Multi-Agent Logic

We present a formalism for reasoning about the information properties of multi-agent systems. Multi-agent logic allows us to represent what is objectively true of some environment, what is visible, or accessible of the environment to individual agents, what these agents actually perceive, and finally, what the agents actually know about the environment. The semantics of the logic are given in terms of a general model of multi-agent systems, closely related to the interpreted systems of epistemic logic. After introducing the logic and establishing its relationship to the formal model of multi-agent systems, we systematically investigate a number of possible interaction axioms, and characterise these axioms in terms of the properties of agents that they correspond to. Finally, we illustrate the use of the logic through a case study, and discuss issues for future work.