ABSTRACT
Human, Social, Cultural, Behavioral (HSCB) models present unique challenges for verification and validation (V&V). Unlike models of hard science systems, HSCB models cannot be readily or robustly tested through experimentation in the real world, nor fully verified or validated using historical data. In order to assess the V&V of an HSCB model, one must better capture and understand the information the model uses to represent the world. A novel way of extracting this kind of information is to decompose an HSCB model into its epistemology. A model’s epistemology is defined as pieces of knowledge describing the model’s construction, ranging from social ontology to raw data. Each of these pieces of knowledge is referred to as an epistemological element of the model. V&V of the model becomes a product of V&V of each epistemological element that define the model. This work describes the decomposition of the model epistemology as a hierarchy, as well as a process for establishing such decomposition. To address the cost of V&V, this work presents an assessment called V&V Early Checkpoint via Epistemological Critique (E-CHEC) designed to mitigate V&V cost. Once the epistemological hierarchy is defined for a model, V&V of each level of the hierarchy is considered with respect to both the focus and the context of the model (i.e., focal V&V and contextual V&V).
