Basic Information

  • Project Title: Learning in Resource Pooling
  • Name: Harsh Shah
  • Project: Learning in Resource Pooling
  • Guide: Prof.Jayakrishnan Nair and Prof.Vivek Borkar

Description of project

Find Nash/Wardrop Equilibrium points for a Resource Pooling game using Stochastic Approximation

Worked With:

For the theoretical part i had to refer to lot of research papers and books on Stochastic Approximation

Relevant Courses

EE621, EE734, EE736

How did you connect with prof?

Mailed professor before the summer for BTP

Expectations from Prof:

In the start, most meetings were used in formulating the problem statement. During this time most work was done inside the meeting duration and not much expected outside that but I used to bring new ideas which we could try. After the problem statement was constructed, meetings were used to guide me about what particular methods I could try out to reach the necessary solution. I was expected to try them out and rigorously present the results in the next meeting. Most of this part was theoretical work and I used to perform simulations for my better understanding.

My expectations from professors was to help me wherever I was stuck in the proofs and point me to different literature for understanding new details.

Outcomes of BTP:

The BTP work is still underway and may results in a publication after some more work.

Experience

Yes I did extend to BTP-II. BTP provides you with some extra time for good research which can be used to shape you for higher studies. My main reason was to pursue this interesting problem and complete the work.

Advice:

Theoretical can become tedious at times where you work for hours on some proof which may not work out at the end. I would advice to have some patience and try out different methods on pen and paper because they will teach you a lot as well. In theoretical work we try to prove something and if it works then well and good but if it doesn’t then try to find out why it doesn’t and give that reason to the professor who can guide you how to go over it.

Documentation is a must and try to keep a running latex document for your results. Document your meetings as well so that you don’t forget to do required things.

Simulations are big help to verify if your theoretical claims are true or not. Use this and also show plots in your presentation.

During the meeting, clear all your doubts. Remember that meeting is for you to clear your doubts rather than telling the professor what you have done.

Last advice about reaching out to professors if they forget to reply to your emails - meet them after their lectures and discuss there.