Dear all students,
I just wanted to mention one more point in regard to your overall voyage to theoretical Ph.D. courses at KTH. I tend to think my course covers relatively rudimentary theories in a fashion tailored to Ph.D. students who are obliged to strike a balance between coursework and his/her research project. While a substantial part of the course (e.g., Modern Statistics, Palm Calculus) can be almost directly applied to your research with some additional efforts (For example, throughput analysis is well exemplified in the IR-UWB paper and so on in Palm Calculus lecture), more than half of the course material serves as an intermediary theoretical skills.
If you are willing to take another theoretical course, I highly recommend you to take EL3300/SF3849 (which I mentioned in foreword to Palm Calculus) rather than the trendy game theory course. That is because (i) optimization theory is just a basic language (e.g., you can't discuss about programming with people who have no concept at all about Java); (ii) game theory should not be applied to lifeless objects which do not have any existential willingness to compete; (iii) there are so many stylized assumptions in game theory, the adoption of which will alienate your research from reality.
Admittedly, there are some research problems where game theory can be crucially exploited. For example, economic interaction between service providers or telcos can be oftentimes best modelled by game theory. Secondly, one part of game theory, called evolutionary game theory (or population game theory) sometimes leads to very desirable optimization algorithms amendable for distributed implementation when a special type of function called potential function exists, to understand which you are still required to study optimization theory.
- Jeong-woo