Melodie: Agent-based Modeling in Python

Agent-based models (ABMs) characterize physical, biological, and social economic systems as dynamic interactions among agents from a bottom-up perspective. The agents can be molecules, animals, or human beings. The interactions can be water molecules forming a vortex, ants searching for food, or people trading stocks in the market.
Agents’ interactions can bring emergent properties to a system and turn it into a complex system. The core reason for using ABMs is usually to model such mechanisms. Besides, taking social economic systems as example, ABMs are also flexible to consider agents’ (1) heterogeneity (e.g., wealth, risk attitude, preference, decision-making rule, etc.) based on micro-data; and (2) bounded rationality and adaptation behavior based on psychological and behavioral studies.
Melodie is a general framework for developing agent-based models (ABMs) in Python. It is published and maintained on the GitHub organization page of ABM4ALL, a developing community among agent-based modelers for sharing ideas and resources. Together with the code repository, we have also published the documentation of Melodie, including a tutorial explaining how a minimum example - an agent-based covid contagion model - can be developed with Melodie step by step.