Stochastic Processes and Their Applications 1st Edition Frank Beichelt (Author) All Chapters Instant Download

Visit https://ebookgate.
com to download the full version and

explore more ebooks
Stochastic Processes and Their Applications 1st

Edition Frank Beichelt (Author)
_____ Click the link below to download _____

https://ebookgate.com/product/stochastic-processes-and-
their-applications-1st-edition-frank-beichelt-author/
Explore and download more ebooks at ebookgate.com

Here are some recommended products that might interest you.
You can download now and explore!
Stochastic Methods and their Applications to

Communications Stochastic Differential Equations Approach
1st Edition Serguei Primak
https://ebookgate.com/product/stochastic-methods-and-their-
applications-to-communications-stochastic-differential-equations-
approach-1st-edition-serguei-primak/
ebookgate.com
Theory and Statistical Applications of Stochastic

Processes 1st Edition Yuliya Mishura
https://ebookgate.com/product/theory-and-statistical-applications-of-
stochastic-processes-1st-edition-yuliya-mishura/
ebookgate.com
Probability Theory and Stochastic Processes with

Applications 2009th Edition Oliver Knill
https://ebookgate.com/product/probability-theory-and-stochastic-
processes-with-applications-2009th-edition-oliver-knill/
ebookgate.com
Basic Stochastic Processes 1st Edition Devolder
https://ebookgate.com/product/basic-stochastic-processes-1st-edition-
devolder/
ebookgate.com
Shore Processes and their Palaeoenvironmental Applications
1st Edition Edward J. Anthony (Eds.)
https://ebookgate.com/product/shore-processes-and-their-
palaeoenvironmental-applications-1st-edition-edward-j-anthony-eds/
ebookgate.com
Large deviations for stochastic processes Jin Feng
https://ebookgate.com/product/large-deviations-for-stochastic-
processes-jin-feng/
ebookgate.com
Adventures in Stochastic Processes 1st Edition Sidney I.

Resnick (Auth.)
https://ebookgate.com/product/adventures-in-stochastic-processes-1st-
edition-sidney-i-resnick-auth/
ebookgate.com
Introduction to Stochastic Processes Second Edition

Gregory F. Lawler
https://ebookgate.com/product/introduction-to-stochastic-processes-
second-edition-gregory-f-lawler/
ebookgate.com
Discrete Stochastic Processes and Optimal Filtering Second

Edition Jean?Claude Bertein
https://ebookgate.com/product/discrete-stochastic-processes-and-
optimal-filtering-second-edition-jeanclaude-bertein/
ebookgate.com
Stochastic Processes and Their
Applications
Stochastic Processes and Their
Applications
Frank E. Beichelt and L. Paul Fatti
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business
Library of Congress Cataloging-in-Publication Data
Catalog record is available from the Library of Congress
First published 2002 by CRC Press
Taylor & Francis Group

6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2002 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
ISBN-13: 978-0-415-27232-2 (hbk)
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the
validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the
copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to
publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let
us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including
photocopying, microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www. copyright.com
(http:/ /www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers,
MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of
users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has
been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis W eh site at

http://www.taylorandfrancis.com
and the CRC Press Web site at

http:/ /www.crcpress.com
Titel de Originalausgabe: Beichelt, Frank: Stochastische Prozesse Fiir Ingcnicure, © B.G. Teubner, Stuttgart 1997.
Translation arranged with the approval of the publisher B.G. Tcubncr from the original German edition into English.
Contents
Preface ix
Symbols and Abbreviations xi
1 Probability Theory 1
1.1 Random Events and their Probabilities 1
1.2 Random Variables 4
1.2.1 Discrete Random Variables 4
1.2.2 Continuous Random Variables 6
1.2.3 Nonnegative Random Variables 8
1.3 Random Vectors 13
1.3.1 Two-Dimensional Random Vectors 13
1.3.2 «-Dimensional Random Vectors 20
1.4 Sums and Sequences of Random Variables 22
1.5 Transformations of Probability Distributions 31
1.5.1 z-Transformation 32
1.5.2 Laplace-Transformation 34
Exercises 37
2 Stochastic Processes 43
2.1 Introduction 43
2.2 Characteristics of Stochastic Processes 47
2.3 Properties of Stochastic Processes 49
2.4 Special StochasticProcesses 53
2.4.1 Continuous-Time Stochastic Processes 53
2.4.2 Stationary Discrete-Time Stochastic Processes 60
Exercises 67
3 Poisson Processes 69
3.1 Homogeneous Poisson Process 69
3.1.1 Definition and Properties 69
3.1.2 Poisson Process and Uniform Distribution 77
3.2 Inhomogeneous Poisson Process 85
3.2.1 Definition and Properties 85
3.2.2 Minimal Repair 89
Exercises 97
v
vi CONTENTS
4 Renewal Processes 99
4.1 Foundations 99
4.2 Renewal Function 102
4.2.1 Renewal Equations 102
4.2.2 Bounds for the Renewal Function 107
4.3 Recurrence Times 111
4.4 Asymptotic Behaviour 114
4.5 Stationary Renewal Processes 118
4.6 Alternating Renewal Processes 120
4.7 Cumulative Stochastic Processes 125
4.8 Regenerative Stochastic Processes 131
Exercises 134
5 Discrete-Time Markov Chains 137

5.1 Foundations and Examples 137
5.2 Classification of States 145
5.2.1 Closed Sets of States 145
5.2.2 Equivalence Classes 147
5.2.3 Periodicity 150
5.2.4 Recurrence and Transience 152
5.3 Limit Theorems and Stationary Distribution 158
5.4 Birth· and Death Processes 163
Exercises 165
6 Continuous-Time Markov Chains 169

6.1 Foundations 169
6.2 Kolmogorov’s Differential Equations 173
6.3 Stationary State Probabilities 183
6.4 Construction of Markovian Systems 187
6.5 Erlang’s Phase Method 189
6.6 Birth- and Death Processes 191
6.6.1 Time-Dependent State Probabilities 191
6.6.2 Stationary State Probabilities 202
6.7 Sojourn Times 207
6.8 Applications in Queueing Theory 209
6.8.1 Introduction 209
6.8.2 Loss Systems 212
6.8.3 Waiting Systems 216
6.8.4 Waiting-Loss-Systems 219
6.8.5 Special Single-Server Queueing Systems 223
6.8.6 Networks of Queueing Systems 229
6.9 Semi-Markov Processes 241
Exercises 250
7 Wiener Processes 257

7.1 Definition and Properties 257
CONTENTS vii
7.2 First Passage Times 263

7.3 Transformations of the Wiener Process 269
7.3.1 Elementary Transformations 269
7.3.2 Omstein-Uhlenbeck Process 271
7.3.3 Wiener-Process with Drift 272
7.3.4 Integral Transformations 285
Exercises 289
8 Spectral Analysis of Stationary Processes 293

8.1 Foundations 293
8.2 Processes with Discrete Spectrum 295
8.3 Processes with Continuous Spectrum 299
8.3.1 Spectral Representation of the Covariance Function 299
8.3.2 Spectral Representation of Stationary Processes 308
Exercises 310
Appendix 1 Landau O rder Symbol 312

Appendix 2 Dirac Delta Function 313
Answers to Selected Exercises 314
References 319
Index 323
Preface
This book is an introduction to stochastic processes and their applications for students in
engineering, industrial statistics, science, operations research, business and public policy
analysis, and finance. It provides theoretical foundations for modeling time-dependent
random phenomena as they occur in physics, electronics, computer science and biology.
Likewise, it takes into account applications in the field of operations research, in
particular in queueing, maintenance and reliability theory as well as in financial markets.
Through numerous, mostly science and engineering-based examples, the subject is
represented in a comprehensible, practically oriented way. Hence the book is also
suitable for self-study. As a non-measure theoretic introduction to stochastic processes,
its study only requires a knowledge of calculus and elementary probability theory which
should be familiar to students after having finished their undergraduate studies in
mathematics and statistics. However, to make the book attractive to mathematically
interested readers as well, some important proofs and theoretically challenging examples
and exercises have been included. Solutions to most of the exercises can be found in an
appendix or the exercises are given together with their solutions. Those sections,
examples or exercises marked with a * symbol are either theoretically more difficult or
of less practical impor- tance. The chapters are organized in such a way that reading a
chapter usually requires a knowledge of the previous ones.
A mathematically rigorous treatment of Wiener processes (Brownian motion processes)
and the spectral analysis of stationary processes, dealt with in chapters 7 and 8,
respectively, is not possible without a basic knowledge of measure theory, Fourier
analysis, and the theory of generalized functions. Therefore, these chapters sometimes
present heuristically motivated explanations and formulas instead of mathematically exact
concepts and derivations. This approach does not detract from the main purpose of this
book, which is to enable readers to apply stochastic modeling in their own field. This
book generally does not deal with the data analysis aspects of stochastic processes. It
can be anticipated that readers will use statistical software for tackling numerical
problems. However, after having studied our book they will be able to work more
creatively with such software and write their own analysis programs. The text may also
serve as a basis for preparing senior undergraduate and graduate level courses.
The authors wish to thank their Honours students of the year 2000, who checked and
tested the exercises and gave many valuable hints. In particular, we acknowledge the
contributions of H. Christoforou, S. Knight, T. Levin, V. Nkwambi and D. Tugendhaft.
Finally, the authors would like to thank Ms. Frances Horrocks and Mrs. Janie Wardle of
Taylor & Francis for their committed, constructive cooperation.
IX
X
The book is a thoroughly checked word for word translation of the German original
published by B. G. Teubner Stuttgart in 1997 under the title "Stochastische Prozesse fur
Ingenieure". The numerous suggestions of German-speaking readers helped very much
to prepare this English edition. Many thanks for their support as well. In this second
reprint, some corrections have been made.
Further helpful comments on this book are very welcome and should be sent to:
University of the Witwatersrand, School of Statistics and Actuarial Science, WITS 2050,
Johannesburg, Republic of South Africa. E-mail: [email protected].
Johannesburg, September 2003 Frank E. Beichelt, L. Paul Fatti

Symbols and Abbreviations
□ ■ · symbols after an example, a theorem, a definition
/{ /)s c f t ) = c for all t e T
f* g convolution of two functions / and g
/*(«) n-th convolution power of /
Hs), L{f) Laplace-transform of a function /
Φ) Landau order symbol
Φ) Dirac delta function
Probability Theory
X Y Z random variables
E(X), Var{X) expected (mean) value, variance of X
fxto> Ρ Φ ) probability density function, (cumulative probability) distri
bution function o iX
λ(χ), Λ(χ) failure rate, integrated failure rate (hazard function)
Ν(μ,σ2) normal random variable with expected value μ and variance a 2
φ(χ), Φ(χ) probability density function, distribution function of a standard
normal random variable
fx(x l>x2· —>*«) joint probability density function of X = (X \,X 2...... Xn)
F x(x i ,x2, ... ,xn) joint distribution function of X = (A'j.A^,... ,Xn)
Cov(X, Y) covariance between X and Y
p(X, Y) correlation coefficient of X and Y
M(z) z-transform (moment generating function) of a discrete random
variable or its probability distribution
Stochastic Processes
{AXO, t e T}, {A',, i e T } continuous-time, discrete-time stochastic process
with parameter space T
Z state space of a stochastic process
ft(x), Ft(x) probability density, distribution function of X(t)
ftut2,...,tn(x l>x2>··· >*«)> Ft,,t2>. ^ ( x l>x2>··· >x") j ° int probability density
function, distribution function of (X(tf),X(/2) , ... ,X(tn))
m{t) trend function of a stochastic process
C{s,t) covariance function of a stochastic process
C(t) covariance function o f a stationary process
p(s,t) correlation function of a stochastic process
XI
XU SYMBOLS AND ABBREVIATIONS
{M(t), f > 0} point process, counting process

{T \ , T2 , ...} pulse process
H(t), H\(t) renewal function of an ordinary, delayed renewal processs
R(t), V(t) backward, forward recurrence time
A(t), A point, stationary (long-run) availability
L(x) first-passage time of level x
Pij, p)j ’ one-step, m-step transition probabilities of a homogeneous,
discrete-time Markov chain
P transition matrix
first-passage probability; probability that, starting from state i,
the first transition into state j occurs at time t —n
fij probability that, starting in state i, the process ever makes a
transition into state j
Pij(t) transition probabilities of homogeneous, continuous-time
Markov chain
<lij, <fo conditional, unconditional transition rates (transition intensities)
of a homogeneous, continuous-time Markov chain
{ π ,; i e Z } stationary state distribution of a homogeneous Markov chain
Xj , \lj birth, death rates
λ, μ arrival rate, service rate (section 6.8)
μ, expected sojourn time of a semi-Markov process in state i
μ drift parameter of a Wiener process with drift
p correlation coefficient; traffic intensity (section 6.8)
η degree of server utilisation
W waiting time in a queueing system
L random lifetime, queue length
L(a,b) first-passage time of level min(a, b)
{W(t), t> 0 \ Wiener-process with drift
(Z(/), t > 0} white noise
ω circular frequency
φ, Φ phase, random phase
.ϊ(ω), 5(ω) spectral density, spectral function
S spectrum S = {ω; ί(ω) > 0}
w band width, w = sup ω - inf ω
<eeS oeS
1 Probability Theory
1.1 Random Events and Their Probabilities
Random events occur in connection with random experiments. A random experi
ment is characterized by two properties:
1) repetitions of one and the same experiment, even if carried out under identical
conditions, generally have different outcomes, and
2) the set of all possible outcomes of the experiment is known.
Hence, the outcome of a single random experiment may not be predicted with cer
tainty. Thus, it only makes sense to carry out random experiments if they are re
peated sufficiently frequently under identical conditions, in order to identify and
quantify stochastic (statistical) regularities. Examples of random experiments are:
1) Tossing a coin. The possible outcomes are "Heads”and "Tails”.
2) Tossing a die. The possible outcomes are the numbers 1, 2 ,..., 6.
3) Counting the number of vehicles arriving at a filling station per day.
4) Counting the number of e-mails arriving in a firm per day.
5) Counting the number of faults occuring per unit time in a computer network.
The possible outcomes are, as in the previous two random experiments, 0, 1 ,....
6) Recording the life lengths of electronic parts.
7) Recording the maximum fluctuation of stock prices per unit time. The possible
outcomes are, as in the random experiment 6, any real numbers in [0, oo).
The following concepts refer to the same random experiment. A possible outcome
a of the experiment is called an elementary (simple) event. The set of all elementa
ry events is called space o f elementary events or sample space. Here and in what
follows the sample space is denoted by M. A sample space is discrete if it is a fini
te or countably infinite su. A random event (briefly: event) A is a subset of M. An
event A is said to have occured if the outcome a of the random experiment is an
element of A, i.e., if a e A . Let A and B be two events. Then the set-theoretic
operations intersection 'W an d union "U" can be interpreted in the following way:
A r\B is the event that both A and B occur and A KJB is the event that A or B (or
both) occur. If A ε B, i.e. if A is a subset of B, then the occurence of A implies the
occurence of B. A\ B is the set of all those elementary events which are elements
of A, but not of B. Hence, A \ B is the event that A occurs but not B. The event
1
2 1 PROBABILITY THEORY
A = Μ \Λ is the complement o f A. Thus, if A occurs, then not A and vice versa. Let
A i ,Α ϊ , ...,An be a sequence of events. Then the rules of de Morgan hold:
( 1. 1)
In particular, if n = 2, A[ - A and A 2 = B,
( 1.2)
The empty set 0 is the impossible event, since, not containing any elementary
events, it can never occur. By definition, M contains all elementary events so that
it must always occur. Hence M is called the certain event. Two events A and B are
called mutually exclusive if their joint occurence is impossible, i.e. if A r\B = 0 .
In this case the occurence of A implies that B does not occur and vice versa. In
particular, A and A are mutually exclusive.
Let “
Mi be the set of all events which can occur when carrying out the random ex
periment. Further, let P = P(·) be a function on ")tt with the following properties:
I) P (0 ) = 0, P (M )= 1.
II) For any event A, 0 < P(A) < 1.
III) For any sequence of pairwise mutually exclusive events
(1.3)
The number P(A) is the probability of the event A. P(A) characterizes the degree
of certainty of the occurence of A. This interpretation of the probability is justified
by the following implications of the the properties I) to III).
1)
2) If
3) If A and B are mutually exclusive events, i.e. if A η B = 0 , then
4) For any events A and B,
Hint: It is assumed that all events which arise from applying the operations r>, U, £ and
\ to any subsets of 1ft are also elements of "ffi.
1.1 RANDOM EVENTS AND THEIR PROBABILITIES 3
The probabilites of random events are usually unknown. However, they can be
estimated by their relative frequencies. If, in a series of n repetitions of the same
random experiment the event A has been observed m = m(A) times, then the
relative frequency of A is given by
Generally, the relative frequency of A tends to P(A) as n increases:
Therefore, the probability of A can be estimated with any required level of accura
cy from its relative frequency by sufficient repetitions of the random experiment.
Two events A and B are called independent if
0 .4 )
The events A i,A 2 ,...,An are independent if for any subset ,Λ,-^jof
In particular, if k = n, then the independence of the A/ implies
(1.5)
Let Λ and B be two events with P(B) > 0. Then the conditional probability of A gi
ven B is
0 .6)
is called an exhaustive set of events if
Let {A \,A 2 ,...,An} be an exhaustive and mutually exclusive set of events. Then
the formula
(1.7)
is called total probability rule and the formula
is called Bayes' theorem.

1.2 R andom V ariables

Given a random experiment with sample space M, a random variable X is a real
function on Μ: X =X(a), a e M . Thus, a random variable associates a number
with each elementary event, i.e. with each outcome of the random experiment.
Any value which X can assume is called a realization of X. The set o f all realiza
tions of X is called the range of X. Because it is random which elementary event
will be outcome of the random experiment, the resulting realization of X is also
random. By introducing a random variable X one passes from the sample space M
of a random experiment to the range {X(a), a e M} of X, which is a set of real
numbers. This transition makes sense if the elementary events are not real numbers
so that a direct quantitative evaluation of the random experiment is not possible.
However, in many cases the elementary events are themselves real numbers, in
which case they can be considered to be realizations of a random variable. Hence a
random variable may be interpreted as the (random) outcome of a random experi
ment the elementary events of which are real numbers.
A random variable X is completely characterized by its (cumulative) distribution
fiinction F(x):
Thus, F(x) is the probability of the random event that X assumes a realization
which is less than or equal to x. For a <b,
( 1 -8 )
Any distribution function F(x) has the following properties:

1) F(~co) = 0, F(+co) = 1.
2) F(x) is nondecreasing in x.
Conversely, every function F(x) satisfying these two properties can be considered
to be the distribution function of a random variable X. The distribution function of
a random variable characterizes its probability distribution.
1.2.1 Discrete Random Variables
The range of a discrete random variable is a finite or countably infinite set. Exam
ples of discrete random variables were given in section 1.1 (examples 2 to 5).
Let {xo, X\, *2> "·} be the range o f X, x,· < x* for i < k. Further, letp, be the pro
bability of the random event that X assumes the realization x,·:
1.2 RANDOM VARIABLES 5
The set {Po>Pi>-··} characterizes the probability distribution of A. Since A must

assume one of its realizations,
Conversely, any sequence of nonnegative numbers [pq,p | ,...} satisfying this con
dition can be considered to be the probability distribution of a discrete random va
riable.
The distribution function of A is
If the range of X is finite and if xn is the greatest realization of X, then this defini
tion has to be supplemented by
F(x) is a piecewise constant function with jumps of size p, at x = x, - 0. There

fore,
Thus, given {Po.Pi. ···} the distribution function of A-can be constructed and, vice
versa, given the distribution function of X, the probabilities p, = P(X = x,) can be
obtained.
The expected value (mean value) E(X) and the variance Var(X) of X are given by
and
respectively, provided the sums exists. The arithmetic mean of n realizations of X

obtained from n independent repetitions of the random experiment tends to E(X)
as n tends to infinity. The variance VariX) is the expected squared deviation of X
from its expected value E(X).
In particular, let A be a binary random variable with range {0,1} and probability
distribution {P(X= 1) = p , P(X= 0) = 1 - p }. Then,
(1.9)
The n th moment μ« of A is the expected value of A” :

Table 1.1 Probability distributions of discrete random variables X
Distribution Range o f X
uniform
distribution
geometric
distribution
binomial
distribution
negative bino
mial distribution
Poisson
distribution
1.2.2 Continuous Random Variables
A random variable X with distribution function F(x) is called continuous if F(x)

has a first derivative
Rx) is called probability densityfimction of X (briefly: probability density or only

density). According to its definition,
In particular,
Conversely, every nonnegative function f(x) satisfying this condition can be con
sidered to be the probability density of a random variable X. As with its distribu
tion function, a continuous random variable is also completely characterized by its
probability density.
Expected value (mean value) E(X) and variance Var(X) of X are defined by
provided the integrals exist.

The n th moment of X is
provided the integrals exist. A useful representation of Var(X) is

( 1. 10)
This relationship also holds for discrete random variables. For a continuous ran
dom variable Ar the probability (1.8) can be written in the following form:
The range of X coincides with the set of all those x for which f(x) > 0.
Table 1.2 Probability distributions of continuous random variables X
Distribution Range o f X
Uniform distrib
ution over [c,d]
Exponential
distribution
Gamma
distribution
Beta distribution
in [0,1]
Erlang
distribution
Normal (Gauss-)
distribution
Comment The densities have the given functional forms over the ranges specified in the
second column. Elsewhere they are identically zero. The Gamma function Γ(χ) and Beta
function B(x,y) are defined by
1.2.3 Nonnegative Random Variables
If X is a nonnegative discrete random variable in the range {0, 1,...} with proba
bility distribution (p,· = P(X =i); i = 0,1,...}, then its expected value
can be written in the form
( 1. 11)
In the remainder of this section, X is assumed to be a continuous nonnegative ran

dom variable with distribution function F(x) and density /(x). Since both functions
F(x) and f[x) are equal to 0 for x < 0, it is sufficient to define them for x > 0. An
analogous formula to (1.11) for E(X) given that E(X) < oo can be obtained by par
tial integration:
Hence,
( 1. 12)
An important example of a nonnegative random variable is the lifetime of techni

cal systems, i.e. the time span to its first failure. (In this context, a failure is assu
med to occur instantanously.) The terminology used in what follows refers to this
example. Consequently, F(x) is called failure probability and
survival probability because F(x) and F(x) are the respective probabilities that the
system does or does not fail in [0, x\.
Let us now consider the distribution function of the residual lifetime of a system
which has already worked for t time units without failing (Figure 1). This condi
tional failure probability will be denoted by F<(x):
Hence, according to (1.6),

Figure 1.1 Illustration of the residual lifetime
Formula (1.8) then yields the desired result:
(1.13)
The corresponding conditional survival probability is given by
(1.14)
Example 1.1 (uniform distribution) Let the random variable X be uniformly dis
tributed over [0, T]. Then its density and distribution function are (see Table 1.2)
elsewhere,
Hence, the conditional failure probability is
Thus, the residual lifetime after the time point t is uniformly distributed over the
interval [0, T -t].
Figure 1.2 Density and distribution function of a random variable being uniformly
distributed over [0, T]
Obviously, the conditional failure probability is increasing with increasing t. Spe

cifically, for T = 100 the system fails in the interval (SO, 60] with probability 1/10.
However, if it is known that the system has not failed in the interval [0,40], then it
fails in the interval [SO, 60] with probability
Example 1.2 (exponential distribution) Let the random lifetime X of a system be

exponentially distributed with parameter λ, i.e. its density and distribution func
tion are (see Table 1.2)
Given t, the corresponding conditional failure probability is
(1.15)
Thus, the residual lifetime of the system has the same distribution function as the
lifetime of a new system: it is exponentially distributed with parameter λ. The ex
ponential distribution is die only continuous probability distribution which has this
so-called memoryless property or lack of memory property. Consequently, the age
of a system with exponential lifetime has no influence on its future failure beha
viour. Or, equivalently, if the system has not failed in the interval [0, /], then, with
respect to its failure behaviour in [t, oo), it is "as good as new". Electronic hardwa
re often has this property after the "early failure time period".
The relationship (1.15) can also be written in the form
(1.16)
It can be shown that the exponential distribution is the only one which satisfies
(116). □
The practical background of the conditional failure probabilit motivates the follow
ing definition:
Definition 1.1 A system is aging in the interval [f j , /2], *i < / 2, if and only if for
an arbitrary, but fixed x and for increasing t, fj < f2 , the conditional failure
probability Ft(x) (conditional survival probability Ft(x)) is increasing (decrea
sing). ·
The following considerations provide another approach to modeling the aging be
haviour of systems: When the conditional failure probability Ft(At) of a system in
the interval [/, ί+ Δ /] is considered relative to the length At of this interval, one
obtains a conditional failure probability per unit time Ft(At)/At, that is, a "failure
probability rate". As At -> 0, this rate tends to a function λ(ί), which gives infor
mation on the instantaneous tendency of the system to fail:
(1.17)
Hence,
Integration on both sides of this relationship yields
The function λ(ί) is called failure rate. The integrated failure rate or the hazard
function is given by
By making use of A(x) we obtain

(1.18)
(1.19)
(1.18) implies an important property of the failure rate:
A system ages in the interval [t \ , tf\, t\ < tj, i f and only if itsfailure rate is
I increasing in this interval.
Example 1.3 ( WeibuU distribution) A random variable Λ' has a Weibull distribu
tion with parameters β and Θ if it has density
(Figure 1.3). Hence its distribution function is
The parameter Θ is a scale parameter. It is equal to 1 if Θ is the measurement unit

for x. Thus, Θ is unessential for characterizing the Weibull distribution. The struc
ture of the distribution function immediately yields the corresponding hazard func
tion:
By differentiation,
Consequently, a system with a Weibull-distributed lifetime ages if and only if β> 1.

Figure 1.3 Densities of the Weibull distribution
From (1.12), the expected value of X is
This result shows once more that Θ is a scale parameter. Special cases of the Wei
bull distribution are the exponential distribution (β = 1) and the Rayleigh distribu
tion (β = 2). Analogously to the exponential distribution, the distribution function
of a Weibull distributed random variable is sometimes written in the form
Thus, λ = (1/Θ)Ρ. A disadvantage of this representation is that λ is not a scale pa

rameter. □
For many applications, the following property of the failure rate λ(χ) is important:
( 1.20)
o(h) is Landau's order symbol with respect to h -» 0, i.e. any function of h satis
fying
(see Appendix 1). Therefore, for Δχ sufficiently small, λ(χ)Δχ is approximately the
probability of a system failure in (x, χ+Δχ] provided that it has operated in [0, x]
without failing. This property of the failure rate can be used for statistical estima
tion: At time t = 0 a specified number of systems with independent, identically
distributed lifetimes start working. Then the failure rate of these systems in the in
terval (x. χ+Δχ] is approximately equal to the number of systems having failed in
(x, χ+Δχ] divided by the number of systems which are still operating at time x.
1.3 RANDOM VECTORS 13
1.3 Random Vectors

Outcomes of random experiments frequently have to be described by more than
one random variable to obtain the required information. In these cases, the out
comes are vectors (X \,X 2,...,Xn), the components X{ of which are random vari
ables. (X],X 2 , -..,Xn) is called a random vector or a multidimensional random vari-
iable. If, after having carried out the random experiment, A', = x,·; i = 1,2,..., λ;
then (X\,X2 ,■··,*«) is called a realization of (Χχ,Χι, -,Χη)· Generally, the proba
bilistic treatment of random vectors cannot be done by considering their compo
nents separately from each other, they have to be considered together. This requir
es the introduction of the joint or common probability distribution of random vec
tors. The joint probability distribution of (X\ ,X 2 ,...,Xn) contains both complete
information on the probability distributions of all the components AT,· and complete
information on their statistical dependence.
1.3.1 Two-Dimensional Random Vectors
This section considers random vectors with two components X\ =X and A^ = Y.

Random vectors with discrete and continuous random variables will be dealt with
separately from each other, although -as in case of one-dimensional random vari
ables- the important concepts and properties are usually identical.
Discrete component* Let X and Y be two discrete random variables with re
spective ranges } and {y$,y I, } and respective probability distributions
Furthermore, let
The set of probabilities {r^·; i,j = 0,1,...} is the joint or two-dimensional proba
bility distribution of the random vector (X\ Y). The individual probability distribu
tions of X and Y are referred to as the marginal distributions of the joint distribu
tion of (X, Y). The following relationship holds between the joint probability dis
tribution and its marginal distributions:
( 1.21)
The conditional probability of Ar=x, given Y=yi is, according to (1.6),

The sets
are the conditional distributions o f X given Y=yj and o f Y given X = xit respec-
tively. Hence, the corresponding conditional expected values are
Thus, the conditional expected value of X given Y is
Because the roles of X and Y can be exchanged, (1.21) yields
( 1.22)
The expected value of the sum X+Y is
Thus, in view of (1.21)
(1.23)
The random variables X and Y are said to be independent if
i.e. X and Y are independent if the random events "X = x " and "Y=yj are inde·
pendent for all i,j = 0,1,... (see section 1.1).
If X and T are independent, then, for all i,j = 0,1,...
The expected value of die product XY is
If X and Y are independent, then

(1.24)
Continuous components Let the components X and Y of the random vector

(X, Y) be continuous random variables with respective distribution functions and
densities
and
and
The joint distributionJunction of the random vector (X, Y) is defined by the proba
bility
as a function of x andy; x,y e (-oo,+oo). The joint distribution function of the ran
dom vector (X Y ) characterizes its joint or two-dimensional probability distribu
tion. (In case o f discrete components the joint distribution function is, of course,
defined in the same way.) F ^ y(x,y) has properties
1)
2)
3) (1.25)
4)
Conversely, any function of two variable which has these properties is the joint
distribution function of a random vector (X, K).
Assuming its existence, the second partial derivative of F(x,y) with respect to x
and y,
is called the joint probability density of (X, Y). The joint density can equivalently
be defined by
(1.26)
Every joint (probability) density has property

Conversely, any nonnegative function of two variables x and y satisfying this con
dition can be considered to be the joint density of a random vector (X, Y).
The probability that the random vector (X, Y) assumes a realization in the region B
of the (x,y)-plane is given by the area integral
(1.27)
Putting y = +oo and x = +oo, respectively, one obtains the marginal distribution
fiinctions of Fx y(x,y)
Thus, the marginal distribution functions belonging to Ρχ y{x,y) are simply the
distribution functions of X and Y, respectively. Similarly, the densities of X and Y
are the marginal densities belonging to fxty(x,y). In view of (1.26),
(1.28)
Two random variables X and Y with joint distribution function Pxty{x,y) are inde
pendent if, for all x and y,
or, equivalently,
If the joint density fxty(x,y) of {X, Y) exists, then the independence of X and Y is
equivalent to fx ty(x,y) being the product of the marginal densities:
If X and Y are not independent, then the conditional distribution function of X

given Y = y,
and the corresponding conditional density of X given Y = y,
are of interest. It can be shown that
(1.29)
Hence,
Thus, Fx(x) can be interpreted as the expected value of the conditional distribu
tion function of X given Y:
(1.30)
The conditional expected value o f Xgiven Y - y is
Therefore, the expected value of X given Y is
In view of (1.29),
(1.31)
The expected values of the sum and of the product of X and Y are given by
Taking into account (1.28) one obtains the same formulas as in case of discrete
components:
0.32)
and for independent A'and Y
(1.33)
The covariance Cov{X, Y) between the random variables X and Y is defined by
An equivalent representation for the covariance is
In particular,
From (1.33) it follows that if X and Y are independent, then Cov(X, f) = 0. But if
Cov(X, Y) = 0, then X and Y are not necessarily independent.
The correlation coefficient between X and Y is defined by
(1.34)
The correlation coefficient has the following properties:

1) If A”and Y are independent, then p(X, Y) = 0 .
2) If Y - aX + b for any constants a and ft, then p(X, Y) = ±1.
3) For any random variables X and T, -1 < p(Y, K) < 1.
The correlation coefficient is, therefore, a measure o f the linear relationship bet
ween random variables. X and Y are said to be uncorrelated if p(X, Y) = 0. Other
wise they are called positively or negatively correlated depending on the sign of
p(X, Y). Obviously, X and T are uncorrelated if and only if
Thus, if X and Y are independent, then they are uncorrelated. But if X and Y are
uncorrelated, they need not be independent.
Example 1.4 (bivariate normal distribution) The random vector (X, Y) is bivar
iate normal distributed with parameters
Px, Py, σχ , cy and p; -oo < μ*,μ^, < oo, σ* > 0, > 0, - 1 < ρ < 1,
if it has joint density (Figure 1.4)
The corresponding marginal densities are obtained by using (1.28):
Corollary If (X, F) is bivariate normally distributed with parameters μ*, σχ, μ^,
ay , and p, then X and Y are normally distributed with parameters μ*,σ* and μ^,
ay ; respectively. X and F are independent if and only if p = 0. (Note that the in
dependence of X and Y is equivalent to fxyy(x,y) =.ίχ(χ)ίγ(γ) ·)
Figure 1.4 Probability densities of the bivariate normal distribution
It can be shown that the parameter p is equal to the correlation coefficient between
A'and T(p = p(X,f))). Therefore,
I f the random vector (X, Y)has a bivariate normal distribution, then X and Y
I are independent i f and only i f they are uncorrelated.
The conditional density of X given Y=

(1.29):
Thus, the random variable X given = is normally distributed as
The parameters of this distribution are the conditional expected value of X given
that Y - y and the conditional variance of X given that Y = y:
O f course, in these formulas the roles of X and Y can be changed. Sums of normal
ly distributed random variables are considered in section 1.4. □
20 PROBABILITY THEORY
1.3.2 /i-Dimensional Random Vectors
Let X \,X 2 , ... ,Xn be continuous random variables with distribution functions
^*,(*1 )>Ρχ2(χ2), ...,FXn(xn) and densities /> ,(* i ), f x 2(xi), ■■■Jx„ (*«)· The
joint distributionfiinction of the random vector X = (X], X2, ... ,Xn) is
Provided its existence, the n th mixed partial derivative of the joint distribution
function with respect to the x \ , x2, ... ,xn:
is called the joint {probability) density of the random vector X. The functions
Fx (xh x 2, ... ,x n) and f\( x i,x 2, ... , X n ) are also called the n-dimensional distri
bution junction and n-dimensional (probability) density, respectively, of X. They
determine the n-dimensional probability distribution of X. The characteristic pro
perties of two-dimensional distribution functions and probability densities can be
extended in a straightforward way to n-dimensional distribution functions and den
sities. Hence they will not be given here.
The distribution functions and the densities of the Y, can be obtained from the jo
int distribution function and density, respectively, analogously to the two-dimen
sional case:
(1.35)
The joint distribution function of (Xj,Xj), i<j, is given by
The random variables X \,X 2, ..., Y« are said to be independent if
Assuming the existence of the densities ίχ.(χΐ), independence of the Y, is equiva

lent to
(1.36)
Conditional densities can be obtained analogously to the two-dimensional case:

For instance, the conditional density of X given Y,· = x, is
whereas the conditional density of X given
In this formula,
is the joint density of (X\ ,Xf)·
The expected value of the product Χ \Χ χ ' "Xn is defined by
In view of (1.35) and (1.36), for independent Xj this «-dimensional integral sim
plifies to
(1.37)
The expected value o f the product o f independent random variables is equal

I to the product o f the expected values o f these random variables.
Starting with (1.33), this fact can be more easily proved by induction. Hence it is
also valid for discrete random variables X(.
Let
be the covariance and the correlation coefficient between Xt and Xj, respectively.
(Note that σ,·,· = VatiX,) and p,·,· = 1.) It is useful to combine these parameters in
the covariance matrix Σ and in the correlation matrix p, respectively:
Example 1.5 {μ-dimensional normal distribution) Let Χ\,Χι,···,Χη be random

variables with vector of expected values μ = ( μ |,μ 2 ,···,μ«) and with covariance
matrix Σ. Furthermore, let be |Σ| the positive determinant of Σ, Σ'* the inverse of
Σ, and x = (x j ,* 2>·">*«)· Then the random vector X = (Χ\,Χι, —,Xn) is «-dimen
sionally normally distributed if it has joint density
is the transpose of the vector x - μ.
In detail, this density is
where Σ,-y is the cofactor of a ,y .

As a special case, for n - 2, xj =x and x 2 = y one obtains the density of the biva
riate normal distribution (example 1.4). Generalizing from the bivariate case it can
be shown that the Y, are Ν(μ,-,σ?)-distributed random variables; i = 1, 2 ,..., n. If
the Y, are uncorrelated, then Σ = ((a ly)) is a diagonal matrix (σ^· = 0 for i * j) so
that the product form (1.36) of the density and, therefore, the independence of the
Y, follows:
□
Theorem 1.1 Let the random vector (Yj ,Y2, .... Y«) be n-dimensionally normally
distributed and let the random variables Y\t Τ2,..., Ym be linear combinations of
theY,:
Then (T ,,Y 2, mYm) is m-dimensionally normally distributed. ■
1.4 Sums and Sequences of Random Variables

Let Y i,Y 2 ,"·,Y« be a finite sequence of random variables with distribution func
tions Fx.(xi), densities ίχ.(Χ() , finite expected values E(Y,) and finite variances
); i - 1,2, ...,n. Furthermore, let /χ(χι,*2> X n ) be the joint density of the
V a t iX j
random vector X = (Y i,Y 2, ...,Y n)· Next expected value and variance of the sum
of the random variables Y j, Y 2 , ···, Y« are determined.
1.4 SUMS AND SEQUENCES OF RANDOM VARIABLES 23
Expected value of a sum The expected value of the sum X\ +X2 + ··· + Xn
is defined by
From (1.35),
Hence,
(1.38)
The expected value o f the sum o f any random variables is equal to the sum
I o f the expected values o f these random variables.
Using formula (1.32), this fact can be more easily proved by induction. In view of
(1.23), formula (1.38) is also valid for discrete random variables Λ”, .
Variance of a sum The variance of the sum X\ +X2 + ··· + Xn is
Since
this formula implies
Thus, for uncorrelated X t ,
(1.39)
The variance o f a sum o f uncorrelated random variables is equal to the sum

o f the variances o f these random variables.
Let a i , <X2 , —, a« be any sequence of finite real numbers. Then,
If the Xt are uncorrelated,
(1.40)
The random variables X \,X i, ... are said to be identically distributed as X if all of
them have the same probability distribution as X. From a probabilistic point view,
there is no difference between identically distributed random variables. In the case
of independent, identically as X distributed random variables, formulas (1.38) and
(1.39) simplify to
Distribution of a sum Let Z = X + Y be the sum of two independent discrete

random variables with probability distributions
Then,
Letting rk = P{Z = k) yields for all k = 0, 1,...
(1.41)
A discrete probability distribution &= 0, l,...}with this structure is called a

convolution of the probability distributions {p,;/ = 0 , 1,...} and {qf,j= 0 , 1,...}.
Now let X and Kbe two independent, continuous random variables with densities
ίχ(χ) and fy(y). Then their sum Z = X+ Y has density
(1.42)
Integration yields the corresponding formula for the distribution function Fz(z) of Z:
0.43)
dF(x\
Since fix) = —— ^ or dF(x) =f(x)dxythis relationship can be written in the form
If X and Y are nonnegative, then (1.42) and (1.43) become
(1.44)
(1.45)
The integrals in (1.42) and (1.43) are called convolutions of the densities f% and fy
and the distribution functions F x and Fy, respectively. Hence, analogously to dis
crete random variables, the following statement is valid:
The probability density (distribution fiinction) o f the sum o f two independent

I random variables is g*ven by the convolution o f their probability densities
(distribution functions).
The density and distributions function of a sum Z=X\ +X 2 + — +Xn of n inde

pendent, continuous random variables X) are obtained by repeated application of
the formulas (1.42) and (1.43), respectively. The resulting functions are respective
the convolutions of the probability densities fx. and of the distribution functions
Fx , / = l , 2 , ..., n. They are denoted by
respectively. If, in particular, the Xj are independent and identically distributed

with density fix), then the density of the sum Z - X \ +X2 + —+Xn is denoted by
/*(">(z). This n th convolution power of fix) can be recursively obtained from
(1.46)
Analogously, the distribution of a sum of n independent, identically distributed

random variables X j , each with distribution function F(x), is given by the n th con
volution power F*^n\z ) of F(x), which can be recursively obtained from
(1.47)
For nonnegative X, the formulas (1.46) and (1.47) become
(1.48)
(1.49)
Example 1.6 (Erlang distribution) Let the independent random variables X\ and
X 2 be exponentially distributed with parameters λ i and λ 2 :
From (1.44), Z - X j +X2 has density
If λ 1 = λ 2 = λ , then
This is the density of an Erlang distribution with parameters n - 2 and λ (see

Table 1.2).
If λ | * %2 , then
Now let X \,X 2 ,...,Xn be independent, identically distributed exponential random

variables with parameter λ. Then the sum Z - X \ +X 2 + ··· + Xn is Erlang distri
buted with parameters n and λ. This statement can be easily proved by induction:
The assertion has been already proved for n - 2. Assuming X\ +X 2 H— *Xn-\ *s
Erlang distributed with parameter η - 1 and λ. Then, according to Table 1.2,
where f(x) = λβ **, x > 0. On condition that this assumption is true, the density of
(1.44)
Thus,
But this is the density of an Erlang distributed random variable with parameters n
and λ. The corresponding distribution function is
(1.50)
□
Example 1.7 {Normal distribution) Let the random variables Xt be independent
2
and normally distributed with parameters μ,- and <sj; i = 1, 2 :
According to (1.42), the probability density of the sum Z=X\ +X2 is
Substituting u = x - μι and ν = ζ - μ \ - μ 2 yields

(
Replacing the term (v - u)^/a\ + « 2/σ? in the exponent of the integrand by
and substituting
28 PROBABILITY TH EORY
yields
Since ef2/2 dt = J i n , the density o f becom es
By theorem 1.1, a sum o f norm ally distributed random variables is also norm ally
distributed. For independent random variables, example 1.7 yields, by induction, a
sharpening o f this statement:
Corollary Let Z - X \ +X 2 + ... +Xn be the sum o f independent random variables
X(= Ν ( μ ,,σ 2); 1 = l,2 ,...,n . Then,
From theorem 1.1 we know that if —,aΧ

hs a jo in t normal distribution,
then the sum X \ + X 2 + — +Xn has a normal distribution. In particular, if ( ,
has a bivariate norm al distribution with param eters μ*, μ^, ox ,a y and p, then
X + Y has a normal distribution with
Standardized random variables A random variable X is called standard

ized if E(X) = 0 and Var(X) = 1. If Xsi any random variable with finite expected
value and finite variance, then
is a standardized random variable, since obviously 0 and Var( = 1. The

random variable Y is called the standardized version o f X.
If X is norm ally distributed, then the standardized version o f X is also norm ally
distributed: Y = 0N
( ,1). The density and distribution function o f an 0, ^ -d istri
buted random variable, which is referred to as a standard normal random variable,
are denoted by φ(χ) and Φ (χ), respectively:
(1.51)
Theorem 1.2 {Central limit theorem) Let X\ ,Χ ι , ... be an infinite sequence of in
dependent random variables with finite expected values E{Xj) = μ and finite var
iances VaiiXj) = σ 2 ; i = l , 2 , .... Furthermore, le t Zn =X\ +X2 + ·" +Xn and
Then,
■
Corollary For n sufficiently large, Zn is approximately normally distributed with

expected value «μ and variance no2 :
Theorem 1.3 deals with the sum of a random number of random variables. To state
this theorem, another important concept has to be introduced:
Definition 1.2 {Stopping time) An integer-valued, positive random variable N is

said to be a stopping time for the sequence of independent random variables
X \,X 2 ,.~ if the random event ”N = n" is independent of all Xn+\,X n+2,··· for
« = 1, 2 ,... ·
Comment A random event A is independent of a random variable X if for all x the ran
dom events .4 and ”X i x" are independent.
Intuitively, the notation "stopping time" can be motivated by assuming that the
random variables Xj are observed in turn. For N - n , this process is stopped after
having observed X\,Xi,...,Xn, i.e., before observing ArM+j,AT„+2 ,...
Example 1.8 Let Xj = 1if after the rth tossing of a coin a "head" is observed and
Xi = 0 otherwise. Further, let P{Xj = 1) = P{Xt = 0) = 0.5 for all i - 0 ,1 ,... Then
is a stopping time for the sequence X\,Xi,.... □
Theorem 1.3 {Wald's equation) Let X \,X 2, ... be a sequence of independent ran
dom variables and let N be a stopping time for this sequence. Assuming the Λ',· to
be identically distributed as X and the expected values E{X) and E(N) to be finite,
then
(1.52)
Proof Let the binary random variables F,· be defined by
F, = 1 holds if and only if no stopping has occured after observing the random va
riables X \,X 2, ...,Y ,_i. Since N is a stopping time, F, is independent of the ran
dom variables X{, X i+\ ,... In particular, F,· is independent of Xt so that
Since £(F,·) = PiN 't i), formula (1.11) yields
Obviously, formula (1.52) holds if N is independent of all X\,Xiy ...
Since E(X{) = 1/2, applying Wald’s equation to example 1.8 yields
According to the definition of N, X\ +X2 + —+Xn = 8 . Hence, E(N) = 1 6 .

Now, let Xj denote the random gain (loss) of a player when tossing the i th coin
(depending on whether "head" or "tail" is observed. If all Y, are identically distri
buted according to P(Y, = -1 ) = P(Y, = +1) = 1/2; i = 1,2,...; then
is a stopping time with respect to X \ , X 2, ... A formal application of Wald's equa

tion would yield
The left hand side of this equation is equal to 3. The right hand side contains the
factor E(X) = 0. This situation requires the assumption E(N) = oo. Therefore, in
this case Wald's equation is not applicable.
1.5 Transformations of Probability

Distributions
The analytical treatment of stochastic models can frequently be facilitated by go

ing over to transformations of the probability distributions of the underlying ran
dom variables. This section deals with the z-transformation for discrete random
variables and with the Laplace transformation for continuous random variables.
Taking into account the applications dealt with in this book, the discussion of these
transformations is restricted to nonnegative random variables. As usual, the proba
bility distribution of a discrete random variable is identified with the set of the
probabilities of its realizations, and the probability distribution of a continuous ran
dom variable is identified with its distribution function or its probability density.
The set of all transformable probability distributions is called the pre-image space,
whereas the set of all images of probability distributions is called the image space.
The following fact is the basis for the successful application of the transformations
mentioned:
There is a one-to-one-correspondence between image space

I and pre-image space.
This fact allows the application of the following algorithm for tackling stochastic
models: The problem, originally given in the pre-image space, is transformed into
the image space. There it is solved. The solution obtained in the image space is re
transformed into the pre-image space to obtain the solution of the original prob
lem. Hence the application of the transformations is only efficient if the problem
in die image space can a priori be more easily solved than the original problem.
For instance, if a system of linear differential equations of the first order with con
stant coefficients has to be solved for the time-dependent probability distribution
of a discrete random variable, then the z-transformation reduces this problem to
solving a linear algebraic equation system with constant coefficients.
Retransforming the solution obtained in the image space is usually the most diffi
cult step when applying the algorithm. But even without retransforming, the image
in many cases provides important or even sufficient information on the pre-image.
For instance, knowledge of the z-transformation and the Laplace transformation of
probability distributions makes it possible to compute moments of all orders of
probability distributions. Therefore, these transformations are also called moment
generatingfunctions. In particular, expected values and variances of random varia
bles can be immediately obtained from the images of their probability distribu
tions.
1.5.1 z-Transformation
Let X be λ discrete random variable with range {0,1,2,...} and probability distri
bution
Definition 1.3 The z-transform of the random variable X and its probability dis
tribution {po>p i ,/>2 >—} >respectively, is defined as the infinite series
Clearly, M(z) is the expected value of zx :
(1.53)
M(z) converges absolutely for |z| < 1:
Therefore, M(z) can be differentiated (and integrated) term by term:
Letting z = 1 yields
Taking the second derivative of A/(z) gives
Letting z = 1 yields
Thus, the first two moments of X are

(1.54)
Continuing in this way, all moments of X can be expressed by means of A/(z).
Since (1.10) is also valid for discrete random variables, the variance of X is
Random documents with unrelated
content Scribd suggests to you:
The Project Gutenberg eBook of Salvage
This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.
Title: Salvage
Author: Roy Norton
Release date: March 24, 2024 [eBook #73244]
Language: English
Original publication: New York: Street & Smith Corporation, 1928
Credits: Roger Frank and Sue Clark
*** START OF THE PROJECT GUTENBERG EBOOK SALVAGE ***

Salvage
By Roy Norton
Author of “Mr. Catlin’s Weakness,” “Widders Come First,” Etc.
Captain Drake again proves, and this time most brilliantly,

his right to the title: “The Opportunist.” With remarkable
sagacity he scents an opportunity on the wind.
Piræus, that historical port of Greece, lay drenched and sweltering in

sunshine. Its great water front, whence galleys had sailed bravely
forth in ancient days, was packed with shipping, most of it idle; for
trade was in the doldrums. Docked between two big “smoke boats”
lay the very trim and neat steam schooner, Malabart, Captain Eli
Drake, owner and commander; and there was nothing in the
Malabart’s physical appearance to indicate that she, too, was
yawning for a cargo of any sort, or to any port, though the charter
rate might be so low as to barely pay expenses. Captain Drake,
whose sobriquet of “The Old Hyena” had survived the days of sail,
was ashore, harassed by cares.
He had been eating into capital to keep his ship in commission,
and his crew, which he had gathered in the course of many years,
from being disbanded. He prized his crew and, after his ship, they
came first in his affections. In quest of cargo—any cargo—he had
scoured the port, made daily trips over the short drive to Athens,
and spent liberal sums on cablegrams to many agents, without avail.
He felt like cursing the big steamship companies, which, with their
army of organized runners, were rapidly driving the independent
owners and tramps off the seas.
In a mood of sullen obstinacy he had tramped almost the length
of the docks when, unexpectedly, he heard the clatter of a cargo
winch; and the sound was so unusual that, like a magnet, it drew
him in its direction. He found a rusty tramp that was lading.
“Now what—how did I miss getting that cargo?” he reflected.
“The Rhodialim, eh?” And after a moment’s thought he muttered:
“Oh, yes. Belongs to that firm of Hakim & Letin. Got her and one
other schooner, doing mostly Levantine and East African coast trade.
Wish I could have got that cargo. My luck’s out.”
Without thinking, or observing that at the dock gates there was a
watchman, who, at the moment, had his back turned and was in
voluble altercation with one of his countrymen, Captain Eli strolled
inward. He finally halted, and with hands in pockets stared, suddenly
discovering something else that made him curious.
“That’s blamed funny!” he thought. “Big cases marked ‘Mining
machinery,’ but a couple of stevedores chuck ’em into the slings as if
they were empty. Also cases of merchandise put up like heavy prints
that seem just as light, and as—— Good Lord! Up there on the
bridge! If that ain’t Bill Morris, I’m dotty. So he’s got a ship out here,
eh, after it got too hot for him about everywhere else on salt water!
Lost two ships in the Pacific trade, under mighty suspicious
circumstances, and had his ticket taken away, last I heard of him.
Ummh! Ten or twelve years ago, that was. So he’s skipperin’ this
craft, eh? If him and me hadn’t locked spars two or three times, I’d
go over and rile him up with a leetle light, airy banter. I guess he’s
——”
“Hey you! Got any business here? How’d you get past me at the
gate?” a voice disturbed him. And although the fellow spoke bastard
Greek, Drake, who, with a sailor’s facility, had picked up considerable
of the tongue, understood, and turned to see the watchman
glowering at him.
“Why?” he asked. “Can’t anybody come onto your dock? Nothing
secret about it, is there?”
The watchman sputtered something about none without a pass
from Hakim & Letin being allowed in, and somewhat peremptorily
ordered Drake to clear out. Not being accustomed to such
treatment, disgruntled, affronted, but recognizing the weakness of
his position and the futility of retort, Drake turned and, swearing
under his breath, obeyed.
It is possible that the episode might have passed from his mind
entirely, but for an encounter that followed some hours later, when,
just as he was turning toward the Malabart, a man whose face bore
the almost indelible stamp of the engine rooms of ships, with grease
worked deeply into the pores of the skin, respectfully touched his
cap peak and accosted him in fairly good, though accented, English.
“Captain Drake,” he said. “Excuse me, sir, for stopping you, but I
am a good man out of work, and want a job on your ship, sir.”
“Sorry, my man, but we’re full up,” Captain Eli replied. “Too full,”
he added, and would have proceeded on his way, had not the
applicant insisted.
“I am good man, sir. First-class engineer; but I would take
anything in your engine room. Because me, I have big family, and
ships are all full now, it seems to me, sir. I lose job when not my
fault. Not at all. When I took engines of ship Rhodialim anybody tell
you they scrap heap. I make ’em good. And now, without word,
since that Captain Bill Morris come, I am fired. He say have his own
engineer and——”
“Huh? What’s that?” Captain Eli, who had been slowly moving
forward with the insistent one at his side, stopped and stared at the
man. “Do you happen to know the name of the new engineer?”
“It is Simmons, or Simons, or something like that.”
Drake’s mouth pursed itself as if to whistle an exclamation, and
for a moment he stood absent-mindedly staring at the stones
beneath his feet. But his thought ran: “Simmons! Simmons! That
was the engineer of the ship that Bill lost last, and he was one of the
chief witnesses at the insurance investigation. Something funny
about this business!”
He abruptly started away, saying as he did so: “You come on
board with me, and I’ll learn if there’s anything can be done. Let me
see your ticket.” And then, a moment later: “Beltramo—Giuseppe
Beltramo is your name, eh? And your ticket shows a long, clean
record. No wonder they didn’t want you on that boat. Never mind
the questions, now. I’ll ask all the questions myself.”
As a rule the relations between Captain Eli Drake and his chief
mate, William Catlin, were of two separate characters, inasmuch as
afloat they observed the distinctions in station and Drake brooked no
interference; but it was well known that ashore they were more
intimate than brothers usually are, and confidants in nearly
everything. Hence, when the commander sent for Catlin upon his
arrival aboard, and on his entry into the cabin addressed him as
“Bill,” Catlin thought: “Something’s turned up.” Aloud he said:
“Landed something, skipper?”
“Landed enough to set me to a heap of thinking, Bill,” Drake said.
And then he bent forward and in a confidential tone told of his
experiences, ending with: “It seems to me there’s some sort of a job
being put up by Hakim & Letin; and—well, there might be some way
for us to make something out of it.”
“Sort of an opportunity, eh?” Catlin grinned, remembering that the
Cape Cod man had earned the reputation of being an opportunist.
“Maybe you can see one, but I can’t. Don’t mind my thick-
headedness. I can get anything when it’s explained, all right.”
But Drake seemed to have become absorbed in some thought of
his own. He stared absently through the cloud of pipe smoke; and
finally chuckled, as if he had reached a solution of some problem.
“I think I sort of grab an idea,” he said at last, getting to his feet.
“You’re a hell of a good friend, Bill, but as a helper in working out a
puzzle you don’t amount to much. Never mind. Think I got it, myself.
So just talking it over with you did have some use, after all. You go
down and keep that feller I brought aboard interested, while I slip
below and see the chief. Most likely be in his cabin, I expect.”
They went out together and Drake sought the engineer.
“Forbes,” he said to that gray-haired veteran. “Can you find
something for an engineer out of a job to do for a few days?”
“Can’t find enough to do myself, let alone make work for a new
man. If this keeps up—— Hold on. While I think of it, that chap Flint,
my third, asked me to-day if I thought there was a chance for him in
the navy. Now if he got a month off to go to the nearest place he
could pass his examinations and file his application——”
“The sure-enough right thing! Let him take a month, and put this
feller on until Flint comes back. Come on up topside and talk to him.”
The result of the conversation was that two men, at least, were
made happy that evening—Flint, who had got unexpected leave for
a month, and Beltramo, who had got a temporary billet.
But Drake was not on the ship when the shift was made. In the
roughest suit of clothes he could muster he had gone ashore and
made his way to a not too-clean bar, where he knew that pilots were
wont to gather. There he patiently waited for the arrival of one he
knew. The man came at last, and Captain Eli drew him into a little
private room at the rear.
“Christophe,” Captain Eli said, “I have done you a favor once or
twice, and you’re the kind of man that likes to repay. Well, the time
has come when you may be of use. Now first, you’ve got to keep
your mouth shut—not one word—not one word to anybody, not even
your wife, of what we say here in this room.”
The pilot, whose face was seamed with years and sea service,
promptly lifted his hand and swore an oath that would have satisfied
any band of conspirators that ever existed.
“First, you know this sea as well as any one, I take it?”
“By Heaven! Better than all save one or two. Was I not a
fisherman in these waters when old enough to float? I know every
foot of it and every reef, and every island and——”
“Good!” Captain Eli interrupted. He leaned across the little table
between them and lowered his voice. “Christophe, if you were going
to sink a ship that was supposed to be bound eastward—say for
Jaffa—where would you do it?”
For a moment the pilot’s mouth hung open and his eyes were
wide, as if he feared for Drake’s sanity.
“But, sir, captain—you—you are not going to sink—— You don’t
mean that——”
“No, of course not! I sink nothing. But you think it over carefully
and answer my questions,” Drake continued. And the pilot, still
wondering, slowly lowered his eyes, shut them as if to ponder such
a case, and then asked: “What time of year, captain?”
“This time of year,” Drake replied.
And again the weather-beaten old pilot shut his eyes and thought.
“Listen, sir,” he said in his quaint but adequate English. “Many
things one must think of. If mens want to sink ship, but not drown
anybody, they must be not too far from land for open boat, eh? Must
be some place where not too much danger big seas for small boats,
eh? Must also be some place where nobody see—away from
fishermen’s boats, or cargo boats, or bigger ships—some place
lonely this time year. Plenty places man could scuttle ship, but few
where get all these things what want, eh? Well, about now most
fishin’ boats work”—he got up and walked to a rough map that was
tacked on the wall and that was almost solidly smeared with the
trails of many fingers across its surface—“works up about here
mostly. In some months, here; some months, there; but now, about
here. So no good up there.” His finger moved as he talked. “No good
through here, because big ships go there. No good there, because
small ships what do island trade work in and out. So, here best place
for all things. Almost only place which fit all I speak between here
and Island of Rhodes. Not too far out of the way. Very good place.
Deep water—plenty water and not much chance boat ever drift
when hit bottom. Yes, captain, sir, that best place anybody can think
of—right about there.”
His gnarled finger ceased to move—pointed at open water off
Nauplia.
They sat down again and, while Christophe eyed him with
perplexed looks, the captain reflected.
“That, you think, is a place a man who knew these waters well
would select?” he said. “But a man who didn’t know them?”
“God knows where!” the pilot exclaimed, lifting his hands and
letting them fall to the table again. “It is the place—the place I say—
where one who knew would choose in, say—seven times out of ten.
As you, sir, know, there are some thousands of islands.”
For half an hour Drake continued to catechize, but without stirring
the old pilot from his conclusions.
“Well, Christophe,” he said at last, arising to go, “I’m going to hire
you for a cruise that may never take place; but I’m taking a little
gamble on certain things. You begin work to-morrow, always with
your mouth shut. Here’s what you are to find out: First, when the
Rhodialim sails. Second, if she’s taking a pilot aboard, and if so, who
and what he is. And third, you’re to report to me aboard the
Malabart each evening just after dark. I don’t care to have too many
notice that you come there. Is it understood? Going wages, of
course,” he concluded, with Yankee thrift.
“Yes, sir, captain. Very well I understood it, and do what you ask.
Maybe some time you tell me why all this, eh?”
“Maybe,” said Drake laconically, as he thumped upon the table to
pay for his bill. And he left behind him one who was still wondering a
little if a certain Captain Drake was all there.
Catlin had a surprise on the following morning, when told that
they were going to take on some supplies. And he was still more
astonished when Drake asked him to muster the crew and learn
whether there was any man aboard who had ever had any
experience in diving. Catlin found a stoker who admitted that years
before he had worked for a salvage company. Drake told the man to
get on shore-going clothes and come with him, and the twain
disappeared. The man returned that afternoon accompanied by a
truck, which duly unloaded and brought aboard a collection of stuff
that made even Catlin scratch his head, and caused conjectures
for’ard as to whether The Old Hyena was going into the wrecking
business. It consisted of a complete diving outfit—air pumps and all
—as well as huge collision mats and handling gear. Drake did not
appear until evening, and seemed unusually speechless, and he
dined and waited for Christophe.
The latter came at last, grinning with self-satisfaction, and was at
once closeted with Drake, who asked: “Well, what did you learn?”
“That Rhodialim, she sail day after to-morrow. She got most her
cargo aboard now. But it’s funny, captain, sir, she got one man who
knew this sea same as me. Long time ago he fisherman, then go
away, and been down Smyrna where not got too good name. Good
man, when sober, but too much drink, so never get good job. That
man I see in saloon. He most full and—— You owe me thirty
drachmas, I spend on him get him fuller, so he talk. Bymeby he
borrow fifty drachmas from me, which also you owes me. He brag
some and say pretty soon he pay back. Pretty soon, maybe two
weeks, he come back with plenty money in pockets. But he shut up
like oyster when I ask how make this so much money, and he say
nobody but him ever goin’ know that. Now what you wish me
make?”
“You go home and keep on keeping your mouth shut. Come
aboard at noon to-morrow. We sail to-morrow afternoon.” Drake was
suddenly decided in his movements.
“How long be gone from my old woman?” Christophe asked.
“Can’t tell. Maybe one week, maybe two. Not likely to be longer, I
think. But all you’ve got to do is to come aboard and I’ll tell you then
where we’re bound. I’m going to clear for Smyrna. There will be no
secret about that.”
On the following morning when Captain Eli went ashore he took
with him the chief engineer. The latter returned with two big
machine cases and armored, high-pressure hose, together with a
case of fittings. Late that afternoon the Malabart slipped out and
away, so palpably light that other sea captains who observed her
shook their heads with understanding. A ship putting to sea in
ballast in dull times evokes the sympathy of the seawise. Aboard the
Malabart there was an air of gloom among the crew.
The captain and owner, walking the bridge, said to Catlin:
“Well, Bill, I’m taking a gamble—thousand to one shot, that’s all.”
When dusk fell the island of Thermia lay close in to starboard, and
the man at the wheel stood ready to port his helm and bring her
over from the sou’-east-by-east to an easterly course to round the
island, that being the route toward Smyrna; but old Christophe,
standing behind him, took the wheel, rang for slow speed and
groped in toward the island. It loomed up about them, a rocky point,
before he said over his shoulder to Captain Eli:
“Here’s where we can lay to, sir. Good anchorage here in this
cove, and no risk of wind.”
All that night she rocked there, gently; on a sea that was almost
without a swell. And when morning came, to the crew’s further
curiosity, she brought in her hook, swung about, and headed due
west, plodding along at slow speed and apparently purposeless. A
liner came out of the north and gave her a passing hoot. Christophe,
eyeing the other boat, said to Captain Eli:
“She be for Messina way, and now not likely be another ship along
here for ten days. That’s what those mens know. If I make good
guess, that’s why they clear Pirzeus to-day, after big ship go, sir.”
“And when will we make that Island of Hydra?” Drake asked,
staring to the westward.
“Just about sunset, captain, sir. Then we slip round it and there
are small islands between it and mainland, and entrance into Nauplia
which so long and so big it is like long gulf. We lay behind them
islands, sir, and—see what shall see about midnight, I think, sir.”
Drake caught his dry, knowing grin, but did not entirely share his
confidence as to the outcome of their strange voyage.
The pilot’s prediction as to progress was fulfilled; just as a hazy
sunset colored the tips of the high, bleak mountains behind which
the day disappeared, they passed the isle with its abandoned and
obsolete fortifications, and hove to in waters that seemed to have
been deserted since the time of ancient wars. Night fell with a thin,
low-lying fog that seemed to sweep down from the great bastions of
Nauplia and rest on the still waters. The stars were obscured and a
new depression engulfed Drake.
“The weather’s against us,” he said gloomily, to the storm-beaten
old pilot. “They could pass us at a couple of cable lengths and we’d
never know it.”
“Not if we were out in a small boat, listening,” Christophe said. “In
small boat hear everything. On ship, no—not so quite well. We must
put out boat and get out maybe two three miles and wait. Yes,
maybe fog too bad, one way, but very good, other. When they pass
we get course then slip quiet, very quiet, same way, with Malabart,
eh?”
Drake pondered. There seemed no other method. He cursed the
fog, but ordered a boat away with Catlin and the pilot aboard, the
latter assuring him that he could find his way back to the ship if the
night were as black as the pits of Satan. A long wait followed after
the boat had disappeared. The gloom of the darkened Malabart, the
lack of the bell striking the hour, the absolute stillness of the ship,
were all upsetting. The very lifelessness of the protected water
where she lay was annoying, for there was not the slightest lapping
whisper of a wave against her hull. Down in the engine room even
the stokers who kept up steam had been cautioned against the
clanging of a furnace door or the ring of a shovel. Had one passed
the Malabart within ten yards he might have thought her the ghost
of some long-abandoned ship. Drake listened from the outer wing of
the bridge, bending over, sometimes with a hand cupped to his ear,
until he was tired. He had about decided that his voyage and
expenditure had been born of folly, when he heard a faint creak,
followed a minute later by another. Then Catlin’s voice below hailed
softly, and the boat pulled around to the side ladder, which had been
lowered and swung barely above the water.
“All right, sir. She passed so close that she almost ran us down.
She had doused her lights and was not doing more than five or six
knots. Christophe says there could be no mistake. She was the
Rhodialim, all right.”
The pilot joined in with: “About a mile and a half out. Long row
back.”
“It won’t do for us to follow too closely on their heels, anyway,”
Captain Eli said. “But are you certain that you can pick her up again,
Christophe, in all this murk?”
“I know the course she will take. I think so, with luck,” the pilot
said. “They not alter course again. Too much else business think of, I
expect. Just keep straight on about five, six miles; then stop. They
not want go much farther. Might meet small fishin’ boats out of
Nauplia. Not take chance of that, eh?”
Captain Eli stood blockily, a dim figure in the darkness, and
seemed making mental calculations.
“I don’t think we’d best be in too much of a hurry,” he said at last.
“We’ve got to take the chances of being too late. If the crew are in
on it with the commander, mate and engineer, there’ll be no time
wasted. If they’re not, the boats won’t be ready to lower, and
besides he’ll have to put up a bluff at saving the ship, to fool the
crew. We’d best give them at least an hour and a half.”
“That crew, captain, sir, are the scum of the water front,”
Christophe put in.
“But just the same, we don’t know that they’re in on it,” Drake
replied. “Bill Morris don’t like to cut too many in on his crooked work.
Seems to me more likely that he’ll try to stampede ’em into the
boats after putting up a great show to save the ship. He’ll call on his
engineer for steam and announce that they must beach her. The
engineer will either pretend to start the engines, or swear that he
can’t turn ’em over. That would stampede the crew, if they’re the
sort one picks up in these parts. I think we’ve got to risk it, and give
’em an hour and a half, certain. After that it depends on how quickly
we can pick her up. Beltramo tells me that she’s fitted with two sea
cocks only into her main hold, because her engines are set well aft.
So she’s not likely to fill within some hours after they’re opened, and
I’ve got it doped out from what I know of Morris’ work that’s the way
he’ll put her under, if that’s what he intends to do; but it’s only little
things that are queer which makes me think that’s what he’s up to.
Big gamble, but——”
“Must be. If not, why he not go on to east’ard?” the pilot asked.
“I’m sure of it, captain, sir.”
But Drake was still doubtful when, still in blackness and running at
slow speed, the Malabart nosed out into the sea with the pilot
himself at the wheel and keeping an eye on both time and compass
as he took up the trail. To the commander’s ears it seemed that with
the ship so light that her blades were barely under water the thrash
of the slow-turning screw must be audible for miles. He saw the
wheel slowly revolving under Christophe’s hands and sensed that the
pilot was now where he thought they might find the sinking ship.
Captain Eli knew that both Catlin, and the second mate, Giles, and
nearly all the crew were forward peering into the dimness ahead,
but it seemed impossible to see anything on such a night. It was a
matter of luck, and he felt a dawning apprehension that his luck was
out. Watching the compass over the pilot’s shoulder he saw that the
ship had made one complete circle and was now holding dead
ahead. The wheel again whirled, and they began another circle, a
mile deeper in that huge bay surrounded by high and forbidding
mountains, when there came a soft whistle from forward and a
pattering of bare feet. Catlin’s muffled voice came from below:
“Hold her, sir, hold her. I think we’ve sighted the Rhodialim about
two points off the port quarter.”
Drake jumped to the engine tube—it having been arranged that a
man was to stand by to obviate the use of bells, inasmuch as the
sound of an engine bell might carry far in such stillness—and now
the Malabart lost way and came to a stop. The boat, which was
swinging barely above the water, was lowered, and Drake, Catlin,
and two men tumbled in and fell to the oars. They rowed quietly.
“There she is, sir,” Catlin whispered.
Exercising still more caution, they drew down on the dim shape
that lay inert and heavy on the water. They came alongside and
listened for voices, but caught no sound. They found the boat davits
hanging idly over the water, and went up the falls noiselessly, and
stood on the deck. Together they ran here and there, making a
search for any human being. Not until then were they confident that
she had been abandoned. Listening down the main cargo hatch they
could hear the swirling and gurgling of water and the soft bumping
of empty cases and crates.
“Get back to the ship, Bill, and rush across all the men that can be
spared; so that if that gang are standing by waiting for the
Rhodialim to sink, we can knock ’em overboard. Tell Christophe to
bring the Malabart alongside twenty minutes after you’ve gone.
That’ll give you time to be back here ahead of her; so if we have to
repel boarders, we’ll have the men to do it. Be as quiet as you can
and get a move on.”
Catlin slipped away and over the side like a ghost. After he had
gone Drake listened attentively for a few minutes, then went back
and again bent over the open hatch. Afterward he tried, by leaning
far over the rail, to estimate how deeply the scuttled ship had
already sunk. It seemed to him that she couldn’t last very much
longer. Taking an electric torch from his pocket, he went below. She
was a fairly deep ship, of good draft, and he was pleased to observe
that the cabin floors were not yet damp. He decided that if the sea
cocks were of the diameter given by Giuseppe, the former engineer,
she had at least an hour and a half longer to float. He knew that her
fires must have been drawn, because Morris would not run the risk
of the sound of a boiler explosion drawing attention to the spot, if
there chanced to be any boat within hearing.
“He knows this business of scuttling ships better than any one I
ever heard of,” Drake soliloquized. “But if he cleared off this time,
without waiting to see her under, he made one hell of a mistake.”
He looked at his watch in the light of his torch and meditated: “If
Bill moves lively and doesn’t lose his way, he should be back here in
half an hour from now. If he loses his way in this blamed fog—I’m
afraid we cut it pretty short!”
He climbed back to the deck, went to the port side, from which
the boat had put off, and listened, prepared to answer a hail, if
Catlin returned groping and had to shout to learn his bearings. Then
from the opposite side of the ship, he heard a single telltale thump,
as if an oar in clumsy hands had slipped from an oarlock and
brought up with a bang.
Drake ran across to the starboard rail just in time to hear a
muttered imprecation, in colloquial Greek:
“Quiet there, you lubber! If the skipper and those two pets of his
are hanging around, we’ve a fine chance of getting away with
anything.”
Drake pursed his lips into a silent whistle, and through his mind
ran the thought: “It’s the crew of this craft come back. Probably
suspected something and are trying somehow to double-cross
Morris, Simmons and whoever they’ve let in on it with ’em. I’m a
fool. Should have kept at least one man with me for such an
emergency.”
Quick as was his thought, his action was quicker. He jerked off his
boots and threw off his jacket. He ran aft in the direction that he
was certain the boat must take to board, and leaned over the rail
just as a man started to climb upward.
“Get back into that boat and sheer off,” he called down. “This ship
is abandoned and is salvage.”
The man hesitated, and a voice from below ordered:
“Go on up! We’ll talk this over on deck.”
“Like hell you will!” Captain Eli declared. “And if any man tries to
come on this ship, he’s looking for trouble. Sheer off, if you want a
talk. If you want a fight, come ahead.”
The man holding the boat fall climbed up and got a foothold on
the strake. He threw a hand inward and caught a rail stanchion and
swung upward, encouraged by muttered comments.
“All right! If you will have it——” Drake growled.
And leaned far over, and struck. In the gloom and darkness he
had not struck well, and instead of knocking his man overboard into
the boat below, he merely shifted him outward just beyond reach of
a second blow. Drake threw himself over the rail and hanging by one
hand struck again with the other. It was a body blow, but the man
was tenacious, clung to the rope, swayed like a pendulum, and, as
he swung back, kicked at Drake with his heavy sea boot. But this
time Drake’s fist smashed home, and the boarder grunted, loosened
his grip on the boat fall, and went slithering down among his
companions. Drake climbed back over the rail just in time to feel a
stunning smash on the back of his head, and was not until then
aware that while he had engaged one assailant, another had climbed
up the opposite boat fall with a monkey-like agility, and had come
behind him.
Infuriated by the attack, he whirled, seized the man, lifted him as
if he were a bundle of waste, and, with a giant’s heave, threw him
far outward. The man shouted as he fell, but Drake did not hear the
splash; for now he found himself fighting desperately with two other
dark shapes who charged silently. Even as Drake fought, he recalled
what he had overheard, which convinced him that these men also
had no wish to recall Morris and his fellow conspirators. Drake
grinned at the humor of that situation—a scalawag crew trying to
steal aboard the ship they had abandoned, Morris and his fellows
somewhere out there in the dark, himself battling for the salvage like
a dog for a bone, and all the time, down there in the hold, the sea
cocks flooding the sinking ship.
The number of his assailants increased. They were urged on by
the leader in a hoarse mutter:
“He’s alone. He must be alone, because no one else comes. Down
him! Down him, because he’s probably got a boat coming!”
Drake fought desperately. Two of his assailants went to the deck
and lay there struggling, as they tried to recover their senses.
Veteran fighter that he was, the participator in events which had
earned for him the sobriquet of The Old Hyena, he used his, head
coolly, his fists heavily, and as he moved here and there slipped out
of the dangers of being cornered and fought for time.
Then came the accident. Retreating, his heels caught over a coil
of rope that had been carelessly left on the deck. He struggled vainly
to recover his balance, but they were on him like a pack of wolves.
And in a fighting, struggling group came to the deck, where they
twisted and turned as he tried to regain his feet, was pulled down,
tried again, was struck heavily over his eyes, saw stars, shook his
head like an enraged bull, and felt himself pinioned to the deck while
one of the men he had previously knocked down arrived in time to
kick him in the ribs. He was now roaring with fury, heedless of all
alarms and thinking of nothing but revenge. He did not hear the
angry shout of Catlin and his men coming on deck. The hold on him
suddenly relaxed. He sat up, rubbing his bruised side and clearing
the blood from his eyes, heard Bill Catlin’s fighting oaths and got to
his feet. Both forward and aft shadowy forms of men in flight flitted
across the decks. He heard Catlin’s shout:
“Don’t let ’em get back to their boat! Knock ’em out and hold ’em.
They’ve probably killed the skipper. If Drake is dead, we’ll drown the
whole damn lot!”
“Yes, don’t let ’em get away, Bill,” Drake shouted, climbing to his
feet and regaining his full senses. “I want ’em. Particularly that
fleabitten rat who gave me the boots. Lash ’em up and get ’em
together. Quick! The ship may sink under us at any time.”
Both he and Catlin ran here and there to bring matters to a
conclusion, and within a few minutes there were seven somewhat
bruised and battered ruffians thrown into the nearest cabins and,
despite their protestations and appeals, locked in. Their leader, who
time and again shouted that he was the second mate of the
Rhodialim, was the first to whine for mercy. He cried, in
comprehensible English:
“You hell of an Ingleeshmans tie us up and put us here to drown.
You let us go we make no more of the fights. We go quiet. But
capitano, please, sir, not drown us.”
“Drown nothing!” Drake growled. “If we see that we can’t save
the ship, we’ll bring you up and turn you loose in your boat, you
damn pirates! And listen here! You keep quiet now. We’ve got no
more time to waste on you.” He turned to Catlin and said: “Lock ’em
in. We’ve got to fall to, if we want to keep this craft afloat.”
They hastily ran out to the deck just in time to hear Giles, the
second mate, calling:
“Ship’s coming, sir. Shall I flash a light for ’em, or hail?”
Drake himself cupped his hands and called: “Malabart, ahoy! This
way!” When he got a response, he ran back to where he had fallen,
struck a match, found his electric torch that had fallen from his
pocket, and with it as a beacon, directed the Malabart to come
alongside.
He called for Beltramo to come aboard to point out the location of
the sea cocks, and for the collision mats to be put across. He set
lookouts to guard against the possibility of other boarders, and
himself took a hand at the work.
“If the others haven’t heard the row, it’s not likely they’ll come
back,” he said. “But we’ll take no chances; we’ll keep as quiet as
possible, just to avoid any more risks of interference. Move lively
now!”
The men of the Malabart ran here and there, their bare feet
pattering, and pulled and hauled a huge, unwieldy mat to the
outward side. Then they ran its looped lines forward and under the
ship’s hull. The Malabart sheered off to give play, and the men fell to
the lines, heaving and tugging, as the mat went over the side and
submerged itself at the point indicated by Beltramo. Throughout
their work, running, and pulling, and hauling, that same air of
noiselessness, of low-spoken orders, was maintained. In the same
muffled silence, filled only with sounds of movement, the other mat
was fixed on the starboard side and drawn taut, and the officers,
listening intently down the hatchway, were encouraged when the
sounds of swirling and gurgling were no longer audible.
A huge cable was brought across from the Malabart, fixed through
the for’ard bits. The Malabart’s screw turned, and she slowly moved
ahead until she took the strain of the tow and headed back for the
shelter of the islands where she had lain in wait. Down on the
engine-room steps Captain Eli held his torchlight against a water
mark and slowly his face lost its grimness. His eyes twinkled when
he saw the ship was no longer taking in an appreciable or dangerous
quantity of water. He mentally estimated the time, and muttered:
“We’ll make it, sure, unless she springs another leak, or the mats
fail!”
Neither accident came, and in the dawn the Malabart towed her
salvage into the sheltered waters, slacked off and came alongside as
the Rhodialim’s anchors splashed into the sea. Drake, going across
to his own ship, where the cook was serving out hot mugs of coffee,
gulped one, and eyed the remnants of the two packing cases that
Forbes had opened on the Malabart’s deck. Two centrifugal pumps,
stocky and powerful, squatted there in the midst of the confusion,
and the engineer was directing the fitting of the steam lines.
“We’ll lash the ships alongside. It’s safe, I think, and it’s so still in
these waters they’ll not chafe,” Drake said to Catlin and the engineer.
And that maneuver was quickly effected. The pump suckers were
hauled across and splashed into the half-drowned hull of the
salvaged ship and a few minutes later two great streams of water
were pouring steadily into the sea. When daylight came the diving
apparatus was planted on the Rhodialim’s deck, and, guided by a
water torch, the man who had abandoned diving made a descent,
found the sea cocks and closed them. And now the salvage was
practically assured.
It was nearly noon when Drake said to Catlin:
“Now we’ll go below and get at the bottom of this business. We’ll
have a little chat with that second mate we’ve got trussed up.”
They brought the man up to the deck. He was sullen, cowed, and
palpably frightened. Drake regarded him coldly for a full minute,
frowning before he said:
“We brought you up to get at the truth of this. Why did you come
back to the ship? Did Morris send you?”
The man started to evade, to stammer, to make palpably false
statements until Drake threatened with:
“Stow that guff! The only chance you’ve got is to come across
with a clean yarn. If you do that, you’ll get away clean. Now quit
your waving the hook, or back below you go, until I can hand you
over to the shore police in Pirzeus. If it suits you better to talk Greek
—— Christophe, come here and tell me what this man says. I want
to get it straight.”
Christophe came, added his own urgings to overcome the man’s
reluctance, and then listened with a dry grin to a voluble confession.
Now and then he interrupted with a question, and although Drake
understood the gist of the mate’s words, Christophe finally turned
and in his own way told what he had learned.
“Thees man, he think maybe he and these other mens can maybe
get lots of little things like chronometers and glasses and such what
left behind; so after lost Captain Morris boat in fog, they row back
see if she still afloat, and come aboard. He swear he not know
anything about how she sink on purpose. Engineer what Morris frien’
run on deck, yell she sprung big leak, and Morris make fuss, and
then say no hope and mus’ take to boats. When these man come
aboard and find you, they thinks maybe ship not sink after all, and if
they can get her back they make lot of money for save her. So, fight
like hell. He swear that all he know. Maybe he spik truth, I think so.”
Drake stared at the man for a moment. Then, with apparent
irrelevance, he asked Christophe:
“How do people go by land from Nauplia to Pirzeus, and how long
does it take?”
“Road over the mountains, sir. Easy go. But take maybe two,
three days.”
“Telephone, I suppose?”
“Sure, captain, sir. Nauplia fine city. One time capital of Greece
and——”
“Good! You tell this man we’re going to keep ’em aboard the
Malabart until we get ready to make it to Piræus, and that nothing
will happen to them, unless they try to leave before we get ready for
them to go.”
The mate of the Rhodialim understood, and broke into profuse
promises; but to make certain that they could not escape, Drake had
all the boats of the Malabart brought around to the salvaged ship,
moored, and the oars taken away, before he liberated his battered
prisoners and told the cook to feed them.
Catlin was still wondering what Drake had in mind when, a few
days later, the Rhodialim was ready to put to sea under her own
steam. Then Drake said to his mate:
“Mr. Catlin, you take Beltramo and whatever scratch crew you
need for the engine room and ship, and go aboard the Rhodialim
and follow us to Piræus; but first have the boat that scum came in
brought around, chuck in grub and water enough to take them to
Nauplia, then chuck them in after it and tell ’em to go and be
damned to ’em.”
The mate’s wonder ceased on the day when the two ships came
to the crowded docks of the Greek seaport, amid the babbling
exclamations of those who recognized the salvaged ship. Drake
called to Catlin to accompany him, and they walked from the docks
to make their official reports.
“We ought to get a neat bit of salvage money out of this trip,”
Catlin said.
“We’ll get that all right. And I’m going to cut it up—half of it,
anyhow, among every man that was with us. Christophe ought to
get a good chunk, and so should Beltramo.”
“But what I can’t get is why you held that gang of beach combers
until we were ready to come here,” Catlin said, observing that The
Old Hyena was in high good humor.
“I waited to give Bill Morris and his pals time to get back and
swear to their story of how the ship was lost,” he said. “It’s about
time they, as well as Hakim & Letin, were put out of business.”
Transcriber’s Note: This story appeared in the February 4, 1928

issue of The Popular Magazine.
*** END OF THE PROJECT GUTENBERG EBOOK SALVAGE ***
Updated editions will replace the previous one—the old editions will
be renamed.
Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying
copyright royalties. Special rules, set forth in the General Terms of
Use part of this license, apply to copying and distributing Project
Gutenberg™ electronic works to protect the PROJECT GUTENBERG™
concept and trademark. Project Gutenberg is a registered trademark,
and may not be used if you charge for an eBook, except by following
the terms of the trademark license, including paying royalties for use
of the Project Gutenberg trademark. If you do not charge anything
for copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.
START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK
To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.
Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund
from the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.
1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law
in the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name associated
with the work. You can easily comply with the terms of this
agreement by keeping this work in the same format with its attached
full Project Gutenberg™ License when you share it without charge
with others.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the
terms of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.
1.E. Unless you have removed all references to Project Gutenberg:
1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears,
or with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.
1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.
1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning
of this work.
1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.
1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1
with active links or immediate access to the full terms of the Project
Gutenberg™ License.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or
expense to the user, provide a copy, a means of exporting a copy, or
a means of obtaining a copy upon request, of the work in its original
“Plain Vanilla ASCII” or other form. Any alternate format must
include the full Project Gutenberg™ License as specified in
paragraph 1.E.1.
1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.
1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:
• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and

personal growth!
ebookgate.com

Stochastic Processes and Their Applications 1st Edition Frank Beichelt (Author) All Chapters Instant Download

Uploaded by

Copyright:

Available Formats

Stochastic Processes and Their Applications 1st Edition Frank Beichelt (Author) All Chapters Instant Download

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stochastic Processes and Their Applications 1st Edition Frank Beichelt (Author) All Chapters Instant Download

Uploaded by

Copyright:

Available Formats

Visit https://ebookgate.

com to download the full version and

Stochastic Processes and Their Applications 1st

_____ Click the link below to download _____

Explore and download more ebooks at ebookgate.com

Stochastic Methods and their Applications to

Theory and Statistical Applications of Stochastic

Probability Theory and Stochastic Processes with

Basic Stochastic Processes 1st Edition Devolder

Large deviations for stochastic processes Jin Feng

Adventures in Stochastic Processes 1st Edition Sidney I.

Introduction to Stochastic Processes Second Edition

Discrete Stochastic Processes and Optimal Filtering Second

Frank E. Beichelt and L. Paul Fatti

Boca Raton London New York

CRC Press is an imprint of the

Catalog record is available from the Library of Congress

First published 2002 by CRC Press

Taylor & Francis Group

© 2002 by Taylor & Francis Group, LLC

No claim to original U.S. Government works

ISBN-13: 978-0-415-27232-2 (hbk)

Visit the Taylor & Francis W eh site at

and the CRC Press Web site at

Symbols and Abbreviations xi

5 Discrete-Time Markov Chains 137

6 Continuous-Time Markov Chains 169

7 Wiener Processes 257

7.2 First Passage Times 263

8 Spectral Analysis of Stationary Processes 293

Appendix 1 Landau O rder Symbol 312

Johannesburg, September 2003 Frank E. Beichelt, L. Paul Fatti

{M(t), f > 0} point process, counting process

4) For any events A and B,

Generally, the relative frequency of A tends to P(A) as n increases:

In particular, if k = n, then the independence of the A/ implies

is called an exhaustive set of events if

is called total probability rule and the formula

is called Bayes' theorem.

1.2 R andom V ariables

Any distribution function F(x) has the following properties:

1.2.1 Discrete Random Variables

The set {Po>Pi>-··} characterizes the probability distribution of A. Since A must

F(x) is a piecewise constant function with jumps of size p, at x = x, - 0. There­

respectively, provided the sums exists. The arithmetic mean of n realizations of X

The n th moment μ« of A is the expected value of A” :

Table 1.1 Probability distributions of discrete random variables X

1.2.2 Continuous Random Variables

A random variable X with distribution function F(x) is called continuous if F(x)

Rx) is called probability densityfimction of X (briefly: probability density or only

provided the integrals exist.

provided the integrals exist. A useful representation of Var(X) is

Table 1.2 Probability distributions of continuous random variables X

1.2.3 Nonnegative Random Variables

can be written in the form

In the remainder of this section, X is assumed to be a continuous nonnegative ran­

An important example of a nonnegative random variable is the lifetime of techni­

Hence, according to (1.6),

Figure 1.1 Illustration of the residual lifetime

Formula (1.8) then yields the desired result:

_ Click the link below to download _

F(x) is a piecewise constant function with jumps of size p, at x = x, - 0. There

In the remainder of this section, X is assumed to be a continuous nonnegative ran

An important example of a nonnegative random variable is the lifetime of techni

Obviously, the conditional failure probability is increasing with increasing t. Spe

Thus, λ = (1/Θ)Ρ. A disadvantage of this representation is that λ is not a scale pa

Assuming the existence of the densities ίχ.(χΐ), independence of the Y, is equiva

The density and distributions function of a sum Z=X\ +X 2 + — +Xn of n inde