Automated Planning: Theory and Practice

Citation

Long, D. (2005), "Automated Planning: Theory and Practice", Industrial Robot, Vol. 32 No. 5. https://doi.org/10.1108/ir.2005.04932eae.001

Automated Planning: Theory and Practice

Morgan Kaufmann, Elsevier1-55860-856-7635 pp.21 May 2004£41.99Malik Ghallab, Dana Nau and Paolo Traverso

Keywords:Automation, Planning

http://books.elsevier.com/bookscat/links/details.asp?isbn=1558608567

Planning has been studied by artificial intelligence researchers for more than 40 years and has usually managed to merit a chapter in each new general book on artificial intelligence in that period. Unlike many other chapters in those books, planning has not been widely covered in more specialised texts. In Automated Planning, Ghallab, Nau and Traverso have rectified that and done so in an ambitious, accessible, broad and technically thorough style. As the authors say at the outset: “Planning is the reasoning side of acting. It is an abstract, explicit deliberation process that chooses and organizes actions by anticipating their expected outcomes.” This makes planning an activity that is highly relevant to automated control and the application of robotic systems, as well as in directing human activities such as rescue missions, logistical operations or evacuations. It is a subject that deserves to be better known amongst the community who build systems that act, but not yet with the capacity to reason about that action.

Planning research has been an exciting and fast-moving field for the past decade, seeing dramatic advances in range and scope of the technology, making this book a timely stock-taking of the current state-of-the-art. At over 600 pages its size alone is fair indication of just how much development the field has seen. However, the authors do not choose to focus primarily on theoretical research advances, but, instead, devote a large part of the book to applications, including illustrative case studies. This makes the book accessible to a wide audience and ties its relevance directly to practitioners in areas where automated planning is ripe for exploitation. Applications considered range from automated manufacturing and robotics through evacuation planning to space applications and intelligent game-playing.

The book is a text book – it is intended to provide an academic coverage suitable to support an advanced undergraduate or specialised graduate course in planning. It is therefore organised as a text book, including clear structural dependencies, precise technical definitions, examples and exercises. However, it is also suitable for use as a reference for researchers in the field and as a deep introduction to the field for those with a technical interest in its potential exploitation. It is certainly not a light read for those with a casual interest and parts of it demand more than a passing acquaintance with technical areas of computer science. Nevertheless, for the reader with a sense of curiosity about what automated planning might offer to their own discipline, this book is the best way to set out to find an answer. Ghallab, Nau and Traverso have filled a significant gap and done the planning research community a great service by writing this book. They have also offered an exciting opportunity to those who do not yet know the field, but for whom its findings are most relevant: those who grasp it will not be disappointed.

Derek LongDepartment of Computer and Information Sciences, University of Strathclyde, Glasgow, UK