Scribd Logo
Upload

Welcome to Scribd!

  • 38 views

    Stochastic Processes

    Uploaded by

    melanocitos
    This document provides an overview of stochastic processes and introduces Poisson processes. It defines stochastic processes as random variables that depend on an additional independent variable like time, space, or energy. Common examples are time series with random impacts. The document then summarizes key aspects of stochastic processes like their dependence on the independent variable and their mean and autocorrelation functions. It presents an example of a random walk process on a lattice. Finally, it introduces Poisson processes and outlines three ways to generate Poisson points for further analysis.

    Copyright:

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

    Stochastic Processes

    Uploaded by

    melanocitos
    38 views6 pages
    This document provides an overview of stochastic processes and introduces Poisson processes. It defines stochastic processes as random variables that depend on an additional independent variable like time, space, or energy. Common examples are time series with random impacts. The document then summarizes key aspects of stochastic processes like their dependence on the independent variable and their mean and autocorrelation functions. It presents an example of a random walk process on a lattice. Finally, it introduces Poisson processes and outlines three ways to generate Poisson points for further analysis.

    Original Description:

    A brief introduction to stochastic processes.

    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 provides an overview of stochastic processes and introduces Poisson processes. It defines stochastic processes as random variables that depend on an additional independent variable like time, space, or energy. Common examples are time series with random impacts. The document then summarizes key aspects of stochastic processes like their dependence on the independent variable and their mean and autocorrelation functions. It presents an example of a random walk process on a lattice. Finally, it introduces Poisson processes and outlines three ways to generate Poisson points for further analysis.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    38 views6 pages

    Stochastic Processes

    Uploaded by

    melanocitos
    This document provides an overview of stochastic processes and introduces Poisson processes. It defines stochastic processes as random variables that depend on an additional independent variable like time, space, or energy. Common examples are time series with random impacts. The document then summarizes key aspects of stochastic processes like their dependence on the independent variable and their mean and autocorrelation functions. It presents an example of a random walk process on a lattice. Finally, it introduces Poisson processes and outlines three ways to generate Poisson points for further analysis.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 6

    i

    i
    9781107035904ar 2014/1/6 20:35 page 147 #161

    9
    Poisson processes and waiting times

    9.1 Stochastic processes


    If a classical particle moves in an external force field F (x) its motion is deterministic
    and it could be called a deterministic process. If the particle interacts with randomly
    distributed obstacles its motion is still deterministic, but the trajectory depends on the random characteristics of the obstacles, and maybe also on the random choice of the initial
    conditions. Such a motion is an example of a stochastic process, or rather a random
    process.
    In a more abstract generalized definition, a stochastic process is a random variable X
    that depends on an additional (deterministic) independent variable , which can be discrete or continuous. In most cases it stands for an index N or time t, space x or energy
    E. An almost trivial ubiquitous stochastic process is given by additive noise (t) on a
    time-dependent signal s(t), i.e. Xt = s(t) + (t). As a consequence, Xt is no longer continuous. The most apparent applications of stochastic processes are time series of any kind
    that depend on some random impact. A broad field of applications are time series occurring, for example, in business, finance, engineering, medical applications and of course
    in physics. Beyond time series analysis, stochastic processes are at the heart of diffusion
    Langevin dynamics, Feynmans path integrals [43], as well as Klauders stochastic quantization [62], which represents an unconventional approach to quantum mechanics. Here
    we will give a concise introduction and present a few pedagogical examples. An important
    example of stochastic processes, which we have encountered already, are random walks
    as discussed in Section 4.2 [p. 57] and Chapter 5 [p. 71]. We will have ample opportunity to encounter further examples of stochastic processes in various sections of this book:
    electron temperature distribution as a function of the laser cord position (Section 23.2),
    energy dependence of a signal on top of a background (Section 23.3), time dependence
    of the magnetic flux in a plasma (Section 24.1.2), surface growth as a function of time
    (Section 24.4), the yearly onset of cherry blossom over time (Section 25.1), and eruptions
    of Old Faithful geyser as a function of time (Section 25.2.2) There are many more examples of stochastic processes covered in this book, namely all those cases where a function
    f (x) is measured which is corrupted by (additive) random noise. An important class of
    stochastic processes are Markov processes. They are introduced in Section 30.3 [p. 544]

    i
    i

    i
    9781107035904ar 2014/1/6 20:35 page 148 #162

    148

    Poisson processes and waiting times

    as the basis for a widespread stochastic integration technique, called Markov chain Monte
    Carlo.
    For the sake of clarity we consider only a single random variable X that might, however,
    depend on an L-component vector of independent random variables. For fixed , X ()
    is an ordinary random variable, for which the CDF and PDF can be defined as usual:
    F (x) := P (X () x),
    d
    F (x).
    p (x) :=
    dx

    (9.1a)
    (9.1b)

    The really new aspects of stochastic processes concern the dependence on the independent
    variable , which, as said before, typically represents time, space or energy.
    Stochastic processes are referred to as continuous (discrete) stochastic processes if they
    are based on continuous (discrete) input variables . That doesnot, however, imply that
    X () as a function of is necessarily continuous (discrete). As a counterexample consider
    a (discrete) Bernoulli experiment, whose outcome o {o1 , o2 } defines the global behaviour
    
    cos(t)
    for o = o1
    Xt (o) :=
    exp(t) for o = o2 .
    This is a case of a discrete stochastic process which is continuous in t. The opposite situation is given by the following process:
    Xt ( ) := 1(t ),
    with being a (continuous) exponential random variable. Apparently, Xt is discontinuous
    if considered as a function of t, but the process is referred to as continuous.
    Interesting aspects of stochastic processes are the -dependence of individual realizations (samples)
    X := X ( )
    for a fixed , or the -dependence of the mean, i.e.
    
    E( ) = X := dx x p (x)
    
    
    L
    dx x p (x|)
    = d p()
      
    (xX ())

    
    =

    d L X () p().

    In the first line we have introduced the independent random variable via the marginalization rule and in the second we have utilized the fact that x is uniquely determined if and
    are specified. Another interesting aspect of stochastic processes is the autocorrelation
    in , i.e.

    i
    i

    i
    9781107035904ar 2014/1/6 20:35 page 149 #163

    9.1 Stochastic processes

    A(,  ) := X  X =

    149

    d L X  () X () p() X  X .

    The autocorrelation plays a crucial role in MCMC, as outlined in Section 30.3 [p. 544]. As a
    simple introductory example we consider a one-dimensional random walk on a lattice, with
    sites x Z. In addition, we assume a parabolic potential, in order to avoid an increasing
    variance, which we found for the unrestricted random walk in Section 4.2 [p. 57]. Let Xt be
    the position of the walker at discrete time t and in each time step the walker has only three
    choices: stay or move one step forward or backward. That is, Xt+1 = Xt + X. A possible
    choice for the probability P (X|Xt , I) that the walker will be at site Xt+1 in the next time
    step can be defined via the MetropolisHastings algorithm (see Section 30.3 [p. 544]),
    resulting in
    
    '
    (
    1
    exp 2 Xt , 1 ,
    P (X = 1|Xt , I) = min
    2
    

    P (X = 0|Xt , I) = 1 P (X = +1|Xt , I) P (X = 1|Xt , I).


    This process is actually a Markov process as the next step depends on the actual position,
    but otherwise it has no memory. In Figure 9.1 the time evolution of a few individual random
    walks Xt with = 0.05 is depicted. The walks always start at X0 = 0. In addition, the
    sample mean Xt and sample variance (Xt )2 , obtained from 1000 independent repetitions
    of such random walks, are displayed. As expected from symmetry, the sample mean is
    Xt 0. The sample variance initially increases linearly with time, according to a diffusion
    process, but then levels off due to the parabolic potential, which prevents the walkers from
    deviating too far from the origin. In Figure 9.1(b) the autocorrelation is depicted, which
    has been computed from the same sample of random walks. Here the autocorrelation of a
    single random walk is defined as

    1
    10

    0.8
    A(t)

    5
    Xt
    0

    0.4
    0.2

    5
    0
    (a)

    0.6

    0.2

    0.4

    0.6
    t /T

    0.8

    1
    (b)

    20

    40

    60

    80

    100

    Figure 9.1 (a) Three individual random walks, all starting at X0 = 0, are depicted as jagged lines in
    different shades of grey. The time is expressed in units of the total time T = 1000. In addition, the
    sample mean (thick solid line) and variance (dashed line) of N = 1000 independent random walks
    are displayed. (b) The corresponding scaled autocorrelation function A(t) versus lag size t.

    i
    i

    i
    9781107035904ar 2014/1/6 20:35 page 150 #164

    150

    Poisson processes and waiting times

    A(t)
    :=

    t1
    
    1
    Xt+t Xt ,
    t1 t0 + 1 t=t
    0

    Xt := Xt Xt .
    The autocorrelation is determined in the time window ranging from t0 = 100 to t1 = 900.
    The first 100 steps have been discarded, in order to enter the region where Xt+t Xt is
    independent of t, then the process is said to be homogeneous. There are a couple more
    subtleties to be accounted for in computing autocorrelations numerically, which are, however, immaterial for the basic understanding of this topic. In Figure 9.1 the scaled autocor

    relation function A(t) := A(t)/


    A(0),
    averaged over the sample, is shown. We see that
    the autocorrelation decays exponentially, which is typical behaviour. A fit of exp{t/ }
    to the data defines the autocorrelation length . We recognize that the random walk has a
    rather long autocorrelation length 20, i.e. there is a long-range correlation between
    fluctuations. For a small autocorrelation A(t) 0, it is necessary that Xt+t Xt changes
    sign frequently as a function of t. Apparently, the autocorrelation length defines a typical
    time window in which the walk is either above or below the mean value. More details on
    autocorrelation functions of Markov processes are given in Section 30.3 [p. 544].
    In the next sections we focus on Poisson processes and various aspects of Poisson points.
    As mentioned before, we will encounter various examples of stochastic processes later in
    this book. For further details on stochastic processes we refer the reader to [8, 159]. Readers particularly interest in mathematically rigorous definitions and properties of stochastic
    processes may look at [45].

    9.2 Three ways to generate Poisson points


    We have introduced the Poisson distribution as a limiting case of the binomial distribution
    by keeping = pN fixed and considering the limit N . We use this situation to
    generate Poisson points as outlined in the next section.

    9.2.1 Global generation of Poisson points


    We consider a finite interval L = (0, L) of length L, in which we generate N independent and uniformly distributed points. They are called Poisson points (PPs). The point
    density is
    :=

    N
    .
    L

    (9.2)

    We now select an arbitrary interval Ix L of length x. The probability for an individual


    point to appear in Ix is, according to the classic definition,
    p=

    x
    .
    L

    i
    i

    i
    9781107035904ar 2014/1/6 20:35 page 151 #165

    9.2 Three ways to generate Poisson points

    151

    The probability that n PPs land in Ix is therefore binomial. The mean number of points
    generated on Ix is
    = p N = x .

    (9.3)

    Next we erase these PPs and start afresh, but this time we double both the length L and the
    number of points N . We retain the interval Ix . As the point density is the same, the mean
    number () of points in Ix is also the same. Upon repeating the doubling process we fulfil
    the requirements of the Poisson theorem ( = pN = fixed and N ). Hence, in the
    limit of N ,1 we obtain a Poisson distribution
    P (n|x, , I) := ex

    (x)n
    .
    n!

    (9.4)

    Distance law of neighbouring PPs


    Next, we are interested in the PDF of the distances between Poisson points p( |, I),
    where stands for the proposition
     : The distance to the next Poisson point is within [, + d ).
    The proposition requires that there is no Poisson point in the interval [0, ), and there must
    be a Poisson point in the following interval d . Both events are logically independent. Both
    probabilities follow from the Poisson distribution P (n|x, , I) given in equation (9.4) with
    (n = 0 ; x = ) and (n = 1 ; x = d ), respectively. This yields the sought-for probability
    P ( |, I) = P (n = 0|, , I) P (n = 1|d, , I) = e d
    and the corresponding PDF is exponential.
    Distance law of Poisson process
    p( |, I) = e .

    (9.5)

    Its mean value is 1/, in accordance with the mean density of .


    Distance law of PPs
    The distances between neighbouring PPs are exponentially distributed. Next we want to
    consider the distribution of the distance dn between Poisson point m and m + n. As the
    individual intervals are independent, the distances are independent of m and they are a
    sum of n independent and exponentially distributed random numbers. According to equation (7.68) [p. 130], such a sum is Gamma distributed.
    1 See equation (6.12) [p. 88].

    i
    i

    i
    9781107035904ar 2014/1/6 20:35 page 152 #166

    152

    Poisson processes and waiting times

    Distance distribution of Poisson points


    p(p(dn |, I) = p (dn |n, ) =

    ( dn )n1 e dn .
    (n)

    (9.6)

    9.2.2 Successive generation of Poisson points


    The derived exponential distribution suggests an alternative view of the generation process.
    So far the Poisson points have been generated by distributing the points homogeneously
    on an interval. In this way it is warranted that the points are independent of each other.
    Similarly, the distances between the points are i.i.d. For this reason also a sequential generation process can be used. We start at x = 0 and determine the next Poisson point x1
    at a distance of 1 according to the exponential distribution equation (9.5). Then the next
    point x2 is generated at a distance 2 , again derived from p( |, I), and so on. The Poisson
    points are then located at
    xn =

    n
    

    i .

    i=1

    The PDF implies that the probability for the next PP being at a distance is given by
    p(d |, I) d = (1 d ) d.
    This expression can be interpreted as the probability for no event (1 dx) in d times
    the probability for one event d in the following infinitesimal interval d . In other words,
    the events in infinitesimal intervals are uncorrelated and the probability Pd for an event in
    d is analytic in d :
    Pd = d.
    As we will see in the next section, these features entail the Poisson distribution.

    9.2.3 Local generation of Poisson processes


    The Poisson process can also be defined by the two axioms derived before.
    The probability for an event happening in an infinitesimal interval d is independent of previous PPs.
    The probability is analytic in d , i.e. P (event in d |, I) = d .

    It would be straightforward to identify the underlying Bernoulli experiment and to derive


    the Poisson distribution as a limiting case. Here, however, we want to present a different
    approach, that is over the top for the present problem but turns out to be very useful in more

    i
    i

    You might also like