ABa-KiTo: Agent-Based Kinetics via Topology

[EF3-6] Labour in an Agent-Based Understanding of Green Growth. This package is situated within the context of the project MATH+ EF6-3.

Agent-Based Kinetics via Topology (ABa-KiTo) is a computational framework that extracts and analyses topological insights to better understand emergent phenomena of agent-based model simulations of complex socio-economic systems. This computational framework extends the MoKiTo framework (Molecular Kinetics via Topology).

The framework involves three stages:

1. Exploration of system state space.
   
2. Construction of the χ-function with the ISOKANN algorithm. We use the Julia package ISOKANN.jl .

3. Clustering of simulation data filtered by the associated χ-function to obtain clusters of states that are dynamically close, i.e., not only in terms of χ-values but also spatially close. A second part of this step is an edge assignment to construct a graph representation which captures the system’s topological structure. The χ-function serves as ordering parameter that highlights the dominant kinetic pathways between macrostates in this graph.

In particular, this package contains the complete data set (ABM, simulations, modules, input and output data) of the ABM analysed in the article: Green Growth Meets Koopman: A Data-Driven Understanding of Economic Green Transitions in an Agent-Based Model .