Describing human decisions in agent-based models - ODD + D, an extension of the ODD protocol

TLDR: The ODD + D protocol may prove helpful for describing ABMs in general when human decisions are included and incorporates a section on 'Theoretical and Empirical Background' to encourage model designs and model assumptions that are more closely related to theory.

Abstract: Representing human decisions is of fundamental importance in agent-based models. However, the rationale for choosing a particular human decision model is often not sufficiently empirically or theoretically substantiated in the model documentation. Furthermore, it is difficult to compare models because the model descriptions are often incomplete, not transparent and difficult to understand. Therefore, we expand and refine the 'ODD' (Overview, Design Concepts and Details) protocol to establish a standard for describing ABMs that includes human decision-making (ODD + D). Because the ODD protocol originates mainly from an ecological perspective, some adaptations are necessary to better capture human decision-making. We extended and rearranged the design concepts and related guiding questions to differentiate and describe decision-making, adaptation and learning of the agents in a comprehensive and clearly structured way. The ODD + D protocol also incorporates a section on 'Theoretical and Empirical Background' to encourage model designs and model assumptions that are more closely related to theory. The application of the ODD + D protocol is illustrated with a description of a social-ecological ABM on water use. Although the ODD + D protocol was developed on the basis of example implementations within the socio-ecological scientific community, we believe that the ODD + D protocol may prove helpful for describing ABMs in general when human decisions are included.