Modeling societies of agents using modal logics

Abstract

Presents a framework that shows how epistemic logic and deontic logic can be used to model a multi-agent system consisting of humans and non-human agents. Modeling agents with epistemic and deontic logic also captures the interdependencies between human and non-human agents. Epistemic logic addresses belief and knowledge (notably common knowledge), and, in many applications in computer science, it specifically addresses knowledge acquired by communication. With deontic logic, we represent obligations, prohibitions and permissions applying to agents. Axioms and rules of inference represent a deductive system for modal logics, and Kripke structures furnish semantic models in which propositions are evaluated to true or false. We are concerned with agents' actions, their attitudes and what they know. Our framework emphasizes the ability to maintain ongoing relationships among agents. Such a system is called reactive. We contrast a reactive system with a transformational system, which, given a set of inputs, produces outputs upon program termination. In developing this framework, we address the concepts of a multi-agent/multi-human collaboration.