By Montenegro R.
Read or Download Mathematical Aspects of Mixing Times in Markov Chains PDF
Best mathematics books
Loads of fiscal difficulties can formulated as limited optimizations and equilibration in their recommendations. numerous mathematical theories were delivering economists with essential machineries for those difficulties coming up in monetary conception. Conversely, mathematicians were motivated via a number of mathematical problems raised by way of fiscal theories.
Optimization is a wealthy and thriving mathematical self-discipline, and the underlying concept of present computational optimization concepts grows ever extra refined. This ebook goals to supply a concise, obtainable account of convex research and its purposes and extensions, for a large viewers. each one part concludes with a frequently broad set of non-compulsory routines.
- Topos Theory
- Equations fonctionelles analytiques
- Elements de Mathematique. Espaces vectoriels topologiques. Chapitres 1 a 5
- Uncertainty Quantification and Stochastic Modeling with MATLAB
- Praktische Lebensversicherungsmathematik: Mit zahlreichen Beispielen, Abbildungen und Anwendungen
- Higher Engineering Mathematics (7th Edition)
Additional info for Mathematical Aspects of Mixing Times in Markov Chains
7) δ ∈ This follows because y = λ (y −δ)+(1−λ) (x+δ) with λ = 1− x−y+2δ [0, 1] and so by concavity f (y) ≥ λ f (y − δ) + (1 − λ) f (x + δ). Likewise, x = (1 − λ) (y − δ) + λ (x + δ) and f (x) ≥ (1 − λ) f (y − δ) + λ f (x + δ). 7). 7) shows that if a bigger value (x) is increased by some amount, while a smaller value (y) is decreased by the same amount, then the sum f (x) + f (y) decreases. In our setting, the condit t tion that ∀t ∈ [0, 1] : 0 g(u) du ≥ 0 gˆ(u) du shows that changing from gˆ to g increased the already large values of gˆ(u), while the equality 1 1 ˆ(u) du assures that this is canceled out by an equal 0 g(u) du = 0 g decrease in the already small values.
12. Goel  solves a card-shuffling problem by comparison methods. He considers a slow card shuffle where either the top card in the deck is put in one of the bottom k positions, or one of the bottom k cards is put at the top of the deck. Mixing time upper and lower bounds are shown by comparison to the relevant quantities for the well studied random transposition shuffle, in which a pair of cards is chosen uniformly and the positions of the two cards are then swapped. The following example illustrates how comparison of spectral profile might make it possible to simplify some difficult results.
In particular, it is one of the few methods for studying relative entropy mixing τD ( ) of a discrete time chain. Work will be done in discrete time, but carries over easily to continuous time, as discussed at the end of the chapter. The results and their proofs are based on the work of Morris and Peres  and Montenegro . We also briefly consider Blocking Conductance, an alternate method which shows better mixing results when set expansion is poor on a small set, but high for larger sets.
Mathematical Aspects of Mixing Times in Markov Chains by Montenegro R.