Human decision making for empirical agent-based models: construction and validation

Results from agent-based or multi-agent simulation (AB/MAS) modelling can provide relevant information for policy makers, scientists and stakeholders about the boundary conditions of rural development and the uncertainties involved in land-use/cover change (LUCC). However, the process of model validation that can build trust in the outcomes for new parameter conditions and in future scenarios is not a trivial problem. Apparently, no common measure of the degree of confounding between parameterization and validation data sets exists. The current lack of success and the effort necessary for validating the models can be traced to the weak theoretical representation of human decision making in current models. Thus, this paper reviews various ways to represent land-use decision making using AB/MAS models. It briefly describes process-based decision making as an alternative approach to address the problem of weak theoretical representation of human decision making, and presents a case study of an agent decision-making model applying an empirical validation technique.