Everyone is different, but does it matter? The role of heterogeneity in empirically grounded agent-based models of alternative fuel vehicles diffusion

There is a large literature on agent-based models (ABMs) to study the diffusion of alternative fuel vehicles (AFVs). Potentially, ABMs could be used to design policies that effectively promote AFVs. Unfortunately, ABMs have several drawbacks related to their complexity - models that are too simple are unrealistic, and models that are too complicated are difficult to describe, verify, and validate. Here we investigate what level of complexity is needed. We focus on the issue of heterogeneity because it is one of the biggest advantages of ABMs, but also one of the main sources of complexity. We begin with a brief review of ABMs for AFV diffusion. We then generalize an empirically grounded ABM of AFVs to analyze the role of different types of heterogeneity related to individual characteristics and social network structure. We show that most of these heterogeneities do not affect the outcome of the model. To facilitate replication of our results, we describe the model and its calibration to empirical data in detail. We also provide a link to a public GitHub repository where the code files, empirical data, and scripts are uploaded to analyze the results.