See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275758413

Logistics Systems Analysis (second edition)

Article  in  Journal of the Operational Research Society · April 1997

DOI: 10.2307/3010274

CITATIONS READS

6 1,834

2 authors, including:

Said Salhi
University of Kent
175 PUBLICATIONS   5,879 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

heuristic search View project

Soft Computing for Modeling and Control of pH in Reactors View project

All content following this page was uploaded by Said Salhi on 12 December 2016.

The user has requested enhancement of the downloaded file.

Journal of the Operational Research Society (1997) 48, 449±452 #1997 Operational Research Society Ltd. All rights reserved. 0160-5682/97 $12.00

Book Selection
Edited by JOHN M. WILSON

PINTER JD. Global Optimization in ActionÿContinuous and Lipschitz Optimization: Algorithms,

Implementations and Applications 449
DAGANZO CF. Logistics Systems Analysis (second edition) 450
SMULDERS SA. Control of Freeway Traf®c Flow 450
DERR KW. Applying OMT ± A Practical Step-by-Step Guide to Using the Object Oriented Modelling Technique 451

Global Optimization in ActionÿContinuous and selected ®elds of expertise of the author. This fact repre-
Lipschitz Optimization: Algorithms, Implementations sents both a strength and a weakness of this volume.
and Applications The book is organized in four parts. Part One contains a
short overview of theory and methods of global optimiza-
Pinter JD tion; Part Two deals with the main research topic of the
Kluwer Academic Publishers, Dordrecht, 1996. xxvii ‡ 478 author, namely the analysis and the development of global
pp. £127.00. ISBN 0 7923 3757 3 optimization algorithms for Lipschitz continuous functions,
both in the unconstrained and in the constrained case. Part
Most books on mathematical programming, especially Three deals with some lower level details on the imple-
those dealing with nonlinear optimization, share two mentation of global optimization methods and on exten-
important defects which prevent them from being really sions, in particular to stochastic problems. Part Four, which
useful outside the advanced theoretical research groups. accounts for more than half the book, deals with several
The ®rst of these is the absence of real applications: almost applications, ranging from systems of equations, to data
all such books start assessing the enormous practical classi®cation, aquifer model calibration, the management
relevance of mathematical programming, sometimes of industrial wastewater and of river pollution and lake
include a list of real-world problems, but, after a few eutrophication. This is, in my opinion, the most peculiar
pages, they deal exclusively with the theory of optimization and interesting part of the book; it is very instructive to see,
and of optimization algorithms. The second important as the title says, optimization models in action in dif®cult
defect, shared by the vast majority of books dealing and socially relevant environmental problems.
with non-linear optimization, is that they deal just with No book is without defects, and this one is no exception.
local optimization: non convexity and multi-extremality Often the style of writing is confused and the book looks
are considered too pathological and intractable to like the result of merging several papers; moreover, I found
deserve attention. Unfortunately most real world problems quite disturbing the frequent insertion of parentheses every
do not possess any kind of convexity. (Think, as a page has at least one; in some pages I counted more than
simple example, of the huge class of discrete optimization 10! The author seems a little bit too con®dent on his
models). algorithms and the software he implemented: very few
This book, at least in part, tries to overcome the above comparisons are made with alternative approaches. Last,
dif®culties: it is a book on global optimization (the 6th in but not least, the price for this volume is extremely high
the series on `Nonconvex Optimization and its Applica- and, given the current budget limitations in most univer-
tions' published by Kluwer) and it extensively reports on sities, I wonder how many libraries will be willing to buy it.
applications, with a special emphasis on applications to Of course it is always easier to speak of defects rather
environmental problems. Most of the book is related to than highlight the positive contributions. I must admit that
research carried out by the author, with many other this book, thanks especially to Part Four, is really interest-
researchers, during the last few years. The book, thus, is ing and stimulating both for reference and as a research
not a broad spectrum of the state of the art in global monograph for the modelling of complex environmental
optimization and its applications, but is more a detailed problems. If supplemented with some good reference to
account of theory, algorithms and applications in some other algorithmic approaches (like, for example, some of
450 Journal of the Operational Research Society Vol. 48, No. 4

the books already published in the same series by Kluwer), This problem can be seen as combined location inven-
it could very well form the basis of an advanced post- tory routing which is an interesting and challenging
graduate course on nonconvex optimization and its appli- research topic. The last chapter reviews the many-to-
cations. many distribution problem with multi-commodities (note
that the case of the single commodity can be transformed
Universita di Firenze Schoen, F into the one studied in the previous chapter). This type of
distribution problem arises in airlines, postal services, less-
than-truckload carriers, railroads, etc. In this chapter break
bulk terminals (BBT) as well as transshipment terminals
Logistics Systems Analysis (second edition) (TT) are needed. BBT are introduced to group local
collections before line haul takes place, whereas TT serve
Daganzo CF as a consolidation terminal as de®ned in the previous
Springer, Berlin. 1996. xvi‡341 pp. DM128.00. chapter. It has been shown that a hierarchy of terminals
ISBN 3 549 60639 4 does usually enhance the transshipment bene®ts.
This book differs from traditional OR approaches in that
Logistics is the science that studies the movements of detailed description of problems and solutions which use
products from production to consumption. This monograph detailed data are avoided and only ef®cient approximate
concentrates on a part of the chain, namely transportation solution procedures requiring as little information as possi-
and inventory. This book is the fruit of over 20 years ble are introduced. The author emphasizes the usefulness of
experience of Professor Carlos Daganzo (University of adopting a two phase approach when tackling real logistical
California, Berkeley) in teaching and researching in this problems, where in stage 1 quick approximate methods are
topic. The book is well organized as it starts introducing the used to limit the number of options and to provide an idea
easier distribution problem namely the one-to-many desti- about the outcome and in stage 2 ®ne tuning is applied as
nations and then extends to the more complex problem only a few options are to be evaluated. This methodology is
which is the many-to-many distribution problem with adopted throughout the book.
transshipments and breakbulk points. This is, in my opinion, a well structured and documented
This book consists of six chapters; the ®rst two introduce book that certainly re¯ects the new era of logistics.
preliminary ideas whereas the remaining four describe the Academics in OR=MS as well as business=manufacturing
theory as requested by complex problems. Suggested exer- managers and consultants will ®nd this book full of
cises are given at the end of each chapter and ®gures, constructive ideas that can be useful in either teaching or
totalling 52, are also provided to illustrate some of the in assisting management in the decision process.
concepts used in the theory.
University of Birmingham Salhi S
The usefulness of simple models is outlined in Chapter 1
and the various costs existing in a logistics operation are
described in Chapter 2. The next chapter introduces the
one-to-one distribution problem with time dependent Reference
demand. The continuous approach of Newell1 is adopted
1 Newell GF (1971). Dispatching policies for a transportation
and some graphs are given to help understand this approx- route Trans Sci 4: 91±105
imation method which is adapted in more complex
problems. Chapter 4 addresses the one-to-many distribution
problem without transhipment terminals. The combined
Control of Freeway Traf®c Flow
inventory routing problem is tackled using (i) the dispatch
time and the corresponding delivery lot sizes (found by Smulders SA
inventory cost and route length approximation) and (ii) the
design of delivery routes at each dispatch time. A staged Stichting Mathematisch Centrum, Amsterdam, 1996,
production schedule based on production subregions is pp 158, Price 40 D¯, ISBN 90 6196 451 2
proposed to reduce total logistic cost by reducing inventory
cost at the origin. In Chapter 5 the problem is extended to This book is described as a monograph and is part of a
include the possibility of having transshipment terminals. series produced by the Centre for Mathematics and Compu-
The main role of these transshipment points (consolidation ter Science in Amsterdam. The book is a focused descrip-
terminals) is to transfer freight quickly using different types tion of four years of research into this topic area by the
of transportation modes with little allowance made for long- author, under guidance from others. In some respects it sits
term storage. An integrated approach that combines inbound between a traditional Ph.D. thesis and a `proper' book in
and outbound transportation schedules to reduce terminal terms of its presentation. The high mathematical content
holding cost is put forward with an illustrative example. and the presence of few worked examples and no chapter
 Book Selection 451

exercises distances the book from the traditional text book sections. A different derivation of the standard deviation
format. was required in very dense traf®c conditions.
The book is arranged in seven chapters entitled: Intro- By this stage of the book, a large and complex modelling
duction; Modelling of Freeway Traf®c Flow; Simulation of system has been constructed to describe the operation of a
Freeway Traf®c Flow; Analysis of Freeway Data; Filtering freeway system. A further requirement is a ®lter which
of Freeway Traf®c Flow; Control of Freeway Traf®c Flow; takes the measurement data of passing times and vehicle
and Conclusions. speeds and produces a minimum variance estimate of the
The ®rst chapter is a gentle introduction to this specialist traf®c state, primarily density. The operation of this ®lter is
topic area, covering the development of freeway infrastruc- tested for bias, robustness with respect to input data,
tures (primarily in the context of the post-war Dutch sensitivity to its parameter values and temporal stability.
experience); modern resistance to further expansion in These tests are made against both simulated and actual
this infrastructure and the resultant problems caused by data.
the mis-match between supply and demand for freeway The ®nal chapter of substance, Chapter 6, considers the
travel. The prime problem is the creation of congestion use of the model to assess the effectiveness of variable
which causes delays, frustrations and accidents. This intro- speed message signs to homogenise the state of traf®c. This
duction is more detailed and expansive than you would smoothing is expected to produce a more stable and
normally ®nd in the con®ned spaces of a conference or potentially less congested traf®c state. A control policy
journal article. where variable messages are used to control traf®c densities
The second chapter proposes a mathematical traf®c on freeway sections is put forward. The control policy
model for freeway traf®c which is an enhancement of an would only become active when the ®lter estimates a
earlier postulated model. In these models the freeway is certain density threshold is exceeded and is deactivated
divided into sections, with density (measured in vehicles when the density is estimated to fall below a lower thresh-
per lane per kilometre) and speed (km=hour) measures old level. This use of two thresholds is to help to reduce the
associated with each section. A set of state (or stochastic occurrence of hysteresis. The results suggest that the on-set
differential) equations describe how these measures change of congestion can be delayed and the ¯ow increased if such
in a section depending on its own measures and those in a control policy is adopted.
upstream and downstream sections. The mathematical The concluding chapter summarizes the work in the
derivations included in this chapter are challenging, requir- other chapters and suggests areas where the material in
ing the reader to concentrate or seek out the background the book may be applied and enhanced.
references, neither of which should be a deterrent to the The index for this book is brief at only 22 entries,
dedicated, interested reader. Casual readers may ®nd this re¯ecting the fact that this book is not a book which is to
chapter off-putting, however. be dipped into to clarify a point of interest to the reader.
In Chapter 3 the partial model developed in Chapter 2 The book has a well de®ned structure and only by follow-
along with assumed parameter values are tested in three ing this structure can a full understanding be gained.
arti®cial `operating' environments: low density traf®c In summary, this is not a text book on methods of
conditions, dense conditions and highly congested condi- freeway control, it is the exposition of a body of work on
tions, but with more interest in the ®rst of these two, more a single theme by the author. This book is best viewed, as
typical, environments. These tests highlighted some unrea- the author describes it, as a thesis. No alternative meth-
listic behaviour in the models, for example uneven distri- odologies are explained in any detail, but references are
bution of traf®c amongst the sections, and some sections supplied. I think the book would have bene®ted from a
having both high speeds and high traf®c densities. Correc- more sedate and gentle route through the mathematical
tive action included the re®nement of some model para- derivations. This is especially so with Chapter 2, given the
meters and respeci®cation of various relationships (for importance of this chapter for the understanding of the rest
example between speed and density) in the model. These of the book.
modi®cations enhanced the performance of the model. The
congestion environment was only considered as a second- University of Leeds Clark SD
ary item of interest.
A ®nal requirement, that of a distribution relationship to
describe the probability distribution of the passing speed of
traf®c is covered in Chapter 4. This is essentially an Applying OMT ± A Practical Step-by-Step Guide to
empirical exercise, using data collected from the Dutch Using the Object Oriented Modeling Techniques
Motorway Control and Signalling System. These investiga-
By Derr KW
tions suggested that the best approximation was a normal
distribution whose parameters were functions of the traf®c SIGS Books, New York, 1996, ISBN 1-884842-10-0
densities and mean speeds in the current and neighbouring pp xv‡542 $44.00
 452 Journal of the Operational Research Society Vol. 48, No. 4

Applying OMT is unique in that it takes you from the start There is a downside however to this book. In approxi-
to the ®nish of a single modelling problem by the creation mately late 1996 this became somewhat dated as the highly
of an electronic ®ling system using the Object Modelling vaunted integrated method fathered by Booch and
Technique (OMT). It also provides you with the code for Rumbaugh is scheduled to be released by Rational Soft-
the solution in both SMALLTALK and C ‡ ‡ . You move ware. Since Rumbaugh is the main spokesman for the
from a statement of the problem to the actual code required OMT, do not expect too many changes, so even though
to implement the OMT model which evolved. In this sense there is some downside to Applying OMT, it may not be as
the book is unique and highly recommended for people just severe as going from WINDOWS 3.1 to WINDOWS 95. I
beginning to use OMT. After having used OMT for over 5 recommend this book for all skill levels involved in Object
years without this book, I really appreciate the frank Oriented Analysis and Design.
manner in which the author dissects the problem and then
proceeds to solve it. I learned how to do just that after 5
years of fumbling and burning the midnight oil. It would
have been most helpful to have had this book 5 years ago. University of Maryland Leake C

View publication stats