Intensive Course on Agent-Based Modelling & Simulation for Business and Marketing

Sydney, July 2011

Overview:

The University of Sydney Business School hosted the first intensive course on agent-based modelling (ABM) in Sydney, Australia, from July 5th to July 28th 2011. The course brought a group of advanced graduate students together with academics interested in developing ABM skills for an intensive four-week course on ABM and its application to marketing, management and the social sciences more broadly. The course was led by Professor David Earnest, Old Dominion University, based on a similar course he offers there. The course combined lectures with hands-on experience in lab sessions, building simulations of business and social phenomena.

Aims:

Agent-based simulation models of business, psychological, social, economic, biological, climate and materials systems are revolutionising the way research is done in many disciplines. This intensive course for research students and faculty was designed to teach participants how to use agent-based simulation methods, including NetLogo, to further their research interests. Students who successfully completed the course were not only familiar with the theoretical and methodological foundations of ABM, they were also able to design, develop, test and implement ABM relevant to their research interests. While the application focus of the course was on marketing and management, researchers from other social science disciplines will also benefit.

Sponsors

  • Australian Research Council DP0881799 IF Wilkinson, RE Marks and LC Young
  • University of Sydney Business School
  • Australia New Zealand Marketing Academy
  • School of Marketing, University of Western Sydney
  • School of Marketing, University of Western Australia
  • Centre for Research in Complex Systems (CRiCS), Charles Sturt University
  • School of Economics, University of New South Wales

Course Leader:

David C. Earnest is an associate professor of political science and international studies at Old Dominion University in Norfolk, Virginia USA. He completed his Ph.D. in political science at The George Washington University in 2004. He also holds an M.A. in security policy studies from the Elliott School of International Affairs at The George Washington University, and a B.A. in political science from Stanford University. He teaches international political economy and political methodology. His substantive research focuses on the political incorporation of migrants in democratic societies, while his methodology interests are in the application of agent-based models to problems of international politics.

Dr. Earnest has published in leading journals, including World Politics and the International Studies Quarterly. He is author of Old Nations, New Voters: Nationalism, Transnationalism and Democracy in the Era of Global Migration (2008, State University of New York Press).

Previously he held an appointment as a Fellow in Political-Military Studies at the Center for Strategic and International Studies in Washington, DC, where he was a specialist in military technology, the defense industrial base, and transatlantic security relations.

Schedule:

The class meets all day every Tuesday and Thursday for four weeks, with additional assignments and exercises to be completed by the participants working individually and in teams. The course includes formal lectures and intensive computer lab workshops. There are also a limited number of spaces available for appropriately qualified students and faculty to audit the formal lecture components of the course. The formal lectures will be recorded for future broadcast and use. The first week will be held as a short course, providing an introduction to ABM and its applications, with additional demonstration of simulations and research seminars on Wednesday.

Background and Rationale:

The subjects of complexity and agent-based modelling (ABM) are fast growing areas of research and theory in many sciences, including business and the social sciences generally. This interest is indicated perhaps most clearly with the award of the Nobel Prizes in Economics to Elinor Ostrom in 2009 for her work in complex systems and the management of common pool resources, to Thomas Schelling in 2005 for his work on the unintended macro outcomes of micro behaviour in complex systems, and to Vernon Smith in 2002 on the role of social relations and the self-organising properties of market systems (not to mention the very early one given to Herbert Simon).

Many types of socio-economic systems as well as psychological, biological and ecological systems are increasingly being understood as complex adaptive systems in which order emerges in a bottom-up self-organising manner from the micro interactions taking place over time among the actors comprising the system, be they chemicals, animals, humans or ideas. Such systems also interact with other systems which buffet and challenge them. There are also feedback effects in which large-scale order feeds back to affect local interactions. There is now a growing industry of pop and scientific books and articles describing complexity and what it means for understanding, researching and operating in complex adaptive systems of different kinds. For more information and resources concerning complexity and agent-based modelling, visit the Agent-Based Computational Economics (ACE) site run by Professor Leigh Tesfatsion. The site covers material and resources related to learning and teaching, research and demonstration programs in a wide variety of disciplines, not just economics.

Simulation using agent-based modelling methods plays an essential role in the study of complex adaptive systems. This is because highly non-linear systems of this kind, though they can in principle be written down in mathematical form, are beyond traditional methods of solution. The only option is to compute the results of the rules over time using computers, hence the use of terms such as computational economics and computational social science. Agent-based models of business and social systems offer a way forward in studying the behaviour and evolution of complex systems that would not be possible otherwise. Such models are not the same as previous types of simulations such as Monte Carlo methods or System Dynamics. The models construct complex systems from the bottom-up by specifying the individual agents' characteristics and values, how they change and interact with other agents, and the environmental agents and conditions in which they operate.

Not only is it impossible to solve the highly nonlinear mathematical equations of motion of such systems with all the interactions taking place, it is also not physically, economically or morally possible to conduct the range experiments required to tease out the effects of different elements of such business and social systems. Once again we are led back to simulation and agent-based models that are now possible given rapid improvements in the power and accessibility of computers and developments in object-oriented programming methods.

A major factor limiting serious science and agent-based modelling of complex adaptive systems in marketing, business and the social sciences is a lack of adequate programming skills among researchers, who are generally not trained in such techniques, which are not easy to acquire. Or, rather, this was the case until quite recently, but as a result of developments in more user-friendly, agent-based modelling systems and programming languages researchers can more readily learn these techniques. The trouble is there are limited opportunities for them to do so. This intensive course is designed to provide a solution to this problem for researchers in our region. We plan to repeat it on a regular basis. There are few courses of this type offered anywhere in the world and none has so far been offered in Australia. David Earnest is one of the few academics willing and able to offer such a course.

107147