A Java Library of Graph Algorithms and Optimization (Discrete Mathematics and Its Applications)

A Java Library of Graph Algorithms and Optimization (Discrete Mathematics and Its Applications)

By Hang T. Lau

Publisher: Chapman & Hall/CRC
Number Of Pages: 386
Publication Date: 2006-10-20
Sales Rank: 436122
ISBN / ASIN: 1584887184
EAN: 9781584887188
Binding: Hardcover
Manufacturer: Chapman & Hall/CRC
Studio: Chapman & Hall/CRC
Average Rating: 3

Because of its portability and platform-independence, Java is the ideal computer programming language to use when working on graph algorithms and other mathematical programming problems. Collecting some of the most popular graph algorithms and optimization procedures, A Java Library of Graph Algorithms and Optimization provides the source code for a library of Java programs that can be used to solve problems in graph theory and combinatorial optimization. Self-contained and largely independent, each topic starts with a problem description and an outline of the solution procedure, followed by its parameter list specification, source code, and a test example that illustrates the usage of the code. The book begins with a chapter on random graph generation that examines bipartite, regular, connected, Hamilton, and isomorphic graphs as well as spanning, labeled, and unlabeled rooted trees. It then discusses connectivity procedures, followed by a paths and cycles chapter that contains the Chinese postman and traveling salesman problems, Euler and Hamilton cycles, and shortest paths. The author proceeds to describe two test procedures involving planarity and graph isomorphism. Subsequent chapters deal with graph coloring, graph matching, network flow, and packing and covering, including the assignment, bottleneck assignment, quadratic assignment, multiple knapsack, set covering, and set partitioning problems. The final chapters explore linear, integer, and quadratic programming. The appendices provide references that offer further details of the algorithms and include the definitions of many graph theory terms used in the book.

Review:

A very handy collection

There are many well-written textbooks that cover the theory
and algorithms on graphs and combinatorial optimization.
Very few provide the computer code for the methods. This
book offers an extensive collection of Java programs in
this area. Each program is self-contained and can be used
independently through parameter passing. The drawback of
the book is that the coding style is not object oriented,
and the programs would be difficult to maintain. The
description of the methods and their implementations is
terse. Hence the book is not intended as a learning text.
But the library of programs is a very convenient handy
device for students and researchers in locating solutions
to classroom didactic problems in graphs and optimization,
which apparently is the main objective of the book.

Review:

Printout of Java programs

This is my third review; my two previous reviews have been removed. I have already notified Amazon about this fact

As I have stated, book is just a printout of Java program, without any explanation how program is doing what is doing, what are program limitations in terms of memory, time and complexity. Programming style is mostly Fortran IV like. Programs are without single line of comment and with non-intuitive variable names, what makes modificatios difficult or impossible. Book can be useful for somebody who needs "black box" library, doesn't need to understand programs and trusts the author that programs fave no flaws

Review:

Useful problem-solving tool

This library of ready-to-use programs is extremely useful. I have used the programs with very minimal effort in obtaining solutions to some graph optimization problems. Unfortunately the programs are not well documented; it would be a challenge to make modifications to the code. However, the library serves as an ideal black box tool in solving most of the pedagogical graph theory and optimization problems, especially well suited for users who are not of much concern for the underlying methodology and implementation.

Review:

A mere compendium of poorly written algorithms

There are so many problems with this book, it's hard to know where to begin. So I don't come across as all and only negative, I will first give it credit for gathering together, at least in name a large number of graph processing algorithms.

That said, here are the problems:

The book is just a catalog of graph algorithms with poorly done documentation and even worse actual code. To wit:

*Each algorithm is preceded by a very brief explanation of what it does and some of the issues involved. Suffice it to say that it's the sparsest and most minimal explanation imaginable; if you don't already understand the issues involved, you probably won't after reading the short paragraph or two that precedes each algorithm / method.

*There is but ONE class and every bit of functionality is contained in its own individual, single static method. This "design" causes not a few of the methods to literally run to a thousand and more lines and contain dozens and dozens of (cryptically named) member variables.

So for instance, if you are interested in planarity testing, there's a "method" called planarityTesting that takes four parameters and returns true or false.

All well and good until you actually look at that method and see declared 51 , that's fifty-one, member variables. Each of these variables has poorly chosen names like, "wkpathfind2" and "store2" and "store3" and of course "store4" and "sortptr1" and "sortptr2". I thought this tactic of vowel-conserving naming of variables went out with the 8 + 3 DOS naming convention. At any rate, the cryptic naming scheme combined with the lack of javadoc combine to render each variable's purpose completely opaque. This makes it all but impossible to relate the code to the underlying graph theory.

Then comes the code.

Imagine a thousand and more lines, literally page after page after page of streaming code, all one single method, manipulating these cryptic variables in virtually uncommented ways.

That is pretty much what you get with this book. One algorithm after another after another.

I would say the following:
1) the author codes as if from another time. There is NO object-oriented design to this code whatsoever. None. Zero. Zip.

2)The methods are hundreds or thousands of lines of what amounts to undocumented symbol manipulation. There is small chance to learn anything from this book with respect to relating the code to graph theory.

3) I can say that, having implemented many of the algorithms in this book myself prior to buying this book, the book has contributed nothing to my understanding and further, that already understanding the issues surrounding many of these methods, that is being a qualified reader, is NOT sufficient to allow the reader to follow and understand the algorithms.

4) If you only want to use the (static) methods to return a value or ascertain some property of a graph and you don't care to understand how it works or why it works, then perhaps you'll be happy with this book, but then , why not release the object code as blackbox library? If the code was never meant to be read, and there is no attempt at explaining graph theory as it relates to the code, then what of value is left for the reader?

5) Finally, if the purpose of the book is deliver a good "black-box" library, readers should know that the actual implementation of the graph "object" chosen in this book makes will make that problematic. The book uses an adjacency matrix to represent the graph, a well known data structure in graph theory. Unfortunately, this data structure has the following well-known problem: it is only suitable for the rare instance of dense graphs. The runtime performance and memory demands of this data structure make it unsuitable to any but very very small graphs. Most graphs are neither very very small nor very very dense, (as dense is defined in graph theory), and for that reason almost all graph drawing packages opt for a linked-list data structure to represent the graph.

This is solidly the worst book on this subject I have yet encountered. Amazon offers a number of alternative books, including the fine Graph Algorithms, Third Edition by Robert Sedgewick and Michael Schidlowsky, a book I have no connection with whatsoever and two authors who are otherwise unknown to me. Bundle of Algorithms in Java, Third Edition (Parts 1-5): Fundamentals, Data Structures, Sorting, Searching, and Graph Algorithms, Third Edition.pass: gigapedia.org

http://rapidshare.com/files/59598828/A_Java_Library_of_Graph_Algorithms_and_Optimization_1584887184.rar