Forecasting the diffusion of electric vehicles: An agent-based model including household choice and social effects of coherence and communication

Abstract

To address the problem of forecasting the diffusion of innovations, we propose an agent-based model to forecast the electric vehicle (EV) market, that considers fuel type choices, simultaneously tackling household decision-making and cognition (in the form of coherence evaluation), as well as processes of communication between households in their social network, which are the drivers of diffusion in the social network. We use a latent class discrete choice model to characterise fuel type choice at the household level, including preference heterogeneity, along with attitudinal and emotional effects with the hot coherence (HOTCO) model of cognitive consistency. Our agent-based model also incorporates direct and indirect communication between households in the network. Results highlight the need for diverse measures targeting vehicle purchase, ownership, and usage dimensions to accelerate EV market adoption.