Modeling the evolution of generativity and the emergence of digital ecosystems

Abstract

Recent literature on sociotechnical systems has employed the concept of generativity to explain the remarkable capacity for digital artifacts to support decentralized innovation and the emergence of rich business ecosystems. In this paper, we propose agent-based computational modeling as a tool for studying the evolution of generativity, and offer a set of building blocks for constructing agent-based models in which generativity evolves. We describe a series of models that we have created using these building blocks, and summarize the results of our computational experiments to date. We find in several different settings that key features of generative systems can themselves evolve endogenously, including "core" components and reusable parts. Moreover, we find that boundedly rational designers without coordination or foresight can evolve business ecosystems that satisfy a diverse range of consumer preferences and exhibit robustness to changes in these preferences over time.