張超權(quán) 唐勝達(dá) 秦永松
摘要譜負(fù)MAP是應(yīng)用概率論領(lǐng)域的重要內(nèi)容之一.利用Asmussen-kella鞅推廣了譜負(fù)MAP(X,J)的波動(dòng)理論,給出譜負(fù)MAP在與之獨(dú)立且服從Erlang分布的隨機(jī)時(shí)刻點(diǎn)上水平與極值的聯(lián)合變換所滿(mǎn)足的等式,進(jìn)而由Erlangization方法,給出譜負(fù)MAP(X,J)的水平與極值的聯(lián)合變換的瞬時(shí)趨近算法.
關(guān)鍵詞譜負(fù)MAP;波動(dòng)理論;Erlangization方法;趨近計(jì)算式
中圖分類(lèi)號(hào)O211.5文獻(xiàn)標(biāo)識(shí)碼A文章編號(hào)1000-2537(2014)02-0078-06
Levy過(guò)程是應(yīng)用概率領(lǐng)域內(nèi)的重要隨機(jī)過(guò)程之一,但是Levy過(guò)程的平穩(wěn)性使其在實(shí)際應(yīng)用中受到一定的局限性.在實(shí)際建模中,過(guò)程會(huì)因長(zhǎng)時(shí)間的演變、外界隨機(jī)因素的干擾等原因而不再具有平穩(wěn)性,如價(jià)格的季節(jié)性,行為的模式化等.由此,可將Levy過(guò)程推廣為機(jī)制轉(zhuǎn)換模型(regime-switching model):連續(xù)時(shí)間的Markov加過(guò)程(Markov additive process),簡(jiǎn)稱(chēng)MAP,這是Levy過(guò)程的一個(gè)自然推廣,MAP已成為隨機(jī)復(fù)雜系統(tǒng)的重要建模工具之一,它已被廣泛應(yīng)用于網(wǎng)絡(luò)通訊、存儲(chǔ)論、交通管理、風(fēng)險(xiǎn)過(guò)程、金融工程等領(lǐng)域[1-2].
許多學(xué)者對(duì)MAP的相關(guān)性質(zhì)作了深入的研究,Cinlar,Ney,Asmussen [3-5]給出了MAP的基本結(jié)構(gòu)及性質(zhì),Ivanovs[6]給出了MAP的指數(shù)矩陣特征值的性質(zhì),DAuria等[7]給出了MAP首達(dá)時(shí)過(guò)程的轉(zhuǎn)移率矩陣的結(jié)構(gòu),并將其應(yīng)用于單邊反射MAP及雙邊MMBM[8],Breuer[9]給出了首達(dá)時(shí)過(guò)程的轉(zhuǎn)移率矩陣的迭代計(jì)算方法,Ivanovs[10]給出了MAP的scale 矩陣,Kypianou等[11]對(duì)MAP波動(dòng)理論進(jìn)行了研究.
Avram[12]在研究風(fēng)險(xiǎn)過(guò)程中得出破產(chǎn)時(shí)刻的Laplace變換等價(jià)于指數(shù)隨機(jī)時(shí)間內(nèi)的破產(chǎn)概率,Asmussen等[13]采用fluid embedding方法將這一結(jié)果推廣并得出服從Erlang(n,q)分布的隨機(jī)時(shí)刻內(nèi)的破產(chǎn)概率,當(dāng)給定分布期望不變時(shí),隨著分布的階數(shù)趨于無(wú)窮,這一隨機(jī)時(shí)刻趨于它的期望定值,利用這一方法,Asmussen等得到了在有限時(shí)刻內(nèi)破產(chǎn)的趨近算法,且這一算法具有良好的穩(wěn)定性且收斂速度快,這一方法稱(chēng)為Erlangization方法;Stanford[14]將這一方法推廣為PH分布情形;Ramaswami等[15]將這一方法應(yīng)用于隨機(jī)流體理論,用于各種有限時(shí)刻內(nèi)的各種首達(dá)時(shí)的研究,Woolford等[16]將這一方法用于分析山火的控制研究.
本文基于上述研究,主要給出譜負(fù)MAP(X,J)的水平與極值的聯(lián)合變換的瞬時(shí)趨近算法.這一結(jié)果在實(shí)際數(shù)值計(jì)算中具有十分重要的意義,本文利用Asmussen-kella鞅,推廣了MAP的波動(dòng)理論,將MAP在指數(shù)時(shí)刻的相關(guān)量推廣至Erlang分布的隨機(jī)時(shí)刻上,繼而由Erlangization方法,給定任意時(shí)刻時(shí)的MAP相應(yīng)量的趨近計(jì)算式.從而解決了譜負(fù)MAP(X,J)瞬時(shí)波動(dòng)理論的瞬時(shí)時(shí)間點(diǎn)上的計(jì)算問(wèn)題.
參考文獻(xiàn):
[1]ASMUSSEN S. Applied probability and queues[M]. 2nd Ed. Berlin:Springer, 2003.
[2]PRABHU N. Stochastic storage processes—queues, insurance risk, dams, and data communication[M].New York:Springer-Verlag, 1998.
[3]INLAR E. Markov additive processes Ⅰ[J]. Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete, 1972,24(2):85-93.
[4]INLAR E. Markov additive processes Ⅱ[J]. Probability Theory Related Fields,1972,24(2):95-121.
[5]NEY P, NUMMELIN E. Markov additive processes Ⅰ. Eigenvalue properties and limit theorems[J]. Ann Probab, 1987,15(2):561-592.
[6]IVANOVS J, BOXMA O, MANDJES M. Singularities of the matrix exponent of a Markov additive process with one-sided jumps[J]. Stochastic Pro Appl, 2010,120(9):1776-1794.
[7]DAURIA B, IVANOVS J, KELLA O, et al. First passage of a Markov additive process and generalized Jordan chains[J]. J Appl Probab, 2010,47(4):1048-1057.
[8]DAURIA B, IVANOVS J, KELLA O, et al. Two-sided reflection of Markovmodulated Brownian motion[J]. Stochastic Models, 2012, 28(1):316-332.
[9]BREUER L. First passage times for Markov additive processes with positive jumps of phase type[J]. J Appl Probability, 2008,45(3):779-799.
[10]IVANOVS J. PALMOWSKI Z. Occupation densities in solving exit problems for Markov additive processes and their reflections[J]. Stochastic Processes Appl, 2012,122(9):3342-3360.
[11]KYPRIANOU A, PALMOWSKI Z. Fluctuations of spectrally negative Markov additive processes[C]//Séminaire de Probabilités XL[M]. Springer: Berlin, 2008,121-135.
[12]AVRAM F, USABEL M. Finite time ruin probabilities of processes with phase type jumps[J]. Insurance, Math Economics, 2003, 32(3):371-377.
[13]ASMUSSEN S, AVRAM F, USABEL M. Erlangian approximations for finite horizon ruin probabilities[J].Astin Bull, 2002,32(2):267-281.
[14]STANFORD D A, AVRAM F, BADESCU A B, et al. Phase-type approximations to finite-time ruin probabilities in the Sparre Andersen and stationary renewal risk models[J]. Astin Bull, 2005,35(1):131-144.
[15]RAMASWAMI V, DOUGLAS G, WOOLFORD D, et al. The erlangization method for Markovian fluid flows [J]. Ann Oper Res, 2008,160(1):215-225.
[16]STANFORD D L, ATOUCHE G, WOOLFORD D, et al. Erlangized fluid queues with application to uncontrolled fire perimeter[J]. Stochastic Models, 2005,21(23):631-642.
[17]JAGERMAN D L. An inversion technique for the Laplace transform with application to approximation[J]. Bell Syst Tech J, 1978,57(3):669-710.
[18]JAGERMAN D L. An inversion technique for the Laplace transform[J]. Bell Syst Tech J, 1982,61(8):1995-2002.
[19]ASMUSSEN S, KELLA O. A multi-dimensional martingale for Markov additive processes and its applications[J]. Adv Appl Probab, 2000,32(2):376-393.
(編輯沈小玲)
[10]IVANOVS J. PALMOWSKI Z. Occupation densities in solving exit problems for Markov additive processes and their reflections[J]. Stochastic Processes Appl, 2012,122(9):3342-3360.
[11]KYPRIANOU A, PALMOWSKI Z. Fluctuations of spectrally negative Markov additive processes[C]//Séminaire de Probabilités XL[M]. Springer: Berlin, 2008,121-135.
[12]AVRAM F, USABEL M. Finite time ruin probabilities of processes with phase type jumps[J]. Insurance, Math Economics, 2003, 32(3):371-377.
[13]ASMUSSEN S, AVRAM F, USABEL M. Erlangian approximations for finite horizon ruin probabilities[J].Astin Bull, 2002,32(2):267-281.
[14]STANFORD D A, AVRAM F, BADESCU A B, et al. Phase-type approximations to finite-time ruin probabilities in the Sparre Andersen and stationary renewal risk models[J]. Astin Bull, 2005,35(1):131-144.
[15]RAMASWAMI V, DOUGLAS G, WOOLFORD D, et al. The erlangization method for Markovian fluid flows [J]. Ann Oper Res, 2008,160(1):215-225.
[16]STANFORD D L, ATOUCHE G, WOOLFORD D, et al. Erlangized fluid queues with application to uncontrolled fire perimeter[J]. Stochastic Models, 2005,21(23):631-642.
[17]JAGERMAN D L. An inversion technique for the Laplace transform with application to approximation[J]. Bell Syst Tech J, 1978,57(3):669-710.
[18]JAGERMAN D L. An inversion technique for the Laplace transform[J]. Bell Syst Tech J, 1982,61(8):1995-2002.
[19]ASMUSSEN S, KELLA O. A multi-dimensional martingale for Markov additive processes and its applications[J]. Adv Appl Probab, 2000,32(2):376-393.
(編輯沈小玲)
[10]IVANOVS J. PALMOWSKI Z. Occupation densities in solving exit problems for Markov additive processes and their reflections[J]. Stochastic Processes Appl, 2012,122(9):3342-3360.
[11]KYPRIANOU A, PALMOWSKI Z. Fluctuations of spectrally negative Markov additive processes[C]//Séminaire de Probabilités XL[M]. Springer: Berlin, 2008,121-135.
[12]AVRAM F, USABEL M. Finite time ruin probabilities of processes with phase type jumps[J]. Insurance, Math Economics, 2003, 32(3):371-377.
[13]ASMUSSEN S, AVRAM F, USABEL M. Erlangian approximations for finite horizon ruin probabilities[J].Astin Bull, 2002,32(2):267-281.
[14]STANFORD D A, AVRAM F, BADESCU A B, et al. Phase-type approximations to finite-time ruin probabilities in the Sparre Andersen and stationary renewal risk models[J]. Astin Bull, 2005,35(1):131-144.
[15]RAMASWAMI V, DOUGLAS G, WOOLFORD D, et al. The erlangization method for Markovian fluid flows [J]. Ann Oper Res, 2008,160(1):215-225.
[16]STANFORD D L, ATOUCHE G, WOOLFORD D, et al. Erlangized fluid queues with application to uncontrolled fire perimeter[J]. Stochastic Models, 2005,21(23):631-642.
[17]JAGERMAN D L. An inversion technique for the Laplace transform with application to approximation[J]. Bell Syst Tech J, 1978,57(3):669-710.
[18]JAGERMAN D L. An inversion technique for the Laplace transform[J]. Bell Syst Tech J, 1982,61(8):1995-2002.
[19]ASMUSSEN S, KELLA O. A multi-dimensional martingale for Markov additive processes and its applications[J]. Adv Appl Probab, 2000,32(2):376-393.
(編輯沈小玲)