Extending OCL with Subjective Logic.

Abstract

The logic of the UML and OCL modeling languages is based on crisp values, e.g., true or false. However, when modeling systems that work in physical environments or where human actors are involved, different users may have subjective opinions about the reality that they perceive, and thus may need to assign different levels of confidence to the logic predicates of the models. These different views, or opinions, may also be subject to uncertainty when there is a lack of knowledge about the system, adding the dimension of ignorance to the traditional belief-disbelief dichotomy. This paper proposes an extension of the OCL/UML datatype Boolean that enables the representation of subjective uncertain opinions, together with a set of logical operators for reasoning with uncertain propositions in order to reach better informed decisions. The proposal has been implemented as an extension of the UML-based Specification Environment (USE) tool, and validated with several applications and case studies.