In this paper we attempt to provide a historical overview of the development of agent-based models from a network perspective. We trace two distinct stages in the development of agent-based models – first-generation models (prior to the 1990s) and second-generation models (since the late 1990s), with a salient break in the middle. We highlight the nexus and synergies between these two approaches. We show that the early developments of agent-based models in the cellular automata tradition contained crucial ingredients concerning networks in them even before they were applied to the social sciences. We argue that networks or interconnectivity (coupling) is a fundamental feature that has enabled agent-based modelers to develop a unique version of complexity science, both in terms of ontology and epistemology.
Economic Foundations for Social Complexity Science, Springer, Singapore, pp.89-106