Dear all,
I hope you liked the mini-project. The course competition will be held on Thursday, December 19. At the day of competition, you will receive a new .p function similar to the ones that you already worked with for the mini-project - e.g. chain_1.p. By executing this function, you are able to sample from a "time-homogeneous Markov chain". Consider this chain as the base chain for the Metropolis-Hasting algorithm for sampling from an arbitrary distribution pi_a. You should find the pi_a for which you will have "the fastest convergence". So, all you should submit for the competition is a 1*5 vector. The specific meaning of "the fastest convergence" and the way scores will be attributed is going to be announced during the competition. You should be prepared for any scenario!
Please note that:
1. The competition score is just a bonus: the team with the highest score will get automatically the maximal grade for the project.
2. You are allowed to use any method and algorithm that you prefer (within the framework described above, of course).
Best wishes and good luck,
Alireza
