. Though the classes, implementing and reading about such field is very entertaining to me. I would like to work with more math than CS as to say.

You have some catching up to do. But nothing too much. I would start reading up on numerical methods as soonas you can.

Regarding my grades, throughout my CS degree I have been a B+~A average student. About the recomendation letter, does it have to be from people working in the numerical methods field, or can it be, say, from my current advisor (combinatorial optimization)?

Your current advisor is ok as a recommendation. But you will also want a letter from somebody in numerical methods for your PhD.

I'm very aware that my knowledge has lots of gaps, so being realistic, would it be hard to get accepted in a MSc or Phd? And would it be a better choice to perhaps drop my current MSc and try another program on numerical methods instead (risking being half or one year delayed)?

I think you have a chance. But to drop out of your Msc for that? I wouldn't do it. Well, you know the risks, it's up to you. It's not like you can't get into a PhD in numerical if you do this masters.