This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Poisson Processes Class Notes for ISEN 618 September 27, 2007 1 Definitions and Preliminaries 1.1 Stochastic Processes Definition 1 A realvalued stochastic process is a family of random variables X = { X ( t ) ,t ∈ T } indexed by a set T , all defined on the same probability space (Ω , F ,P ) . For our purposes, we will typically take the set T to be either R + = [0 , ∞ ) or N + = { , 1 , 2 ,... } . The random functions X ( · ,ω ) : T → R , ω ∈ Ω are called sample paths or trajectories . A stochastic process will be described as, for example, rightcontinuous, increasing, nondecreasing, of bounded variation, etc., if the sample paths have these properties (a.s.). A stochastic process is said to be integrable if sup ≤ t< ∞ E  X ( t )  < ∞ ; square integrable if sup ≤ t< ∞ EX ( t ) 2 < ∞ ; bounded if there exists a finite constant Γ such that P ( sup ≤ t< ∞  X ( t )  < Γ) = 1 . 1 1.2 Filtrations Definition 2 A family of sub σalgebras {F t ,t ≥ } is called increasing if s ≤ t implies F s ∈ F t . An increasing family of sub σalgebras is called a filtration ; we will typically denote a filtration by F . When F = {F t ,t ≥ } is a filtration, the σalgebra T h> F t + h is denoted F t + . The corresponding limit from the left, F t , is the smallest σalgebra containing all the sets in S h> F t h and is written W h> F t h . A filtration is right continuous if, for any t , F t + = F t . The most natural filtration for a stochastic process X is called the “history”; i.e. F t = σ ( X ( s ) , ≤ s ≤ t ), the smallest σalgebra with respect to which each of the variables X ( s ) , ≤ s ≤ t is measurable. A stochastic process { X ( t ) ,t ≥ } is adapted to a filtration {F t ,t ≥ } if for every t ≥ ,X ( t ) is F tmeasurable. 1.3 Point Processes and Counting Processes A point process is a random distribution of points in a Euclidean space. For example, the points may represent shocks that occur randomly over time and that cause damage to a system (here, if we start observing the shocks at time 0, the space is R + ). Or, the points may represent the location of stars in a galaxy (here the space is R 3 ), or the locations of factories that may produce airborne pollutants, or the locations of defects on a semiconductor wafer (here the space is R 2 ). For ease of exposition, in this discussion, we’ll consider point processes on R + , and we will think of the points as times of occurence of a particular phenomenon. Let us begin as usual with a probability space (Ω , F ,P ). We have the fol lowing formal definition of a point process. Definition 3 A point process T = { T n ,n ∈ N + } is an ordered sequence of nonnegative random variables 0 = T < T 1 < ··· denoting the locations of 2 points....
View
Full Document
 Fall '07
 Klutke
 Probability theory, Exponential distribution, Poisson process, Tn

Click to edit the document details