AND Ming, I, W.: Editor'S

RECENT PUBLICATIONS
EDITOR'S NOTE: L i s t i n g o f a publication i n t h i s section i s f o r record

and reference only and does not c o n s t i t u t e an endorsement o f point of
view or advocacy o f use.
MANAGEMENT MODELS AND INDUSTRIAL APPLICATIONS OF LINEAR PROGRAM-

MING, Volume I, by Abraham Charnes and William W. Cooper. John Wiley and Sons, New York,
1961. xxiii + 467 pp. $11.75.
The world now seems to be deluged by books on linear programming, many of them less
than mediocre, and there is little sign of slackening in their rate of arrival. It is, then, an
event of some magnitude when an authoritative work on the subject is offered to us by two of
the leading contributors to the field.
Even by itself, this book, which is only the first of two volumes, is extremely wide
ranging. Input-output analysis, Koopmans' efficiency and pricing, nonlinearity, and a wide
variety of linear programming problems all make their appearance. Aspects of all of these
are treated in considerable depth. Almost every chapter contains at least touches of originality
and, indeed, some of them are taken, with slight revision, from the authors' original contribu-
tions to the journals.
The logic of the organization of the book is not entirely clear. There is no straight line
of development which the reader can easily spot and use as a guide to the volume. For exam-
ple, the input-output chapter does not seem to be connected with anything which follows or pre-
cedes it, and the same can be said of some other portions of the book. The level of exposition
also is uneven. Thus Chapter II, on the "stepping-stone" method of computation for the trans-
portation problem, succeeds in remaining as elementary as the authors mean it to be, while
many of the later chapters are likely to be considered extremely difficult by the novice. Inci-
dentally, a very minor point of criticism can be raised in relation to Chapter 11. The wording
of its introduction is likely to perpetuate the widespread misunderstanding of the relationship
between the simplex and the "stepping-stone" methods. For these both consist essentially of
the same algorithm; the latter is just the special version of the simplex procedure which
applies to transportation problems.
If any single theme can be said to characterize the volume it is the use of a device
whose essential simplicity is symptomatic of its brilliance. This is the procedure which the
authors call "complete regularization." It is a method for adjoining an artificial constraint to
any (consistent) linear programming problem in such a way that the modified program is guar-
anteed to have a solution. The point is best illustrated diagramatically: in Figure 1 we have
the (solution space) representation of a simple profit maximization linear programming prob-
lem with the is0 profit lines (the parallel dotted lines) as shown, and the feasible region repre-
sented by the shaded area. It is clear that point M represents the optimal solution-it is the
63
64 RE CENT PUBLICATIONS
A
!
A'
X 0 X
Figure 1 Figure 2
point of maximum profits. But in Figure 2, since the shaded feasible region is not bounded
from above, the problem has no solution-there is no limit to the amount of profits which can
be made, and therefore no "best" combination of values of x and y. Suppose now, that we
AA', to both problems, at a considerable distance from the ori-
adjoin an artificial constraint, -
gin. In the first diagram, since AA' lies outside the feasible region (it is a redundant con-
straint) it makes no difference to the solution. But in Figure 2, this constraint transforms the
problem from one which is unsolvable into one which has an optimal solution - P. Moreover, in
the nature of the case, the solution point, P, must lie on the artificial constraint line, AA'.
The authors then propose the following test:
1. To any linear programming problem, adjoin a sufficiently liberal artificial constraint.
(The authors point out that this can always be done, e.g., by employing the constraint x + y 5 K,
where K is the sum of the coordinates of all of the original extreme points.)
2. Solve this new "regularized" problem.
3. To test whether fie original problem possessed a solution, examine whether the
solution of the regularized problem lies on the artificial conetraint line.
This simple device is used with great skill in deriving all sorts of deeper results,
including general theorems on such problems as existence and degeneracy.
The reader will have gathered that the book will represent an important addition to the
library of anyone who works in these areas. That it is not an equally good vehicle for the stu-
dent's initial immersion in the subject is surely no ground for complaint. It is merely men-
tioned here to avoid misunderstanding of the nature of the work.
(Reviewed by William J. Baumol,

Princeton University)
* * *
~5U.S. GOVERNMENT PRINTING OFFICE : 1 9 6 2 0 -640210

AND Ming, I, W.: Editor'S

Uploaded by

Copyright:

Available Formats

AND Ming, I, W.: Editor'S

Uploaded by

Document Information

Original Title

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

AND Ming, I, W.: Editor'S

Uploaded by

Copyright:

Available Formats

RECENT PUBLICATIONS

EDITOR'S NOTE: L i s t i n g o f a publication i n t h i s section i s f o r record

MANAGEMENT MODELS AND INDUSTRIAL APPLICATIONS OF LINEAR PROGRAM-

(Reviewed by William J. Baumol,

~5U.S. GOVERNMENT PRINTING OFFICE : 1 9 6 2 0 -640210

You might also like