CRL Speaker Series - Dr. Przemysław Szufel

Understanding Complexity: An Introduction to Agent-Based Modeling Simulations

Agent-based modeling (ABM) has emerged as a powerful tool for simulating complex systems across diverse areas such as transaction markets, social interactions, epidemiology, and urban modeling. The development of such models enables us to understand complex dependencies between actors and to design optimal policies for scenarios such as market design, social cohesion, pandemic prevention, or congestion management. Two particular features of the ABM approach are its ability to capture the heterogeneity of actors (their unique properties) and to include network interactions (e.g., networks of transactions, information flow, or transportation) in the model. In the ABM approach, each agent observes the environment and independently makes decisions that lead to the optimal achievement of their objectives.This talk aims to introduce participants to the fundamental principles of ABM, highlighting its unique ability to model complex interactions between autonomous agents and their environments. Through several use case scenarios, we will discuss its key components, including agents, environments, and the rules governing agent behaviour and interactions.Subsequently, we will demonstrate how ABM has been effectively applied to understand and predict the behavior of complex systems in real-world scenarios. Additionally, we will delve into the methodologies for designing and implementing these models.Finally, some sample source code in Julia for a simple ABM will be presented and discussed to explain the typical implementation approaches.

About the speaker:

Przemysław Szufel is an Assistant Professor at the Warsaw School of Economics, Poland and an Adjunct professor at the Cybersecurity Research Lab, Toronto Metropolitan University. He does research in simulation modelling, graph analysis and optimization. His favourite tool for numerical computing in research is the Julia programming language. He actively participates in the Julia language community – is an co-author of one book, maintains 4 official Julia packages devoted to spatial data processing and graph analysis, and holds 2nd place on the StackOverflow portal answering Julia-related questions. Since 2018, he has been a regular speaker at JuliaCon, the major conference devoted to the development of the Julia language. Utilizing the Julia tool he has completed several industrial and academic data analytics projects in the areas of simulation, optimization, production planning, manufacturing, and process design.

Przemysław conducts workshops on advanced data analysis methods worldwide, he teaches a Julia course each year at the Massachusetts Institute of Technology, and has led Julia modelling workshops in, among others, University of California, Berkeley and Fields Institute located at University of Toronto.  Przemysław has also conducted training and workshops in the field of data analysis for the European Commission, the Central Bank of Poland, major telecoms, utility providers, banks, and strategic consulting companies.