My biggest criticism of this book relates to its title, which, to say the least, is confusing. The conjoining of “statistical,” “systems,” and “GPSS” implies that these terms are mutually exclusive foci of simulation, whereas GPSS is only one vehicle for the implementation of certain simulations. A much more apposite title would have been The design and implementation of discrete event simulations using GPSS.

The authors have set out to introduce the theory and implementation of discrete event simulations. To achieve this successfully, they say, a text should provide a theoretical basis for the simulation methodology, together with sufficient details of an implementation language. These should be exemplified together in case studies. On the whole, Karian and Dudewicz have been successful in their quest.

The text begins with an overview of computer simulation that attempts to place discrete event simulation in a wider context. The first chapter focuses on the single-server queue. Some aspects of the discussion, such as queue representation and sorting techniques, might seem superfluous in a text unconcerned with the implementation of simulation languages. This material reflects the somewhat low-level nature of GPSS, the language used throughout the book. GPSS is introduced in the second chapter.

The design and testing of a basic generator of uniformly distributed pseudorandom numbers is treated in depth in chapter 3. The succeeding chapter deals with the generation of random variables from a range of distributions commonly used in discrete event simulation. These chapters contain some sound advice, linked to information that should be understood by all discrete event simulationists. An important omission from these chapters is the consideration of the selection or automatic generation of random number generator seeds that ensure well-separated streams. If this is ignored, then the simulation results could be useless, however carefully the generators themselves have been selected.

GPSS is treated in successively greater depth in chapters 5 and 7. Each chapter has a number of examples, which illustrate most of the features described. Sandwiched between this treatment of the language is an important overview of the statistical design and analysis of simulations. Many texts discuss how to implement simulations, but unlike this one, few address the design of the simulations, so that the reader can determine the simulation time required to achieve the goal of the study. A few more worked examples might have been useful for the reader to whom statistics is not second nature. The book concludes with an extensive case study that brings into play a lot of the material introduced earlier.

The book comes with a disk (for DOS machines) that contains GPSS/PC, a limited version of GPSS for educational use. The examples illustrated in the text, or simplifications thereof, are included on the disk.

Before concluding, I must add my own caveat, implied earlier, that some other software tools for discrete event simulation are much easier, and consequently safer, to use than GPSS. Having said that, GPSS is more widely available than many alternatives. For those who wish to learn how to design discrete event simulations and implement them in GPSS, this book is a good resource. The book could be used in upper-division undergraduate or graduate classes, or by individuals, both professional and amateur. For the reader not committed to GPSS, much of the text will be superfluous.