AgentsX.jl — An Extended Julia Framework for Exploring Urban and Social Systems

Published in In Review, 2022

Agent-Based Modelling (ABM) is applied successfully in various use cases, including but not limited to economic modelling, socio-behavioural modelling, ecological modelling, public health, and urban design. We draw attention to the emerging ABM platform Agents.jl, written in Julia — an accessible, high-level programming language. We identify Agents.jl as a promising package for interfacing, customising and extending for specialised uses. We present the key design ideas for a proposed extension and interface to the Agents.jl framework for ABM for urban and social systems simulation — AgentsX.jl — that provides greater flexibility of agent definitions for urban and social researchers. Our primary motivation is to formalise ABM design through a “Code as the Model” approach, reducing barriers to documentation and increasing reproducibility. Our proposed design entails a structured means of defining an ABM based on layers modelled after spheres of influence, clearer constructs to coding an ABM as interfaces to Agents.jl, and an insightful visualisation toolkit that uses dimension reduction techniques.

Manuscript will be available for download after the completion of the review process.

Available after publication

BibTeX

Available after publication