Logics for Cooperation , Actions and Preferences

In this thesis, a logic for reasoning about cooperative ability, actions and preferences is developed. It is an extention of a cooperation logic with actions, developed by Sauro et al., which is a modular modal logic consisting of an environment module for reasoning about actions and their effects and an agents module for reasoning about the cooperative ability of agents to perform actions. In this thesis, that logic is combined with a preference logic with unary preference modalities. In the resulting logic, we can reason about the abilities of groups to enforce some state of affairs in an explicit way: It is explicitly represented how exactly a group can achieve some state of affairs and how this achievement relates to the preferences of single agents. It is shown that the developed cooperation logic with actions and preferences is sound and complete with respect to the class of multi-agent systems with preferences, which are setlabelled transition systems with a preference relation over the set of states and an attached model in which it is specified which actions groups of agents can perform. The cooperative ability of agents in this framework is investigated in detail. It is shown what is the relation between the actions single agents can perform and the role the agents play within a group when the group is trying to enforce some state of affairs. Moreover, it is shown that the semantic structures of the logic provide a formal framework for mechanism design. Apart from standard game theoretic results, we obtain results that show the relationship between the distribution of action abilities among agents and the properties of implementable choice rules.