This interesting book presents a wide-ranging view of its subject. However, the sequence of chapters could have been organized better for increased readability.
After an introductory chapter on the issue of uncertainty, the second chapter provides a review of modeling techniques. I found this chapter easy to read and a very good starting point. However, the author seems to have neglected giving fuzzy logic and multi-valued logics their own sections, as he did in his review of techniques, such as belief functions and possible worlds. Nonetheless, references to fuzzy logic are provided in the references section at the end of the book.
Chapter 3 covers the issues of conditioning uncertainty measures: belief, possibility, and plausibility. This chapter seems to come too early. A discussion of updating belief would perhaps be better placed before a discussion of belief revision (chapter 9). Similarly, chapters 7 and 10 would be better placed at an earlier place in the text,
namely, after chapter 2.
While chapter 4 covers Bayesian networks, including different versions, chapter 5 covers the likelihood representation of uncertainty. Chapter 6 appears earlier than expected. It covers the issues of uncertainty in multiagent systems. Agents rely heavily on the notion of belief in their knowledge representation. Given that this book
is heavily biased toward notions of belief and plausibility in modeling uncertainty, I would expect it to cover multiagents. However, it would have been better to put the chapter on multi-agents at the end, where different elements of belief modeling and systems are covered.
Chapters 7, 8, and 9 make up the most logically continuous part of the book. They cover the logical aspects of modeling belief, starting from propositional logic, and leading to modal epistemic logic. Chapter 8 covers the logical aspects of belief inferencing, such as defaults, conditional logics, and counterfactuals. Chapter 9 is a continuation of chapter 8, covering belief revision. Belief revision is one of the most important topics of belief systems. Since agents live in dynamic environments, they need to reason about change, and adapt their belief sets to accommodate changes. Chapter 10 would have been better placed before chapter 7, since it covers the basics of first-order logic (FOL). The chapter goes beyond the usual coverage of FOL, however, with sections that discuss using FOL to reason about probability.
Chapter 11 is an essential part of the book. It covers topics on the statistical treatment of uncertainty, and how that links to beliefs. The extended discussion about random worlds in this chapter is interesting and quite refreshing.
The references section is extensive, while the exercises at the end of each chapter make it ideal for use as a textbook. The index is a valuable additional help.
I would recommend the book to postgraduate students who are studying reasoning about uncertainty and modeling.