October 2018
« Mar    

Amineh Ghorbani

Amineh Ghorbani

Amineh Ghorbani

I am an assistant professor at the faculty of Technology, Policy and Management (TBM) of Delft University of Technology. I obtained my M.Sc. in Computer Science from University of Tehran (Iran) (2009) and my PhD from Delft University of Technology (2013). MY main interests are institutions and institutional emergence in agent-based simulations, social simulation in general, policy analysis and decision support.
For more information, you can download my CV from here.

Current Research

During my PhD, I developed a meta-model for agent-based modelling of socio-technical systems. You can find me thesis here. The meta-model, called MAIA, describes various concepts and relations in a socio-technical system. I have been greatly inspired by Elinor Ostrom and her IAD framework.

For more information see the MAIA website.


Me as a Researcher

When I started my PhD at the Faculty of Technology, Policy and Management (TPM), one word sounded familiar to me: ‘agent’.

As a Master Student I was interested in the Artificial Intelligence line of research in the computer science domain. Besides learning and using bio-inspired algorithms (e.g., neural networks, ant colony optimization) to study consumer behavior, I carried out research in the multi-agent systems lab. I studied various agent communication protocols, languages, architectures and agent-oriented systems development methodologies.

After obtaining my Masters Degree, I applied for a position in agent-based modeling of complex socio-technical systemsat Delft University of Technology. I was greatly interested in applying my knowledge in agent research to real world application as the project description explained. However, I had no idea that the definition of agents the people at the faculty use, is quite different from what I had learnt. TPM was a faculty of social sciences including management and economics. The people who made agent-based models were highly intellectual in the social science domain but hardly knew the science behind agent research. Their models were extremely simple and still they managed to gain very interesting results from their simulations of ‘complex socio-technical systems’.

This is what made me think, that with the knowledge I have in agent-oriented research, I can advance the modeling paradigm for the social scientist community and apply this line of research to where it is actually being used and appreciated. However, there was one problem. The agent literature was highly complex for people who came from different domains. Therefore, they were quite reluctant to be introduced to this line of research.

By talking to many people in the faculty from different sections and departments, I found out that there are many socio-analytical theories being used independently from simulation. They had the power to explain social phenomenon and address complex problems. I studied many social and economic theories and frameworks to gain knowledge in this field of research. Some frameworks like the Institutional analysis and Development framework of ElinorOstrom (Noble laureate in Economics, 2009), even described a social system in quite a formal manner. This is where I started thinking; why not use the social science frameworks to build agent-based models.

In my PhD research, I developed a high-level modeling language called MAIA (Modeling Agent system based on Institutional Analysis) to describe socio-technical systems for social simulation. MAIA conceptualizes those systems, where thegoal is to understand the effect of some strategic decision making or policy on the behavior of thousands of individuals and the emergent outcomes. I tested my modeling language with several case studies and developed some tools that would actually help build simulations from a MAIA model.

The way MAIA is formalized opens possibilities for developing rich cognitive agents who are able to adapt to evolving social, institutional or physical environments and create emergent patterns of behavior. MAIA connects to existing ABM platforms such as Repast, taking their functionalities in parallel processing, analysis of results and graphical representations of simulations etc. on board.

One of the most valuable outcomes of my PhD research has been,to learn to work in a multi-disciplinary environment, talking to people with different expertise, putting their knowledge into use and actually transferring my own knowledge in a useful and understandable way.  I believe this research has opened new possibilities for research in complex adaptive systems since there is now a strong body of knowledge from three separate domains: social science, artificial intelligence and software engineering. I have also built a large network by attending conferences in all these domains and introducing new ideas to researchers from domains other than their own.

From what I have pursued during these years of research, I feel that I have just started a journey with many possibilities and ideas that I have high ambition to accomplish. Therefore, I see myself in the academia for many years to come, further developing and maturing tools for agent-based modeling because I believe that this tool is highly instrumental in studying human behavior.

© 2011 TU Delft