Reviewed by
Nicole Ronald
Swinburne University of Technology,Department of Information and Communication Technologies
Models for decision making often focus on one part of the system, for example,
the vehicles travelling along a road rather than why the vehicles are actually
travelling. Looking at the whole system, i.e., the interactions between the
participants, each in charge of their own smaller components, and between the
participants and the technological aspects of the system, could lead to a
better understanding of the system and therefore improved decision-making. This
book, written by members of the Energy and Industry group in the Technology,
Policy and Management faculty at TU Delft, Netherlands, describes how these
simulations could be developed.In part one of this book, by I. Nikolic and J. Kasmire, the theory behind complex systems, adaptivity, and agent-based modelling is discussed, setting the scene for the complexity of modelling interacting actors and technology. This contribution provides comprehensive background, explaining complex adaptive systems separately (systems, adaptivity, and complexity) and as a whole, before covering current research in modeling complex adaptive systems and agent-based modelling. Following this, a practice chapter by I. Nikolic, K.H. van Dam, and J. Kasmire, works through a methodology for the modelling of socio-technical systems step by step, using a greenhouse case study. It is not assumed that the reader has programming skills, so pseudocode is used and explained. The methodology itself is the result of a decade of developing models and updating and refining the methodology.
Part two contains five fully-worked through case studies, introduced by K.H. van Dam, I. Nikolic, and Z. Lukszo. Thes cases look at supply chains (by B. Behdani, K.H. van Dam, and Z. Lukszo), consumer lighting (by E.J.L. Chappin and M.R. Afman), CO2 policies and electricity generation (by E.J.L. Chappin and G.P.J. Dijkema), and mobile phone lifecycles (by L.A. Bollinger, C.B. Davis, and I. Nikolic). Each chapter works through the ten steps described in the practice chapter. The types of decisions being explored range from short-term (responding to unexpected events in supply chain management, such as a delayed shipment) to long-term (power production companies buying/selling plants and generating power over 50 years).
The book concludes with a chapter on future directions (by A. Chmieliauskas, C.B. Davis, and L.A. Bollinger) focusing on data sharing using Semantic Web technologies. While this is a useful chapter, a chapter reflecting on and comparing the case studies would have been appreciated to provide some closure to the book.
The methodology presented is sound. Some small criticisms, however, are the lack of relation to other modelling processes (even non-agent-based modelling processes) and also the lack of loopbacks in the process. Iteration is mentioned briefly in steps 4 (model formalisation: the research questions might need to be reworked at this stage) and 6 (verification: the code from step 5 might need revision). However, the chapter introduction and the case studies give the impression, especially to those new to modelling, that if these ten steps are worked through in order then a usable model will result. Lessons learned from and difficulties in the case studies also would have been beneficial to mention in each chapter; the difficulty of getting stakeholders on board with the model and its outputs is discussed in one case study, and suggestions for long-term stakeholder managements are made in the methodology chapter.
Overall, six case studies are presented using the methodology. This is ideal, especially for students to see the process applied to different models in a coherent manner. Graduate students would be able to begin designing a model after reading. However, they would need some programming skills or knowledge of a toolkit, such as NetLogo or Repast, in order to develop the entire model without assistance.
To conclude, this is a very useful book for those developing (and teaching) agent-based modelling of socio-technical systems and is a step towards developing more useful models for decision-makers.
Return to Contents
of this issue