Creation of Mesa, An Agent-Based Modeling Framework: Simulating the World in Python

Agent-based models enable researchers to create digital twins of real-world phenomena and test alternative policies computationally. This dissertation describes the creation of Mesa, an agent-based modeling framework in Python, one of the most popular programming languages in the world. Since 2014, I led the open-source software development effort called Mesa, articulating the basic design philosophy for agent-based models in Python and developing an efficient software architecture. Driven by volunteers, Mesa has facilitated research in the following domains: healthcare and medicine, energy markets and climate change, technology security and operations, biological systems and epidemiology, belief and value systems, supply chains, economies, autonomous systems, and more. Mesa has been cited in more than 350 published papers by more than 790 authors, and the number of projects using the framework runs in the thousands. The core package has had more than 133 contributors. Lastly, I discuss the opportunity to take Mesa from a general tool to an integrated ecosystem to increase research efficiency and standardize the approach to data and building models. Mesa could accelerate research further through community engagement via coding sprints, user research, and a digital community hub.

Author

• Kazil, Jacqueline L.

• ABM
• agent-based modeling
• complex system modeling
• digital twin
• Python framework
• simulation framework

Identifier

hdl:1920/13778

OAI identifier oai:identifier

oai:MARS:1920/13778