Order this book
AIT Austrian Institute of Technology GmbH
Apart from the colourful mix of applications gathered in this volume, some overarching current developments and research endeavours in the field of social simulation can be identified. The first one to be highlighted is the shift from hypothetical and theoretical models towards empirical agent-based models (ABMs) or simulation models in general, and by this, the aim to build models not only reflecting real-world systems but also using real-world data of many kinds during the design and development of the models. For instance, in Chapter 2, web-based questionnaire surveys are used to empirically initialize, calibrate and validate the ABM, the use of demographic data to initialize the agent population (Chapter 12), the application of field-study data during the validation process (Chapter 20), as well as the use of experimental findings in the design of the model (Chapter 16).
The second direction of current research endeavours addressed by the book is the integrated use of simulation models, networks and social network analysis. In the context of financial networks, Chapter 5 analyses the flow of financial capital to evaluate the security or vulnerability of financial systems. Chapters 6 and 7 are both dedicated to opinion dynamics; whereas, in Chapter 6 a co-evolutionary process in the context of voter dynamics is modelled, in Chapter 7 the authors study the dynamic consequences of the topological network structures (representing different structures of social spaces) underlying social simulations.
A third current topic focuses on applications of participatory modelling, i.e., stakeholder engagement in the modelling and simulation process. A particular challenge in this respect, which is explicitly dealt with in Chapter 18, is how to involve decision-makers, who mostly have no, or limited understanding of an agent-based modelling language. To cope with this issue, the authors propose a hybrid approach of agent-based and gaming simulations. Chapter 17 presents an ABM targeted at managers and employees in a business environment, assisting them in decision-making processes.
Fourth, the use of simulation models, especially ABMs, as a tool for ex-ante evaluation, forecasting and prediction gains more and more importance, especially with stakeholders and decision-makers (e.g., policy makers, industry representatives) striving for alternative ways to a priori assess possible impacts and effects of interventions. The model presented in Chapter 10 aims to simulate and evaluate policies for small and medium-sized enterprises, by means of a macroeconomic model. Also, on a macroeconomic level, in Chapter 11 the effects of a corporate tax rate on GDP under different firm characteristics are evaluated. In the field of health care, the results derived from the model presented in Chapter 13 are intended to demonstrate possible governmental health care interventions.
Finally, the book touches some interesting methodological aspects, such as using ABMs with genetic algorithms for calibration and validation (Chapter 19), handling large amounts of data by means of Principal Component Analysis (PCA) within a simulation framework (Chapter 1), and the use bootstrapping for agent initialization (Chapter 2). Moreover, Chapter 21 proposes a simulation development process to compare and re-implement existing models in a chosen problem domain.
Overall, this volume provides highly interesting insights in new and innovative domains of computational social science, such as consumer-generated advertising (Chapter 2), e-government (Chapter 3), and vulnerability of the financial system (Chapter 5). However, this richness of applications is – at some points – also a slight drawback of the volume, being a melting pot of quite heterogenous articles, making it difficult for the editors to establish a completely homogeneous common thread for the volume. However, given the fact that the book is mostly a summary of conference proceedings, this is not that surprising, and not a big problem in general, as the reader is aware of it from the beginning.
In essence, cross-reading the volume as a whole – without a doubt – it discusses important advances and trends in computational science. The book reflects the increasing awareness for calibration and validation issues and highlights the use of real-world data as important means to this regard. Moreover, integrating statistical tools (e.g., PCA, bootstrapping) and ABM, constitute important advances that elevate ABMs and simulation models in general, to a new level in social sciences by increasing the credibility of such models immensely.
To conclude, this volume is a typical post-conference proceeding, primarily addressed to researchers that are interested in the state-of-the-art across various fields of simulation applications in social sciences. As the book was published in 2014, some of the “advances” might unfortunately already be outdated by now.
Return to Contents of this issue
© Copyright JASSS, 2018