Automated planning for numerous co-existing agents, with uncertainty caused by various levels of their predictability, observability and autonomy, is a complex task. One of the most significant issues is related to explosion of the state space. This paper presents a formal framework which can be used to model such systems and proposes the use of formally-modeled agents' preferences as a way of reducing the number of states. A detailed description of preference modeling is provided, and the approach is evaluated by examples.
展开▼
