Research projects
Agent-based modelling: methodological
NWO Digging Into Data MIRACLE (2014-2015) - project ‘MIning Relationships Among variables in large datasets from CompLEx systems’, €100.000
Social scientists have used agent-based models (ABMs) to explore the interaction and feedbacks among social agents and their environments. The bottom-up structure of ABMs enables simulation and investigation of complex systems and their emergent behavior with a high level of detail; however the stochastic nature and potential combinations of parameters of such models create large non-linear multidimensional “big data,” which are difficult to analyze using traditional statistical methods. Our proposed project seeks to address this challenge by developing algorithms and web-based analysis and visualization tools that provide automated means of discovering complex relationships among variables. The tools will enable modelers to easily manage, analyze, visualize, and compare their output data, and will provide stakeholders, policy makers and the general public with intuitive web interfaces to explore, interact with and provide feedback on otherwise difficult-to-understand models.
I work on this project together with Dr. Ju-Sung Lee and Prof. Alexey Voinov.
More: http://wici.ca/new/research/digging-into-data-did-research/
Collaborating institutions: University of Waterloo (Canada), Arizona State University (USA), University of Dundee (UK)
Research output: see paper #23
Growing spatial and socio-economic dynamics in empirical agent-based models using artificial intelligence algorithms (2013 - 2017)
Agent-based models (ABMs) often imply various learning algorithms (LA) for agents to form expectations and opinions about the environment, prices, and future trends of other variables of interests. In ABMs the latter is an emergent result of social, economic and spatial interactions between heterogeneous adaptive agents. Various learning techniques, often based on artificial intelligence (AI) principles are being used in ABMs. However, the choice of a particular method is often ad-hoc and is driven by subjective preferences of a researcher. Yet, the type of a LA would largely impact the way expectations are formed and, thus, model emergent results of models. Despite this fact, formal quantitative research that would explore the impacts of the choice of LA algorithm on the model outcomes is lacking.
Spatial ABMs often use naive deterministic algorithms, which are rule-based or condition-based, to simulate behavioral change in agents. While agents in spatial ABMs are sometimes endowed with memory the actual learning in AI style is rarely implemented. The endogenous switching of expectations formation strategies using LA is underdeveloped in spatial ABMs.
The PhD student - Ms. Shaheen Abdulkareem - focuses on testing the effects of various AI algorithms on the emergent outcomes in spatial agent-based models.
We collaborate with Ellen-Wien Augustijn and Prof. Ahmed Tahir within this project.
Agent-based models (ABMs) often imply various learning algorithms (LA) for agents to form expectations and opinions about the environment, prices, and future trends of other variables of interests. In ABMs the latter is an emergent result of social, economic and spatial interactions between heterogeneous adaptive agents. Various learning techniques, often based on artificial intelligence (AI) principles are being used in ABMs. However, the choice of a particular method is often ad-hoc and is driven by subjective preferences of a researcher. Yet, the type of a LA would largely impact the way expectations are formed and, thus, model emergent results of models. Despite this fact, formal quantitative research that would explore the impacts of the choice of LA algorithm on the model outcomes is lacking.
Spatial ABMs often use naive deterministic algorithms, which are rule-based or condition-based, to simulate behavioral change in agents. While agents in spatial ABMs are sometimes endowed with memory the actual learning in AI style is rarely implemented. The endogenous switching of expectations formation strategies using LA is underdeveloped in spatial ABMs.
The PhD student - Ms. Shaheen Abdulkareem - focuses on testing the effects of various AI algorithms on the emergent outcomes in spatial agent-based models.
We collaborate with Ellen-Wien Augustijn and Prof. Ahmed Tahir within this project.