Hybrid Agent-Based Simulation of Adoption Behavior and Social Interactions: Alternatives, Opportunities, and Pitfalls

Agent-based modeling and simulation (ABMS) is a powerful analysis tool that has led to significant contributions in the field of innovation diffusion. In this article, we examine the potential and pitfalls of extending adoption models used in agent-based diffusion via machine learning (ML) and soft computing (SC) techniques. More specifically, we 1) classify features related to agents' decision-making and social interactions that are generally not considered in current adoption models; 2) present, along with illustrative examples, an assessment of the potential of hybrid ABMS involving ML and SC to incorporate and model these features; and 3) identify essential considerations for the implementation and applicability of such adoption models. To support future efforts in developing computational systems based on these hybrid ABMS, the article also highlights research areas to further investigate at the intersection of ABMS, ML, and SC.