1 / 19

Introduction to Discrete-Time Markov Chain

Introduction to Discrete-Time Markov Chain. Motivation. many dependent systems, e.g., inventory across periods state of a machine customers unserved in a distribution system. excellent. good. fair. bad. time. Motivation. any nice limiting results for dependent X n ’s?

kimberly
Download Presentation

Introduction to Discrete-Time Markov Chain

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Introduction to Discrete-Time Markov Chain 1

  2. Motivation • many dependent systems, e.g., • inventory across periods • state of a machine • customers unserved in a distribution system excellent good fair bad time 2

  3. Motivation • any nice limiting results for dependent Xn’s? • no such result for general dependent Xn’s • nice results when Xn’s form a discrete-timeMarkov Chain 3

  4. Discrete-Time, Discrete-State Stochastic Process • a stochastic process: a sequence of indexed random variables, e.g., {Xn}, {X(t)} • a discrete-time stochastic process: {Xn} • a discrete-state stochastic process, e.g., • state {excellent, good, fair, bad} • set of states  {e, g, f, b}  {1, 2, 3, 4}  {0, 1, 2, 3} • state to describe weather {windy, rainy, cloudy, sunny} 4

  5. Markov Property • a discrete-time, discrete-state stochastic process possesses the Markov property if • P{Xn+1 = j|Xn= i, Xn−1 = in−1, . . . , X1 = i1, X0 = i0} = pij, for alli0, i1, …, in1, in, i, j, n  0 • time frame: presence n, future n+1, past {i0, i1, …, in1} • meaning of the statement: given presence, the past and the future are conditionally independent • the past and the future are certainly dependent 5

  6. One-Step Transition Probability Matrix • pij 0, i, j  0, 6

  7. Example 4-1 Forecasting the Weather • state {rain, not rain} • dynamics of the system • rains today  rains tomorrow w.p.  • does not rain today  rains tomorrow w.p.  • weather of the system across the days, {Xn} 7

  8. Example 4-3 The Mood of a Person • mood  {cheerful (C),so-so (S), or glum (G)} • cheerful today  C, S, or G tomorrow w.p. 0.5, 0.4, 0.1 • so-so today  C, S,or G tomorrow w.p. 0.3, 0.4, 0.3 • glum today  C, S, or G tomorrow w.p. 0.2, 0.3, 0.5 • Xn: mood on the nth day, such that mood  {C, S, G} • {Xn}: a 3-state Markov chain (state 0 = C, state 1 = S, state 2 = G) 8

  9. Example 4.5A Random Walk Model • a discrete-time Markov chain of  number of states {…, -2, -1, 0, 1, 2, …} • random walk: for 0 < p < 1, • pi,i+1 = p = 1 − pi,i−1, i = 0, 1, . . . 9

  10. Example 4.6A Gambling Model • each play of a game a gambler gaining $1 w.p. p, and losing $1 o.w. • end of the game: a gambler either broken or accumulating $N • transition probabilities: • pi,i+1= p = 1 − pi,i−1, i = 1, 2, . . . , N − 1; p00= pNN= 1 • example for N = 4 • state: Xn, the gambler’s fortune after the n play  {0, 1, 2, 3, 4} 10

  11. Limiting Behavior of Irreducible Chains 11

  12. Limiting Behavior of a Positive Irreducible Chain • cost of a visit • state 1 = $5 • state 2 = $8 • what is the long-run cost of the above DTMC? 0.1 0.9 0.2 0.8 2 1 12

  13. Limiting Behavior of a Positive Irreducible Chain • j = fraction of time at state j • N: a very large positive integer • # of periods at state j  j N • balance of flow • j N  i (i N)pij  j= i ipij 13

  14. Limiting Behavior of a Positive Irreducible Chain • j = fraction of time at state j • j= i ipij • 1= 0.91 + 0.22 • 2= 0.11 + 0.82 • linearly dependent • normalization equation: 1+ 2= 1 • solving: 1= 2/3, 2= 1/3 C 0.1 0.9 2 1 0.2 0.8 14

  15. Limiting Behavior of a Positive Irreducible Chain • 1 = 0.752 + 0.013 • 3 = 0.252 • 1 + 2 + 3 = 1 • 1 = 301/801, 2 = 400/801, 3 = 100/801 1 0.25 3 1 2 0.75 0.99 0.01 15

  16. Limiting Behavior of a Positive Irreducible Chain • an irreducible DTMC {Xn} is positive  there exists a unique nonnegative solution to • j: stationary (steady-state) distribution of {Xn} 16

  17. Limiting Behavior of a Positive Irreducible Chain • j = fraction of time at state j • j = fraction of expected time at state j • average cost • cj for each visit at state j • random i.i.d. Cj for each visit at state j • for aperiodic chain: 17

  18. Limiting Behavior of a Positive Irreducible Chain • 1 = 301/801, 2 = 400/801, 3 = 100/801 • profit per state: c1 = 4, c2 = 8, c3 = -2 • average profit 1 0.25 3 1 2 0.75 0.99 0.01 18

  19. Limiting Behavior of a Positive Irreducible Chain • 1 = 301/801, 2 = 400/801, 3 = 100/801 • C1 ~ unif[0, 8], C2 ~ Geo(1/8), C3 = -4 w.p. 0.5; and = 0 w.p. 0.5 • E(C1) = 4, E(C2) = 8, E(C3) = -2 • average profit 1 0.25 3 1 2 0.75 0.99 0.01 19

More Related