Paper Menu >>
Journal Menu >>
Applied Mathematics, 2011, 2, 908-911 doi:10.4236/am.2011.27122 Published Online February 2011 (http://www.SciRP.org/journal/am) Copyright © 2011 SciRes. AM Definition of Laplace Transforms for Distribution of the First Passage of Z er o Lev el o f th e Semi-Mark o v Ra nd om Process with Positive Tendency and Negative Jump Tamilla I. Nasirova, Ulviyya Y. Kerimova Baku State University, Baku, Azerbaijan E-mail: shaxbazi_a@yahoo.com, ulviyye_kerimova@yahoo.com Received May 5, 2011; revised May 26, 2011; accepted May 29, 2011 Abstract One of the important problems of stochastic process theory is to define the Laplace transforms for the distri- bution of semi-markov random processes. With this purpose, we will investigate the semi-markov random processes with positive tendency and negative jump in this article. The first passage of the zero level of the process will be included as a random variable. The Laplace transforms for the distribution of this random variable is defined. The parameters of the distribution will be calculated on the basis of the final results. Keywords: Laplace Transforms, Semi-Markov Random Process, Random Variable, Process With Positive Tendency And Negative Jumps 1. Introduction There are number of works devoted to definition of the Laplace transforms for the distribution of the first pas- sage of the zero level. (Borovkov 2004) [1] defined the explicit form of the distribution, while (Klimov 1996) [2] and (Lotov V. I.) [3] indicated implicit form of the dis- tribution of the first passage of zero level. The presented work explicitly defines the Laplace transforms for the unconditional and conditional distribution of the semi-markov random processes with positive tendency and negative jump. 2. Problem Let’s assume that k and k , random variables 1k of independent random variable sequence 1 , kk k evenly distributed in probability face are given. Using these random variables we will derive the following semi-markov random process: ,,()FP 1 1 k i i Xtz t , if 1 11 ξξ1, kk ii ii tk X(t) process is the (asymptotic) semi-markov random processes with positive tendency and negative jump. Let’s include the0 1 random variable defined as be- low: 0 1min :0tXt where 0 1 , is the time of the first passage of X(t) process. We need to find Laplace transform for distribution of 0 1 random variable. 3. Definition of Laplace Transform for the Distribution of 0 1 Random Variable Let us set Laplace transform for the distribution of 0 1 random variable as L : 0 1 0 1 0, 0, 0 , LEe LzEe Xzz 1 In this case we can express the equation as 1,1 1 0 1 11 0, , 0, z Tz Thus, T and 0 1 are evenly distributed random vari- ables. Our goal is to find Laplace transform of relative and non-relative distribution of 0 1 random variable. According to the formula of total probability, we can put it as T. I. NASIROVA ET AL.909 00 1 11 1 11 11 :0 :0 |0 ddd T zz EeXze PePeP If to consider the following substitution 11 ;;ss yT we derive 0 1 11 0 11 00 0 11 0 11 00 11 0 11 0 |0d;d d;d; d dd dd dP dd s syzs zs s sy s syzs zs s sy s s s s EeXze Psy ePsyT ePs Py ePyPsLzsy eP szs eLPzsP 0 zs s or 11 0 11 00 ξd dξd s s zs s sy LzeP zsP s eLzsyPyP s (1) Let’s assume that zsy . In this case we will receive the following integral equation: 11 0 1 00 ξd ddξd s s zs s s LzePzsP s eL PzsP 1 s (2) We will solve this integral equation in special case. Let’s assume that 1 random variable has the Erlan- gian distribution of m construction, while 1 ξ random variable has the single construction Erlangian distribu- tion: 1 21 1 ξω 11, 0, 2!1 ! 1(), 0. m t t Pt tt tet m Ptet t where 0, 0 () 1, 0. t tt In this case Equation (2) will be as follows: 1 00 dd (1)! mz m mzs s zm s Lz e ee se mL s (3) We can derive differential equation from this integral equation. For this purpose, we will multiply both sides of Equation (3) by z e : 1 00 dd (1)! m zm mzs sm s eL z eseL m s If we increment both sides by z, we will get: 1 0d (1)! zz mz sms s eLzeLz eeseLzs m s In this case we will receive the differential equation with (m+1) construction: 1 0 1 1 mkk k m k mk mk z mz m CL zLz e eLz (4) The general solution of this differential equation will be t 12 12 z zm kk m zC LeCeCe z k (5) By finding ,, m CC 1 from Equation (3) we will get the following systematic equations: 1 000 ' 1 0 11 0 0 0 d 1! 00 d 1! 001 d mm s sm mss msm s mm kk x km mx k Lese m LL esLss m CLLe Lxx dLs (6) Copyright © 2011 SciRes. AM T. I. NASIROVA ET AL. 910 By exploitation of Equation (5), Equation (6) becomes 1 00 0 1 1 0 11 1 1 1 0 01 01 1! d 1! 1d i i i mm m m s sk m i i ms k i m mm m xkx m ii ii s ii i mm mm mxkx km mm mi ii x ki ii Cese m CkCex Cex m CkCkCeCe Ce x (7) After simplification of the last system we will get 11 11 1 1 11 11 1! 1! 1! 1! 1! 1 mm mm i imm ii ii m mm m ii ii m ii i i m mm mm km mmi mi im ii ii i Cm m Cmk k m Ck CC mk C CkC kCk m (8) or 1 1 1 0 1 () 0 0 m mm mi mi mm m iii m m ii m ii m mkm mi i m ki kC kk kC k Ck kC k (9) By exploitation of (m mii kk Equation (9) becomes 1 1 1 0 1 0 0 m mm mii mm i m iii i mkm mi ii k kC kkC Ckkk C or Copyright © 2011 SciRes. AM T. I. NASIROVA ET AL. Copyright © 2011 SciRes. AM 911 1 1 1 00 00 mmm ii i m i i m i i kC C C (10) Thus, (9) is a linear dependence equation system, as 23 1 1 0 m m m CC C Ck Then the general solution of integral Equation (3) will be as follows: 11 1 1 z z m kk m LzC ee k This expression is the Laplace transform for relative distribution of 0 1 random variable. Then, we will need to find L . In accordance with formula of total prob- ability, z0 LzdPX0L z and as the distribution of X(0) and 1 random variables are same, 1 1 1 1 z0 1 1z0 11 z0 1 1 dP 0 d 1! d 1! z z k m kmz mkz m m m LCeXz Ce ze m Cezz m C k z Therefore, 1 1 m m LC k This expression is the Laplace transform for non-rela- tive distribution of 0 1 random variable. Respectively, we will get the following characteristics for ߣm > ߤ: 0 1 32 3 2 2 3 32 2 0 12 2 3 2 2 32 2 0 1 0 123 0 3 02 1 ( 3 00 2 1 1 z 1( ) z m EL m mm m Lm m mm m mm DL Lm m m mmm m mm mz Em mm mz m Dmm 4. Conclusions In this article we have defined Laplace transforms for relative and non-relative distribution of the first passage of zero level of semi-markov random process with posi- tive tendency and negative jump. 5. References [1] A. A. Borovkov, “On the Asymptotic Behavior of the Distributions of First-Passage,” Mathematics and Stati- stics, Vol. 75, No. 1, 2004, pp. 24-39. doi:10.1023/B:MATN.0000015019.37128.cb [2] V. I. Lotov, “On the Asymptotics of the Distributions in Two-Sided Boundary Problems for Random Walks Defined on a Markov Chain,” Siberian Advances in Mathematics, Vol. 1, No. 3, 1991, pp. 26-51. [3] G. P. Klimov, “Stochastic Queuening Systems,” Nauka, Moscow, 1966. [4] T. I. Nasirova and R. I. Sadikova, “Laplace Trans- formation of the Distribution of the Time of System Sojourns within a Band,” Automatic Control and Computer Sciences, Vol. 43, No. 4, pp. 190-194. doi:10.3103/S014641160904004X |