Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 32 views

    18.445 Introduction To Stochastic Processes: Lecture 16: Optional Stopping Theorem

    Uploaded by

    DevendraReddyPoreddy
    This document discusses stopping times and the optional stopping theorem. It defines stopping times and properties like optional stopping. It states that if a process is a martingale, supermartingale, or submartingale, the optional stopping theorem can be used to relate the expected value of the process at a stopping time to the initial expected value.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    18.445 Introduction To Stochastic Processes: Lecture 16: Optional Stopping Theorem

    Uploaded by

    DevendraReddyPoreddy
    32 views13 pages
    This document discusses stopping times and the optional stopping theorem. It defines stopping times and properties like optional stopping. It states that if a process is a martingale, supermartingale, or submartingale, the optional stopping theorem can be used to relate the expected value of the process at a stopping time to the initial expected value.

    Original Description:

    Fyjkkbv

    Original Title

    MIT18_445S15_lecture16

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses stopping times and the optional stopping theorem. It defines stopping times and properties like optional stopping. It states that if a process is a martingale, supermartingale, or submartingale, the optional stopping theorem can be used to relate the expected value of the process at a stopping time to the initial expected value.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    32 views13 pages

    18.445 Introduction To Stochastic Processes: Lecture 16: Optional Stopping Theorem

    Uploaded by

    DevendraReddyPoreddy
    This document discusses stopping times and the optional stopping theorem. It defines stopping times and properties like optional stopping. It states that if a process is a martingale, supermartingale, or submartingale, the optional stopping theorem can be used to relate the expected value of the process at a stopping time to the initial expected value.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 13

    18.

    445 Introduction to Stochastic Processes


    Lecture 16: Optional stopping theorem

    Hao Wu

    MIT

    13 April 2015

    Hao Wu (MIT) 18.445 13 April 2015 1 / 15


    Recall
    (Ω, F, P) a probability space
    A filtration (Fn )n≥0
    X = (Xn )n≥0 is adapted to (Fn )n≥0 and is integrable
    X is a martingale if E[Xn | Fm ] = Xm a.s. for all n ≥ m.
    X is a supermartingale if E[Xn | Fm ] ≤ Xm a.s. for all n ≥ m.
    X is a submartingale if E[Xn | Fm ] ≥ Xm a.s. for all n ≥ m.
    Today’s Goal
    stopping time
    Optional stopping theorem : E[XT ] = E[X0 ] ?

    Hao Wu (MIT) 18.445 13 April 2015 2 / 15


    Examples

    Pn
    Example 1 Let (ξi )i≥1 be i.i.d with E[ξ1 ] = 0. Then Xn = 1 ξi is a
    martingale.
    Example 2 Let (ξi )i≥1 be i.i.d with E[ξ1 ] = 1. Then Xn = Πn1 ξi is a
    martingale.
    Example 3 Consider biased gambler’s ruin : at each step, the gambler
    gains one dollar with probability p and losses one dollar with
    probability (1 − p). Let Xn be the money in purse at time n.
    If p = 1/2, then (Xn ) is a martingale.
    If p < 1/2, then (Xn ) is a supermartingale.
    If p > 1/2, then (Xn ) is a submartingale.

    Hao Wu (MIT) 18.445 13 April 2015 3 / 15


    Basic properties
    About the expectations
    If (Xn )n≥0 is a martingale, then E[Xn ] = E[X0 ] for all n.
    If (Xn )n≥0 is a supermartingale, then E[Xn ] is decreasing in n.
    If (Xn )n≥0 is a submartingale, then E[Xn ] is increasing in n.
    More
    If (Xn )n≥0 is a supermartingale, then (−Xn )n≥0 is a submartingale.
    If X is both supermartingale and submartingale, then it is a
    martingale.
    If X = (Xn )n≥0 is a martingale, then (|Xn |)n≥0 is a non-negative
    submartingale.

    Lemma
    If (Xn )n≥0 is a martingale, and ϕ is a convex function,then (ϕ(Xn ))n≥0
    is a submartingale.
    Hao Wu (MIT) 18.445 13 April 2015 4 / 15
    Examples

    Suppose that (Yn ) is a biased random walk on Z : p = 6 1/2,


    (
    1 with probability p
    Yn+1 − Yn =
    −1 with probability 1 − p.

    Lemma
    Set µ = 2p − 1, and set
    Xn = Yn − µn.
    Then (Xn ) is a martingale.

    Hao Wu (MIT) 18.445 13 April 2015 5 / 15


    Stopping time
    Suppose that (Fn )n≥0 is a filtration.

    Definition
    A stopping time T : Ω → N = 0, 1, 2, ..., ∞ is a random variable such
    that
    [T = n] ∈ Fn , ∀n.

    Lemma Lemma
    The following are equivalent. If S, T , Tj are stopping times. The
    [T = n] ∈ Fn for all n. following are also stopping times.
    [T ≤ n] ∈ Fn for all n. S ∨ T and S ∧ T
    [T > n] ∈ Fn for all n. infj Tj and supj Tj
    [T ≥ n] ∈ Fn−1 for all n. lim infj Tj and lim supj Tj

    Hao Wu (MIT) 18.445 13 April 2015 6 / 15


    Stopping time
    (Ω, F, P) : a probability space with a filtration (Fn )n≥0 .
    X = (Xn )n≥0 : a process adapted to (Fn )n≥0
    Definition
    Let T be a stopping time. Define the σ-algebra FT by

    FT = σ{A ∈ F : A ∩ [T ≤ n] ∈ Fn , ∀n}.

    Intuitively, FT is the information available at time T .


    If T = n0 , then FT = Fn0 .
    XT 1[T <∞] is measurable with respect to FT .
    Let S and T be stopping times, if S ≤ T , then FS ⊂ FT .
    If X = (Xn )n≥0 is a process, define X T = (XnT )n≥0 by XnT = XT ∧n .
    X T is adapted.
    If X is integrable, then X T is also integrable.
    Hao Wu (MIT) 18.445 13 April 2015 7 / 15
    Optional Stopping Theorem
    Goal : E[XT ] = E[X0 ] ?
    Theorem
    Let X = (Xn )n≥0 be a martingale.
    1 If T is a stopping time, then X T is also a martingale.
    In particular, E[XT ∧n ] = E[X0 ].
    2 If S ≤ T are bounded stopping times, then E[XT | FS ] = XS , a.s.
    In particular, E[XT ] = E[XS ].
    3 If there exists an integrable random variable Y such that |Xn | ≤ Y
    for all n, and T is a stopping time which is finite a.s., then
    E[XT ] = E[X0 ].
    4 If X has bounded increments, i.e. ∃M > 0 such that
    |Xn+1 − Xn | ≤ M for all n, and T is a stopping time with E[T ] < ∞,
    then E[XT ] = E[X0 ].

    Hao Wu (MIT) 18.445 13 April 2015 8 / 15


    Gambler’s ruin
    The gambler’s situation can be modeled by a Markov chain on the
    state space {0, 1, ..., N} :
    X0 : initial money in purse
    Xn : the gambler’s fortune at time n
    P[Xn+1 = Xn + 1 | Xn ] = 1/2,
    P[Xn+1 = Xn − 1 | Xn ] = 1/2.
    The states 0 and N are absorbing.
    τ : the time that the gambler stops.

    Theorem
    Assume that X0 = k for some 0 ≤ k ≤ N. Then
    k
    P[Xτ = N] = , E[τ ] = k (N − k ).
    N

    Hao Wu (MIT) 18.445 13 April 2015 9 / 15


    Optional Stopping Theorem

    Theorem
    Let X = (Xn )n≥0 be a martingale.
    1 If T is a stopping time, then X T is also a martingale.
    In particular, E[XT ∧n ] = E[X0 ].
    2 If S ≤ T are bounded stopping times, then E[XT | FS ] = XS , a.s.
    In particular, E[XT ] = E[XS ].
    3 If there exists an integrable random variable Y such that |Xn | ≤ Y
    for all n, and T is a stopping time which is finite a.s., then
    E[XT ] = E[X0 ].
    4 If X has bounded increments, i.e. ∃M > 0 such that
    |Xn+1 − Xn | ≤ M for all n, and T is a stopping time with E[T ] < ∞,
    then E[XT ] = E[X0 ].

    Hao Wu (MIT) 18.445 13 April 2015 10 / 15


    Optional Stopping Theorem
    Theorem
    Let X = (Xn )n≥0 be a supermartingale.
    1 If T is a stopping time, then X T is also a supermartingale.
    In particular, E[XT ∧n ]≤E[X0 ].
    2 If S ≤ T are bounded stopping times, then E[XT | FS ]≤XS , a.s.
    In particular, E[XT ]≤E[XS ].
    3 If there exists an integrable random variable Y such that |Xn | ≤ Y
    for all n, and T is a stopping time which is finite a.s., then
    E[XT ]≤E[X0 ].
    4 If X has bounded increments, i.e. ∃M > 0 such that
    |Xn+1 − Xn | ≤ M for all n, and T is a stopping time with E[T ] < ∞,
    then E[XT ]≤E[X0 ].

    Hao Wu (MIT) 18.445 13 April 2015 11 / 15


    MIT OpenCourseWare
    http://ocw.mit.edu

    18.445 Introduction to Stochastic Processes


    Spring 2015

    For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
    Optional Stopping Theorem
    Theorem
    Let X = (Xn )n≥0 be a supermartingale.
    1 If T is a stopping time, then X T is also a supermartingale.
    In particular, E[XT ∧n ]≤E[X0 ].
    2 If S ≤ T are bounded stopping times, then E[XT | FS ]≤XS , a.s.
    In particular, E[XT ]≤E[XS ].
    3 If there exists an integrable random variable Y such that |Xn | ≤ Y
    for all n, and T is a stopping time which is finite a.s., then
    E[XT ]≤E[X0 ].
    4 If X has bounded increments, i.e. ∃M > 0 such that
    |Xn+1 − Xn | ≤ M for all n, and T is a stopping time with E[T ] < ∞,
    then E[XT ]≤E[X0 ].
    5 Suppose that X is a non-negative supermartingale. Then for any
    stopping time T which is finite a.s., we have E[XT ]≤E[X0 ].
    Hao Wu (MIT) 18.445 13 April 2015 11 / 15

    You might also like