Global+Optimization+Algorithms+Theory+and+Application_Part25

Global+Optimization+Algorithms+Theory+and+Application_Part25...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 28.4 Some Discrete Distributions 481 distribution are listed in Table 28.3 43 and examples for its PDF and CDF are illustrated in Figure 28.3 and Figure 28.4 . parameter definition parameters λ = μt > (28.120) PMF P ( X = x ) = f X ( x ) = ( μt ) x x ! e − μt = λ x x ! e − λ (28.121) CDF P ( X ≤ x ) = F X ( x ) = Γ ( ⌊ k +1 ⌋ ,λ ) ⌊ k ⌋ ! = ∑ x i =0 e − λ λ i i !! (28.122) mean EX = μt = λ (28.123) median med ≈ ⌊ λ + 1 3 − 1 5 λ ⌋ (28.124) mode mode = ⌊ λ ⌋ (28.125) variance D 2 X = μt = λ (28.126) skewness γ 1 = λ − 1 2 (28.127) kurtosis γ 2 = 1 λ (28.128) entropy H( X ) = λ (1 − ln λ ) + e − λ ∑ ∞ k =0 λ k ln( k !) k ! (28.129) mgf M X ( t ) = e λ ( e t − 1) (28.130) char. func. ϕ X ( t ) = e λ ( e i t − 1) (28.131) Table 28.3: Parameters of the Poisson distribution. 4 8 12 16 20 24 28 l =1 l =6 l =20 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 f (x) X x l =1 l =3 l =6 l =10 l =20 l Figure 28.3: The PMFs of some Poisson distributions Poisson Process The Poisson process 44 [1914] is a process that obeys the Poisson distribution – just like the example of the telephone switchboard mentioned before. Here, λ is expressed as the product of the intensity μ and the time t . μ normally describes a frequency, for example μ = 1 min . Both, the expected value as well as the variance of the Poisson process are λ = μt . In 43 The Γ in Equation 28.122 denotes the (upper) incomplete gamma function. More information on the gamma function Γ can be found in Section 28.10.1 on page 532 . 44 http://en.wikipedia.org/wiki/Poisson_process [accessed 2007-07-03] 482 28 Stochastic Theory and Statistics 0.1 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 x F ( x ) X 5 10 15 20 25 30 l =1 l =3 l =6 l =10 l =20 discontinuous point 0.2 Figure 28.4: The CDFs of some Poisson distributions Equation 28.132 , the probability that k events occur in a Poisson process in a time interval of the length t is defined. P ( X t = k ) = ( μt ) k k ! e − μt = λ k k ! e − λ (28.132) The probability that in a time interval [ t,t + Δt ] 1. no events occur is 1 − λΔt + o ( Δt ). 2. exactly one event occurs is λΔt + o ( Δt ). 3. multiple events occur o ( Δt ). Here we use an infinitesimal version the small- o notation. 45 The statement that f ∈ o ( ξ ) ⇒ | f ( x ) | ≪ | ξ ( x ) | is normally only valid for x → ∞ . In the infinitesimal variant, it holds for x → 0. Thus, we can state that o ( Δt ) is much smaller than Δt . In principle, the above equations imply that in an infinite small time span either no or one event occurs, i.e., events do not arrive simultaneously: lim t → P ( X t > 1) = 0 (28.133) The Relation between the Poisson Process and the Exponential Distribution It is important to know that the (time) distance between two events of the Poisson process is exponentially distributed (see Section 28.5.3 on page 489 ). The expected value of the number of events to arrive per time unit in a Poisson process is...
View Full Document

This document was uploaded on 08/10/2011.

Page1 / 20

Global+Optimization+Algorithms+Theory+and+Application_Part25...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online