Lectures
- Mon 29 Oct
- Introduction and course outline.
-
Transition matrix and initial distribution of a Markov chain.
(LPW Sec 1.1, Numerical example: Frog.R)
- Tue 30 Oct
- Left-multiplication and right-multiplication by the transition matrix.
(LPW Rem 1.3)
- Random walks on graphs.
(LPW Sec 1.4, 1.5.1)
-
Random walk on the hypercube: The Hamming weight and Ehrenfest's urn.
(LPW Sec 2.3)
- Mon 5 Nov
-
Reachability of states and hitting times.
(LPW Sec 1.5.2)
- Random mapping presentation.
(LPW Sec. 1.2)
- Simulation of the initial distribution. Random number generation.
(LPW App. B.3; RANDOM.ORG)
- Tue 6 Nov (Computer lab at Room MaD353)
- Mon 12 Nov
- Existence of a stationary distribution.
(LPW Sec 1.5.1, 1.5.3; Warning: Section 1.5 contains some errors).
- Uniqueness of the stationary distribution.
(LPW Sec 1.5.4, LPW Exercise 1.13)
- Tue 13 Nov
-
Reversible Markov chains.
(LPW Sec. 1.6)
- Birth-and-death processes.
(LPW Sec. 2.5)
- Mon 19 Nov
- Total variation distance.
(LPW Sec 4.1)
- Coupling of random variables.
(LPW Sec 4.2)
- Tue 20 Nov
-
Aperiodic Markov chains.
(LPW Sec 1.3)
- Markov chain convergence theorem.
(LPW Sec 4.3)
- Mon 26 Nov
- Graph coloring and counting problems.
(LPW Sec 3.1)
- Markov chain Monte Carlo: Metropolis algorithm.
(LPW 3.2)
- Tue 27 Nov
- Mixing times and coalescing chains.
(LPW Sec 4.4–4.5)
- Glauber dynamics and Gibbs samplers.
(LPW Sec 3.3)
- Decentralized graph coloring.
K. Duffy, C. Bordenave, D. Leith : Decentralized constraint satisfaction
- Mon 3 Dec (No lecture)
- Tue 4 Dec (Full afternoon from 14:15–17:30, Room MaD302)