29
Appendix A Laplace-Stieltjes Transforms A.1 Laplace-Stieltjes Transforms Let F(t} be a well-defined function of t specified for t 0 and s be a complex number. If the following Stieltjes integral: (ALI) converges on some So. the Stieltjes integral (A.LI) converges on s such that > The integral (A.LI) is called the Laplace-Stieltjes transform of F(t). If the real function F(t) can be expressed in terms of the following integral: F(t) = lot dF(x) = lot f(x)dx, (AL2) then (AL3) which is called the Laplace transform of f (t). Noting that F(t) is in correspondence with F*(s) (ref. Theorem 2.2.2 and Table 2.2.2), F(t) can be uniquely specified by F*(s). The inversion formula for obtaining F{t) from F*{s) can be given by 1 l b + ic eat F(t) = lim -. -F*(s)ds, c-oo 211"z b-ic S (AI.4) where i = A is an imaginary unit, b > ma.x(u,O} and u is a radius of conver- gence. The following two theorems are well-known and of great use as the limit theorems for the Lapla.ce-Stieltjes transform F*(s) of F(t).

Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Embed Size (px)

Citation preview

Page 1: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Appendix A

Laplace-Stieltjes Transforms

A.1 Laplace-Stieltjes Transforms

Let F(t} be a well-defined function of t specified for t ~ 0 and s be a complex number. If the following Stieltjes integral:

(ALI)

converges on some So. the Stieltjes integral (A.LI) converges on s such that ~(s) > ~(so). The integral (A.LI) is called the Laplace-Stieltjes transform of F(t). If the real function F(t) can be expressed in terms of the following integral:

F(t) = lot dF(x) = lot f(x)dx, (AL2)

then

(AL3)

which is called the Laplace transform of f (t). Noting that F(t) is in one-t~one correspondence with F*(s) (ref. Theorem

2.2.2 and Table 2.2.2), F(t) can be uniquely specified by F*(s). The inversion formula for obtaining F{t) from F*{s) can be given by

1 l b+ic eat F(t) = lim -. -F*(s)ds, c-oo 211"z b-ic S

(AI.4)

where i = A is an imaginary unit, b > ma.x(u,O} and u is a radius of conver­gence.

The following two theorems are well-known and of great use as the limit theorems for the Lapla.ce-Stieltjes transform F*(s) of F(t).

Page 2: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

242 APPENDIX A. LAPLACE-STIELT JES TRANSFORMS

Theorem A.l (An Abelian Theorem) If for some non-negative number a,

lim F(t) = C t-+oo to. r(a+1)'

(A.l.5)

then

lim so. F*(s) = C, 8-++0

(A.l.6)

where r(k) = 1000 e-xxk-1dx is a gamma function of order k defined in Eq.(2.4.13).

Theorem A.2 (A Tauberian Theorem) If F(t) is non-decreasing and the Laplace­Stieltjes transform

F*(s) = 1000 e-8tdF(t) (A.l.7)

converges for ~(s) > 0, and if for some non-negative number a,

lim so. F*(s) = C, 8-++0

(A.l.8)

then

lim F(t) = C . t-+oo to. r(a+1)

(A.l.9)

Exampe A.l.l (Elementery Renewal Theorem) The Laplace-Stieltjes trans­form of the renewal function is given in Eq.(4.2.15):

M*( ) = F*(s) s 1 - F*(s)'

where F*(s) is the Laplace-Stieltjes transform of F(t). Let us apply Theorem A.2.

We have

lim sM*(s) = lim F*(s) 8-++0 8-++0 [1 - F*(s)]/s

since

and

lim F*(s) = l. 8-+0

1 J.£'

That is, from the Tauberian Theorem, we have

lim M(t) = .!:.. t-+oo t J.£

Page 3: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

A.2. PROPERTIES OF LAPLACE-STIELT JES TRANSFORMS 243

A.2 Properties of Laplace-Stieltjes Transforms

For the Laplace-Stieltjes transform, we have the following relationship:

(A.2.1)

That is, the Laplace-Stieltjes transform F* (s) can be obtained by s times the Laplace transform of F(t). We can easily obtain the Laplace-Stieltjes transforms from the corresponding Laplace transforms, since most textbooks only discuss the latter.

Table A.l shows the the general properties of the Laplace-Stieltjes transforms. The general properties in Table A.l can be applied in practice. Besides the general properties in Table A.l, Table A.2 shows the important formulas for the Laplace-Stieltjes transforms. Such formulas in Table A.2 can be applied to derive the Laplace-Stieltjes transform F*(s) from F(t), and vice versa.

Table A.l General properties of the Laplace-Stieltjes transforms.

F(t) F*(s) = 1~ e-stdF(t)

Fl(t) + F2 (t) F;(s) + F;(s)

aF(t) aF*(s)

F(t - a) (a> 0) e-saF*(s)

F(at) (a> 0) F*(s/a)

e-at F(t) (a> 0) s

--F*(s+a)

F'(t) = dF(t) s+a s[F*(s) - F(O)]

dt tF'(t) dF·(s) -s ds

l F(x)dx !F*(s) tOt s

10 ···10 F(t)(dtt ~F*(s) sn lim F(t) t-++O lim F*(s)

s-+oo

lim F(t) t-+oo

lim F*(s) s-++O

Page 4: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

244 APPENDIX A. LAPLACFrSTIELT JES TRANSFORMS

Table A.2 Formulas of the Laplace-Stieltjes Transforms

F(t) F·(s) = LX) e-8tdF(t)

6(t - a)t (a > 0) se-SO

l(t - a)* (a> 0) e-SO

l(t) 1

t 1 -s

tn (n : a positive integer) n! -sn

ta (a> -1) r(a + 1)

sa

e-at (a> 0) s --

s+a te-at (a> 0)

s (s + a)2

tne-at (a> 0) n!s

(s + a)n+1

tfje-at (a> 0,{3 > -1) sr({3 + 1) (s + a)fj+1

S2 (!R(s) >1 a I) cos at

S2 +a2

sin at sa

(!R(8) >1 a I) 82 +a2

cosh at 82

(!R(8) >1 a I) S2 - a 2

sinh at sa

(!R(8) >1 a I) 82 - a 2

logt -"'( -logs §

t Dirac's delta function. + Heaviside's unit function. § "'( = 0.57721 ... , Euler's constant.

Page 5: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

A.3. APPLICATIONS TO DISTRiBUTIONS 245

A.3 Applications to Distributions

We discuss the Laplace-Stieltjes transforms of the distributions. As shown in Section 2.3, we have introduced six common discrete distributions. We derive the Laplace-Stieltjes transforms for a few discrete distributions.

Example A.3.1 (Binomial distribution)

Fi(s} = ~ e-s:c (:) p:cqn-:c = (pe-s + qt . (A.3.1)

Applying the formulas for the moments, we have

E[X] = (_I}dFx(s) I = np, ds s=o

(A.3.2)

cPF*(S}1 E[X2] = (_1)2 x2 = n(n - l}p2 + np, ds s=o

(A.3.3)

which imply

Var(X) = E[X2] - E[X]2 = npq. (A.3.4)

Example A.3.2 (Geometric distribution)

(A.3.5)

Example A.3.3 (Negative binomial distribution)

F*(s}=~e-s:c(X-l) r :c-r= [ pe-s ]r X ~ X - r p q 1 - qe-S

(A.3.6)

As shown in Example 2.3.2, if Xl, X 2 , ••• ,Xr are independent and identi­cally distributed random variables with Xi '" GEO(p) , the random variable Sr = Xl + X 2 + ... + Xr is distributed with the negative binomial distribution Sr '" N B(p, r). Here we have verified this fact by using the Laplace-Stieltjes transforms.

We next show the Laplace-Stieltjes transforms for the continuous time dis­tributions in Section 2.4.

Example A.3.4 (Exponential distribution)

F.i(s} = t JO e-stdFx(t} = roo e-st ~e->'tdt = _~_. k k s+~

(A.3.7)

Applying the formulas for the moments, we have

E[X] = (_I}dFx(s) I = l/~, ds s=O

(A.3.8)

Page 6: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

246 APPENDIX A. LAPLACE-STIELT JES TRANSFORMS

which imply

2 1 1 Var(X) = >.2 - >.2 = >.2'

Example A.3.5 (Gamma distribution)

Applying the formulas for the moments, we have

k Var(X) = >.2'

Example A.3.6 (Equilibrium distribution)

(A.3.9)

(A.3.1O)

(A.3.11)

(A.3.12)

(A.3.13)

In Chapter 4, we have introduced the equilibrium distribution in Eq.( 4.3.44):

1 lot Fe(t) = - [1 - F(y)] dy, J.t 0

(A. 3.14)

where J.t is the mean of F(t). The Laplace-Stieltjes transform of Fe(t) is given by

F;(s) = fo'X) e-st dFe(t)

= ~ t'" e-st d [1 - F(t)] (see Table A.l) J.tS io

=~[I-F*(s)]. J.tS

(A.3.15)

A.4 Applications to Differential Equations

Example A.4.1 (Example 6.4.5) We have discussed a two-state Markov chain whose Kolmogorov's forward equations are given by

P~o(t) = ->.Poo(t) + J.tP01(t),

P~l(t) = >.Poo(t) - J.tPOl(t),

(A.4.1)

(A.4.2)

Page 7: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

AA. APPLICATIONS TO DIFFERENTIAL EQUATIONS

with the initial conditions that Poo(O) = 1 and POI (0) = o. Let

P;is) = 1000 e-st dPoj(t)

247

(A.4.3)

be the Laplace-Stieltjes transforms of POj(t) (j = 0,1). Noting the initial con­ditions and using Table A.l, we have the Laplace-Stieltjes transform expressions for Eqs.(A.4.1) and (A.4.2):

sP;O(s) - s = -AP;o(s) + JLP;1(S),

SP;1(S) = AP;o(s) - JLP;1(S),

whose solutions are given by

P.* ( ) s + JL s + _JL_. A + JL 00 s = s+A+JL s+A+JL A+JL s+A+JL'

P.*() A A A+JL 01 s = S + A + JL - A + JL . s + A + JL .

(A.4.4)

(A.4.5)

(A.4.6)

(A.4.7)

Applying the formulas in Table A.2, we have the following inversions of the Laplace-Stieltjes transforms:

POO(t) = _JL_ + _A_e-(>'+I')t, A+JL A+JL

(A.4.8)

( ) A A -(>.+)t POI t = -- - --e 1', A+JL A+JL

(A.4.9)

which have been given in Example 6.4.5.

Example A.4.2 (M/M/2/2/2 queueing modeD We have discussed an M/M/c/c/c queueing model with finite population in Section 9.4 in general. We restrict ourselves to a case of c = 2 (i.e., 2 machines). Let QOj(t) be the probabilities that (2 - j) machines are operating at time t given that 2 machines are operating at t = 0, where j = 0,1,2. Kolmogorov's forward equations are given by

Q~O(t) = -2AQoo(t) + JLQ01 (t),

Q~1(t) = 2AQoo(t) - (A + JL)Q01(t) + 2JLQ02(t) ,

Q~2(t) = AQOl(t) - 2JLQ02(t),

with the initial conditions Qoo(O) = 1 and QOj(O) = 0 (j = 1,2). Let

Q~j(s) = fooo e-st dQoj(t)

(A.4.1O)

(A.4.11)

(A.4.12)

(A.4.13)

be the Laplace-Stieltjes transforms of QOj(t) (j = 0,1,2). The Laplace-Stieltjes transform expressions for Eqs.(A.4.lO), (A.4.l1), and (A.4.12) are given by

sQ~(s) - s = -2AQ~(s) + JLQ~1(S), (A.4.14)

Page 8: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

248 APPENDIX A. LAPLACE-STIELT JES TRANSFORMS

SQ~l (s)

SQ~2(S)

= 2AQ~O(S) - (A + /-l)Q~l(S) + 2/-lQ~2(S),

= AQ~l(S) - 2/-lQ~2(S).

Solving with respect to Qoo(s), we have

* S2 + (A + 3/-l)s + 2/-l2 Qoo(s) = (s + A + /-l) [s + 2(A + /-l)]

= 1- 2A/-l A + /-l (A + /-l)2 s + A + /-l

A2 2(A+/-l) (A + /-l)2 . S + 2(A + JL)"

(A.4.15)

(A.4.16)

(A.4.17)

where the last equation has been derived by applying the partial fraction ex­pansion (or decomposition). Applying the formulas in Table A.2, we have the following inversion of the Laplace-Stieltjes transform in Eq.(A.4.17):

(A.4.18)

where Poo(t) has been given in Eq.(A.4.8). Of course, Eq.(A.4.18) can be easily derived by considering two MIMl11111 queueing models in parallel (cf. Section 9.4). Similarly, we have the following Laplace-Stieltjes transforms:

(A.4.19)

2A2 Q~2(S) = (s + A + JL) [s + 2(A + JL)]' (A.4.20)

which imply

Q01(t) = 2POO (t)P01 (t), (A.4.21)

(A.4.22)

where Poo(t) and P01 (t) have been given in Eqs.(A.4.8) and (A.4.9), respectively.

Page 9: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

A.5. APPLICATIONS TO RENEWAL FUNCTIONS 249

A.5 Applications to Renewal Functions

In Chapter 4 we have developed renewal processes. Let {N(t), t ~ O} be a renewal process with interarrival distribution F(t). The renewal function is given by

00

M(t) = E [N(t)] = L F(n)(t). (A.5.I) n=l

The Laplace-Stieltjes transform of MCt) is given by

M*(s) = F*(s) 1 - F*(s) , (A.5.2)

where F*Cs) is the Laplace-Stieltjes transform of F(t).

Example A.5.1 If we assume that F(t) obeys a gamma distribution, i.e., Xi '" GAM('x, k) in general (k is a positive integer), we have the Laplace­Stieltjes transforms of F(t) and M(t), respectively:

F*(s) = c~'xr, (A.5.3)

M * ( ) _ (sh) k _ 1 s - k - k·

1 - (8~>') (1 + f) - 1 (A.5.4)

Let €r = e¥ be all the distinct roots of an equation Sk = 1, where r = 0,1,2,···, k - 1. By applying the partial fraction expansion, we have

M*Cs) = .! E 'x€r . kr=os+,X(I-€r)

By inversion, we have

,Xt 1 k-l € MCt) = - + - L _r_ [1- e->.t(l-er )] ,

k k r=l 1 - €r

which has been given in Example 4.2.3.

Example A.5.2 If we assume that

F(t) = 1 - ~ (e->.t + e-2>'t) ,

we have the Laplace-Stieltjes transform of F(t):

rcs) = 1 _ .! (_S_ + _S_) . 2 s +,X s + 2,X

CA.5.5)

(A.5.6)

(A.5.7)

CA.5.8)

Page 10: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

250 APPENDIX A. LAPLACE-STIELT JES TRANSFORMS

The Laplace-Stieltjes transform of M(t) is given by

M*(s) = F*(s) 1 - F*(s)

3>.s + 4>.2 s(2s + 3>')

= 4>. + !. 3>./2 3s 9 s + 3>./2

(A.5.9)

by applying the partial fraction expansion. By inversion (i.e., applying the for­mulas in Table A.2), we have

4>.t 1 ( a) M(t) = """3 + 9 1 - e-"2 At . (A.5.1O)

Example A.5.3 If we assume that F(x) obeys a negative binomial distribution of order 2 (refer to Section 2.3), we have the following Laplace-Stieltjes transform:

F* (s) = (--'pe=---_S_) 2

1 - qe-S (A.5.11)

The Laplace-Stieltjes transform M*(s) of the renewal function M(x) is given by

M*(s) = F*(s) 1 - F*(s)

= pe2-s [l_le_s - 1- (1 ~ 2P)e-s ]

= f: !!. [1 - (1 - 2ptl e-s(n+l), n=12

which implies the (discrete) renewal density

m(x) = ~ [1- (1- 2p)X-l] (x = 2,3,·· .).

That is, the (discrete) renewal function is given by

00

M(x) = Lm(j) j=2

= p; - ~ + (1-42P)X (x = 2,3,···)

(see Problems 4.4 and 4.5).

(A.5.12)

(A.5.13)

(A.5.14)

Page 11: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Appendix B

Answers to Selected Problems

Problems 1

1.2

(i) ABccc, (ii) Au B u C, (iii) AB u AC U BC,

(iv) ABC, (v)ABCCC U AC BCc uN BCC,

(vi) ABc U ABcC U AC BC = (AB U AC U BC) - ABC,

(vii) NWCc, (viii) (ABC)c.

1.5

(i) P{A I B} = 1/3, (ii) P{B I A} = 1/4, (iii) P{A - B} = 1/4,

(iv) P{B - A} = 1/6.

1.9

(i) (3x2 - 2y)3 = 27x6 - 54x4 y + 36x2y2 - 8y3,

(ii) (4x + 3y2)3 = 64x3 + 144x2y2 + 108xy4 + 27y6.

1.10

(51°) (~) = 921,200 ways.

Page 12: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

252 APPENDIX B. ANSWERS TO SELECTED PROBLEMS

1.12

(ii) 27 ways.

1.13

(ii) 20 ways.

1.14

n r ( distinguisable ),

1.15

Problems 2

2.7

2.14

n 1 2 3 4 5 6 7 8 9 10 11

(1') E[X] 1 = \ \' 1\1 + 1\12

( n+sS-l) (indistinguisable) .

Theoretical values Data 18.27 14 13.81 12 10.44 12 7.89 8 5.96 7 4.51 4 3.41 2 2.57 4 1.95 4 1.47 3 1.11 2

Page 13: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Problems 3

3.3

3.4

3.10

3.11

(i) t = 0 (9: 00 a.m.), t = 9 (6: 00 p.m.),

E[N(9)] = At = 90 persons,

Var(N{9)) = At = 90 persons2 •

(U) P{N(O.5) = O} = e-5 = 0.00674.

(i) P{N(30) = O} = e-5 ,

P{N(30) = 2} = 25e-5 2 .

P{N(30) = I} = 5e-5 ,

(ii) Upper limit 90 + 3V90 = 118.3 persons.

Lower limit 90 - 3V90 = 71.7 persons.

m(lO) = 440 persons, Variance = 440 persons2 .

F{Xl ~ t} = 1 - e-(at)/l (Weibull distribution).

Problems 4

4.3

4.4

Var(N(t)) = 2M * M{t) + M(t) - [M(t)]2

(i) F*(s) = pe- s /(1 - qe-S ),

(ii) M*(s) = pe-s /(1 - e-S ), M{m) = pm (m = 1,2, ... ).

253

Page 14: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

254 APPENDIX B. ANSWERS TO SELECTED PROBLEMS

4.5

M(m) = pm _ ~ + (1 - 2p)m 2 4 4

(m = 2,3,·· .).

Problems 5

5.1

(i) 11'(2) = [0.68 0.32], 11'(4) = [0.6672 0.3328], 11'(8) = [0.6667 0.3333].

(ii) 11'(2) = [0.66 0.34], 11'( 4) = [0.6664 0.3336], 11'(8) = [0.6667 0.3333].

(iii) 1I'(n) = [2/3 1/3] (n = 2,4,8).

5.2

5.4

5.5

5.8

(iii) pn = [1- hr an

where

(n = 1,2,3,· .. ).

(iii) One class {O, 1, 2}, recurrent and aperiodic.

(ii) i 2

Pi,i-i = N2' i(N - i)

Pii = 2 N2 ' (N - i)2

Pi,i+1 = N

(i=O,1,2,···,N), Pi,i-i + Pii + pi,i+1 = 1.

(iii) One class {O, 1,2,···, N}, recurrent and aperiodic.

Pi: {O, 1, 2} recurrent.

P 2 : {I} recurrent (absorbing), {O,2} transient.

P 3 : {O, I} recurrent, {2} recurrent (absorbing), {3} transient.

Page 15: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

255

5.9

PI: {O, 1, 2} recurrent, periodic with period 3.

P 2 : {O, 1, 2} recurrent, aperiodic.

P 3 : {O, 1,2,3, 4} recurrent, aperiodic.

5.11

(iii) PI: 11'0 = 11'1 = 1/2.

P 2 : 11'0 = 11'1 = 1/2.

P 3 : 11'0 = 11'1 = 1/2.

(iv) r pn [1/2 n.!..rr;., i = 1/2 1/2] 1/2

(i = 1,2).

(v) lim .! fpi = [1/2 1/2] n-HX) n . 3 1/2 1/2 .

0=1

5.12

[1/3 2/3 0 0

o 1 1/3 2/3 0 0 0 lim pn = 0 0 1 0 o . n-+oo

3/5 2/5 0 0 0 0 0 0 3/5 2/5

5.13

2' --+ 0, 4' --+ 1, 0' --+ 4, l' --+ 5, 3' --+ 2, 5' --+ 3.

0 1/3 2/3 0 0 0 0 1 1/2 1/2 0 0 0 0

p=2 0 0 2/3 1/3 0 0 3 0 0 1/2 1/2 0 0 4 0 1/4 0 0 1/4 1/2 5 0 0 3/4 0 1/4 0

Page 16: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

256 APPENDIX B. ANSWERS TO SELECTED PROBLEMS

A state transition diagram of Problem 5.13.

0 3/7 4/7 0 0 0 0 1 3/7 4/7 0 0 0 0

r pn 2 0 0 3/5 2/5 0 0 1m = 0 0 3/5 3/5 0 n-+oo 3 0

4 6/35 8/35 9/25 6/25 0 0 5 3/70 4/70 27/50 18/50 0 0

5.15

o [0 p 0

~l p = 1 q 0 P 2 0 q 0 3 p 0 q

Doubly stochastic (irreducible, recurrent, periodic with period 2).

5.16

1 0 ~ 0 ; 0 q

0 1 0 ~ 0 ~ lim p2n = (1 - ~)

q q q

1 0 ~ 0 ; 0 n-+oo q q

0 1 0 ~ 0 ~ q q q

Page 17: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

257

0 !. 0 ~ 0 ~ q q

1 0 ~ 0 ; 0 q

lim p2n+1 = (1 - p) 0 !. 0 ~ 0 ~ q q q n--+oo q

1 0 ~ 0 ~ 0 q4

Problems 6

6.1

M(t) = e>.t.

6.2 1 1 1

E[N(t) = n) = ~ + 2,\ + ... + n'\)

1 1 1 Var(N(t) = n) = ,\2 + 22,\2 + ... + n2,\2'

6.3

1 1 1 E[X(t) = kJ = n'\ + (n - 1)'\ + ... + (n - k + 1)'\)

1 1 1 Var(X(t) = k) = n2,\2 + (n _ 1)2,\2 + ... + (n _ k + 1)2,\2·

6.4

V · . 2(>,-,,)t ,\ + J.I. [1 e-(>'-/J.)t] anance = ze .... -- - . ,\-J.I.

6.5

{I (,\ < J.L)

(iv) lim PlO(t) = r (,\ > J.L) t--+oo 0 (,\ = J.L).

Page 18: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

258 APPENDIX B. ANSWERS TO SELECTED P8,OBLEMS

6.8

6.9

(i) M'(t) = a + (A - p,)M(t), M(O) = i.

{ ie(.~-I')t + I'~~ (A I: p,)

(ii) M(t) = at + i (A == p,).

{ I'~~ (A < p,)

(iii) lim M(t) = t--+oo

00 (A ~ p,),

6.10

a(a + A)(a + 2A) , " [a + (j - I)A] (J' = 1,2", ,), Pi = j! p.i Po

= [1 ~ a(a + A)(a + 2A) , " [a + (j - I)A]]_l Po +~ " .

;=1 J, p.l

6.11

(i) ap~,s) +p,(S_I)ap~:,s) =A(s-I)P(t,s),

00 [..\ (1 - e-I't)]i (iii) P(t ) =" I' -t(1-e-I'I). j ,s ~ "e s,

j=O J,

[ ~ (1 e-I't)]i n (t) ;; - -4(I-e-I'I) ( 1 2 ) rOj = " e I' j = 0, , , ....

J,

( ~ . ( ') }' n (t) ;;}' _4 (P . d' 'b ' ) IV 1m rOj = -,-e I' Olsson lStn utlOn ,

t--+oo J!

6.12

1 (A)i _4 Pi = j! ~ e I' (j = 0,1,2", .), Poisson distribution,

Page 19: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

6.13

6.14

[-2A

(i) P/(t) = P(t)A = pet) ~

2

[ " + Ae-(~+#')t] (ii) R (t) ,-

00 = A+f.t '

2A -(A + f.t)

2f.t o 1 A "

-2f.t

lJi. + Ae-(~+#')t] [A - Ae-(~+#')t] P01 (t) =2 , \+' , + f.t 1\ f.t

[ A - Ae-(H#')t1 2 P02 (t) = ---.,.\--­

I\+f.t

(iii) lim Poo(t) = (~) 2 , t-+oo 1\ + f.t

lim POl(t) = 2 (~) (~), t-+oo 1\ + f.t 1\ + f.t

lim P02 (t) = (~)2 t-+oo 1\ + J.t

(1"1") Po -_ (1 + 2A + 2A2)-1 2A f.t f.t2 ,PI = /iPo,

Problems 7

7.1

(i) Q(oo) = [Oi4 0"6] o "

(ii) 5f.t 3A Po = 3A+5f.t' PI = 3A + 5f.t "

7.2

(ii) f.t A Po= A+f.t' PI =--"

A+f.t

259

Page 20: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

260

7.3

7.5

APPENDIX B. ANSWERS TO SELECTED PROBLEMS

(ii) 11"0 = 11"1 = 11"2 = 1/3.

(iii) Po = (1 + ~ + ~) -1 ,

J.Ll J.L2

(ii) M*(s) = 1 _ F*tS) G*(S) l G*~~fs;(S) F*7s;~~(s)]' (iii) G~(s) = Gil(S) = G*(s)F*(s),

G~l(S) = G*(s), Gio(s) = F*(s),

P*(t) = [Poo(S) Pi1(S)] Pio(s) Pll(s)

1 [1 -G*(s) = 1 - F*(s)G*(s) F*(s)[l - G*(s)]

G*(s)[l - F*(S)]] 1- F*(s) .

Problems 8

8.2

8.3

(i)

(ii)

P{min(X1, X 2 ,'" Xn) > t} = F 1(t)F2 (t) .. · F n(t)

= exp [- t ri(t)] . • =1

1 1 E[X] = -n-' Var(X) = )2' EAi (~Ai

Page 21: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

8.6

8.7

261

{ ~ [e-A(HZ) _ e-(A+I')t] + ~ [e-AZ - e-A(HZ)] (A :f J.t)

R1(x,t) = Ate-A(Hz) + le-AZ -le-A(t+z) (A = J.t).

lim Ro(x,t) = ~e-AZ, t-+oo 1\ + J.t

Problems 9

9.1

9.2

9.5

(i)

(ii)

(iii)

(iv)

(i)

(ii)

(iii) (iv)

(i)

(ii)

(iii)

P{The attendant is busy} = p = 3/4.

p2 1 Lq = -- = 2-4 persons.

1-p

Wq = P = ~ hour = 3 min. J.t(1- p) 20

P { U > 115} = ~e-l = 0.276.

P{The secretary is busy} = p = 0.8.

L=~ =4jobs. -p W = L/ A = 40 min.

P{U + V ~ 40} = 1 - e-1 = 0.632.

P{The dentist is busy} = 1 - Po = 0.6346.

L = 1.423 patients.

Wq = 23.64 min.

Page 22: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

262 APPENDIX B. ANSWERS TO SELECTED PROBLEMS

9.8

A U = - = 1.0, 2.0 (0.1)

f..£

U P4 1.0 0.0154 1.1 0.0201 1.2 0.0262 1.3 0.0324 1.4 0.0396 1.5 0.0479 1.6 0.0563 1.7 0.0658 1.8 0.0747 1.9 0.0848 2.0 0.0950

A U = - ~ 1.5 ===> P4 = 0.0479 < 0.05 (5%).

f..£

Page 23: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Appendix C

The Bibliography

[1] A. O. Allen, Probability, Statistics, and Queueing Theory with Computer Science Applications, Academic Press, New York, 1978.

[2] L. J. Boon, Statistical Analysis of Reliability and Life-Testing Models: Theory and Methods, Marcel Dekker, New York, 1978.

[3] R. E. Barlow and F. Proschan, Mathematical Theory of Reliability, Wiley, New York, 1965.

[4] R. E. Barlow and F. Proschan, Statistical Theory of Reliability and Life Testing: Probability Models, Holt, Rinehart and Winston, New York, 1975.

[5] U. N. Bhat, Elements of Applied Stochastic Processes, Wiley, New York, 1972.

[6] K. L. Chung, Elementary Probability Theory with Stochastic Processes, Spr­inger-Verlag, New York, 1974.

[7] E. Qinlar, Introduction to Stochastic Processes, Prentice-Hall, Englewood Cliffs, New Jersey, 1975.

[8] D. R. Cox, Renewal Theory, Methuen, London, 1962.

[9] W. Feller, An Introduction to Probability Theory and Its Applications, Vol. I, 3rd ed., Wiley, New York, 1968.

[10] W. Feller, An Introduction to Probability Theory and Its Applications, Vol. II, 2nd ed., Wiley, New York, 1971.

[11] B. V. Gnedenko, Y. K. Belyayev, and A. D. Solovyev, Mathematical Meth­ods of Reliability Theory, Academic Press, New York, 1969.

[12] P. G. Hoel, S. C. Port and C. J. Stone, Introduction to Stochastic Processes, Houghton Mifflin, New York, 1971.

Page 24: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

264 APPENDIX C. THE BIBLIOGRAPHY

[13] S. Karlin and H. M. Taylor, A First Course in Stochastic Processes, 2nd ed., Academic Press, New York, 1975.

[14] S. Karlin and H. M. Taylor, A Second Course in Stochastic Processes, Ac&­demic Press, New York, 1981.

[15] J. G. Kemeny and J. L. Snell, Finite Markov Chains, Springer-Verlag, New York,1976.

[16] L. Kleinrock, Queueing Systems, Vol. I: Theory, Wiley, New York, 1975.

[17] L. Kleinrock, Queueing Systems, Vol. II: Computer Applications, Wiley, New York, 1976.

[18] J. Kohlas, Stochastic Methods of Operations Research, Cambridge Univer­sity Press, Cambridge, 1982.

[19] R. G. Laha and V. K. Rohatgi, Probability Theory, Wiley, New York, 1979.

[20] N. R. Mann, R. E. Schafer, and N. D. Singpurwalla, Methods for Statistical Analysis of Reliability and Life Data, Wiley, New York, 1974.

[21] S. Osaki, Stochastic System Reliability Modeling, World Scientific, Sing&­pore, 1985.

[22] E. Parzen, Modem Probability Theory and Its Applications, Wiley, New York, 1960.

[23] S. M. Ross, Applied Probability Models with Optimization Applications, Holden-Day, San Francisco, 1970.

[24] S. M. Ross, Stochastic Processes, Wiley, New York, 1983.

[25] H. M. Taylor and S. Karlin, An Introduction to Stochastic Modeling, AC8r demic Press, Orlando, 1984.

[26] W. A. Thompson, Jr., Point Process Models with Applications to Safety and Reliability, Chapman and Hall, New York, 1988.

[27] K. S. Trivedi, Probability & Statistics with Queueing, and Computer Science Applications, Prentice-Hall, Englewood Cliffs, New Jersey, 1982.

Page 25: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

Index

A Abelian theorem, ................ 242 Actual arrival rate, ......... 226, 228 Age replacement models, .... 200-203 Age, .......................... 73,98 Arrival rate, ..................... 217 Arriving time, .................... 68 Availability, ................. 194-195

average, ....................... 194 joint, .......................... 195 limiting, ....................... 194 limiting average, ............... 194

Availability theory, .......... 192-199

B Elath-tub curve, ............. 187-188 Elayes'theorem, ................... 12 Elernoulli distribution, ......... 34-35 Eleta distribution, .............. 47-48 Elinomial coefficient, .............. 13 Elinomial distribution, . 35-36, 56, 245 Elirth and death process, ..... 143-155

definition, ..................... 144 linear growth process, ......... 148 linear growth with immigration, 163

Elirthday problem, ................ 17 Elivariate exponential distribution, 53 Elivariate joint distribution, ....... 49 Elivariate normal distribution, ..... 52 Ellackwell's theorem, .............. 94 Ellock diagram, .................. 147 Ellock replacement models, ......... .

200, 203-207

C Central limit theorem, ............ 58 Channel, ........................ 215 Chapman-Kolmogorovequation, .... .

108,136 Characteristic function, ........... 31

Chebyshev's inequality, ........... 57 Coincidences

(See Matches) Combinations, .................... 14 Combinatorial analysis, ........... 13 Communication relation, ......... 112 Conditional density, ............... 50 Conditional distribution, .......... 50 Conditional expectation, .......... 50 Conditional probability mass function,

50 Conditional probability,

definition, ....................... 8 Convolution, ...................... 54 Correlation coefficient, ............ 52 Counting process, ................. 63 Covariance, ....................... 52 Cumulative hazard, .............. 187 Current life, ................... 73, 98 Customers, ...................... 215

D Degenerate distribution, ..... 218-219 Density, .......................... 29 Dependability, ................... 195 DFR, ............................ 187 Directly Riemann integrable, ...... 95 Distribution,

continuous, ..................... 40 definition, ...................... 29 discrete, ...................... :. 33

Doubly stochastic matrix, ........ 133 Down time, ...................... 193

E Elementary renewal theorem, ..... 92 Embedded Markov chain, ........ 168 Empty set

(See Null set) Equally likely, ..................... 6

Page 26: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

266

Equilibrium distribution, 98, 101, 246 Erlang's B formula, ......... 231, 234 Erlang's C formula, .............. 230 Erlang's loss formula

(See Erlang's B formula) Event, ............................. 1 Eventual transition probability, ..... .

114, 168 Excess life, .................... 73, 97 Expectation, ...................... 30 Expected cost rate, .............. 200 Expedited order, ................. 207 Exponential distribution, ........... .

42-44, 188, 245-246

F Failure rate, ................ 186, 233 Failure, .......................... 186

random, ....................... 186 First passage probability, ........ 114 Furry process, ................... 139

G Gamma distribution, ............... .

44-45, 56, 188, 246 Gamma function, ................. 44

Geometric distribution, ............ . 36-37, 191-192, 245

H Hazard function, ................. 187 Hazard rate,

(See Failure rate) Holding time, .................... 216 Hypergeometric distribution, ...... 21

I IFR, ............................. 187 Impossible set

(See Null set) Inclusion and exclusion formula

(See Poincare's theorem) Independent event

(See Mutually independent event) Independent increments, .......... 65 Independent trials, ................ 10 Indicator, ......................... 34 Infant phase, .................... 187 Infinitestimal generator, ..... 154, 157

INDEX

Initial distribution, .............. 107 Initial probability, ............... 107 Intensity function, ................ 77 Interarrival time, ............. 68,217 Interoccurrence times

(See Interarrival times) Inventory cost ................... 207 Item, ............................ 186

J Joint density, ..................... 49

Joint probability mass function, ... 49

K k-Erlang distribution, ........ 217-218 k-out-of-n system, ............... 212

Kendall's notation, .............. 220 Key renewal theorem, .......... 95-97 Kolmogorov's forward equation, .....

138, 142, 146, 153, 155 Kolmogorov's backward equation, ...

154, 155 Kurtosis, ......................... 32

L Laplace transform, ............... 241 Laplace-Stieltjes transforms, ........ .

32, 172, 241 Lifetime, ........................ 186 Little's formulas, ....... 225, 228, 236 Lognormal distribution, .. 47, 190-191

M Maintainability, .................. 192 Maintenance, .................... 192

corrective, ..................... 192 preventive, .................... 192 scheduled, ..................... 193

Marginal density, ................. 50 Marginal distribution, ............. 49 Marginal probability mass function, 49

Page 27: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

INDEX

Markov chain, absorbing, ..................... 116 continuous-time, ............... 135 continuous-time finite-state, ... 155 discrete-time, .................. 105 ergodic, ....................... 119 finite-state, ................ 123-128 irreducible .................... 113 spatially homogeneous, ........ 111

Markov property, ................ 105 Markov renewal function, ........ 170 Markov renewal process,

alternating, ............... 17~181 definition, ..................... 167 stationary, .................... 177

Markov's inequality, .............. 56 Mass function, ................... 167 Matches, ......................... 19 Maximum queueing system capacity, .

219 MDT, ........................... 194 Mean

(See Expectation) Mean recurrence time, ........... 118 Mean value function, .............. 78 Moment generating function, ...... 32 MTBF, .......................... 194 MTBM, ......................... 194 Multinomial coefficient, ........... 20 Multinomial expansion, ........... 21 Multiplication principle, .......... 14 Multivariate distributions, ..... 48-56 MUT, ........................... 194 Mutually independent event, ... 9, 10

N ~fold (Stieltjes) convolution, .. 54, 84 ~step distribution, .............. 109 Negative binomial distribution, ..... .

37-39, 192, 245 Normal distribution, ...... 46-47, 190

standardized, ................... 47 nth moment, ...................... 31

about the mean, ................ 31 about the origin, ............... 31

Null set, ........................... 3 Number of service channels, ...... 219

267

o One-step transition probability, .. 167

Order statistics, ................. 212

Ordering models, ............ 207-211

p

Parallel system, .................. 212

Partition, ......................... 11

Pascal distribution

(See Negative binomial distribution)

Permutation, ..................... 14

Poincare's theorem, ............... 17

Point

(See Sample point)

Poisson arrivals, ................. 217

Poisson distribution, ... 39-40, 67, 192

Poisson process, ............... 65-68

decomposition, ................. 75

definition, ............... 66, 67, 69

nonhomogeneous, ............ 77-79

nonstationary, ............... 77-79

superposition, .................. 75

Population, .................. 216-217

finite, ................ 217, 233-238

infinite, ....................... 216

Probability distribution

(See Distribution)

Probability mass function,

definition, ...................... 29

Probability,

definition, ....................... 5

Pure birth process, .......... 136-141

definition, ..................... 137

Pure death process, .......... 141-143

with linear death rate, ......... 142

Page 28: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

268

Q Queueing discipline, ......... 219-220 Queueing models, ............ 215-220

Infinite server Poisson queue, ... 72 M/M/1, ....................... 150 M/M/1/K/K, ............. 233-234 M/M/1/N, ........... 152,226-228 M/M/1/oo, ............... 221-226 M/M/1/oo queue with impatient cus-

tomers, ..................... . 164

M/M/2/2/2, .............. 247-248 M/M/c/c, ................. 230-231 M/M/c/c/c, ............... 237-238 M/M/c/K/K, ............. 235-236 M/M/c/oo, ............... 229-230 M/M/oo, ...................... 151 M/M/oo/oo, .............. 231-232 multiple server, ....... 219, 228-232 single server, ......... 219, 221-228

Queueing models with finite population, 233-238

R Random trial, ...................... 1 Random variable,

arithmetic, ..................... 94 continuous, ..................... 29 definition, ...................... 29 discrete, ........................ 29 independent, ................... 49 lattice, ......................... 94 standardized, ................... 31

Random walk, on a circle, .................... 133 one-dimensional symmetric, .... 116 two-dimensional symmetric, ... 131 with reflecting barrier, .... 121, 134

Randomization, .................. 159 Rayleigh distribution, ............. 46 Regeneration point, .............. 169 Regular order, ................... 207 Reliability, .................. 193-194

interval, ....................... 194 limiting interval, .............. 194

Reliability function, (See Reliability)

Reliability models, ........... 185-199

INDEX

Renewal density, .................. 95 Renewal equation, ........... 88, 104

in matrix form, ................ 170 Renewal function, ............ 85, 249 Renewal process,

definition, ...................... 83 delayed, ........................ 99 stationary, ................ 101-103

Renewal-type equation, ........... 96 Repair rate, ..................... 233 Repairman problems, ............ 233 Repeated trials

(See Independent trials) Replacement models, ........ 200-207 Residual life, .................. 73, 97

S Sample function, .................. 65 Sample path, ..................... 65 Sample point, ...................... 1 Sample space, ...................... 1 Sample with replacement, ......... 14 Sample without replacement, ...... 14 Sampling inspection, .............. 21 Second moment, .................. 30 Semi-Markov kernel, ............. 167 Semi-Markov process,

definition, ..................... 167 Series system, ................... 212 Server, .......................... 215 Service, .......................... 215 Service rate, ..................... 219 Service time, .................... 219 Shortage cost, ................... 207 Skewness, ......................... 32 Standard deviation, ............... 31 State classification, .......... 112-117

absorbing, ..................... 116 aperiodic, ..................... 113 non-null recurrent, ............ 118 null recurrent, ................. 118 period, ........................ 113 periodic, ...................... 113 positive recurrent, ............. 118 recurrent, ..................... 114 transient, ...................... 114

State space, ...................... 65 State transition diagram, ........ 106

Page 29: Appendix A Laplace-Stieltjes Transforms - Springer978-3-642-84681-6/1.pdf · Appendix A Laplace-Stieltjes Transforms ... from the corresponding Laplace transforms, ... ditions and

INDEX 269

Stationary distribution, .......... 119 Stationary increments, ............ 65 Stationary independent increments, 65 Stationary probabilities, ..... 176-177 Stieltjes convolution, .............. 54 Stieltjes integrals, ................. 30 Stirling's formula, ............... 115 Stochastic process, ................ 64

continuous-time, ................ 64 discrete-time, ................... 64

Strong law of large numbers, ... 57-58

T Tauberian theorem, .............. 242 Telephone line, .................. 216 Total event, ........................ 3 Total life, ......................... 73 Total probability formula, ......... 11 Traffic intensity, ................. 222 Transition probability matrix, .... 106 Transition probability, ...... 105, 135

n-step, ........................ 105 Truncated normal distribution, ... 190 Type I counter, .................. 182

U Unconditional distribution, ...... 170 Unconditional mean, ............. 170 Uniform distribution, 33, 71, 188, 191

continuous, .................. 40-42 discrete, ........................ 34 standard, ....................... 41

Uniformization, .................. 159 Up time, ........................ 193 Useful life phase, ................ 187 Utilization factor, ................ 223

V Variance, ......................... 30

W Waiting time, ..................... 68 Wearout phase, .................. 187 Weibull distribution, . 45-46, 190, 212

y Yule process, 139-140