Handling Prioritized Goals and Subgoals in a Logical Account of Goal Change (Extended Abstract)
Shakil M. Khan, Yves Lespérance · 2009
Most existing formal models of goals [2, 3] assume that all goals are equally important and many only deal with achievement goals. Moreover, they do not guarantee that an agent’s goals will properly evolve when an action/event occurs, e.g. when the agent’s beliefs/knowledge changes or a goal is adopted or dropped. Also, most of these frameworks do not model the dependencies between goals and the subgoals and plans adopted to achieve these goals – subgoals adopted to bring about a goal should be dropped when the parent goal becomes impossible, is achieved, or is dropped. Dealing with these issues is important for developing effective models of rational agency and BDI agent programming languages. Here, we outline a formal model of prioritized goals and their dynamics that addresses these issues. In our framework, an agent can have multiple goals at different priority levels, possibly inconsistent with each other. We define intentions as the maximal set of highest priority goals that is consistent given the agent’s knowledge. Our formalization of goal dynamics ensures that the agent strives to maximize her utility. Our model of goals supports the specification of general temporally extended goals, not just achievement goals, and also handles subgoals and their dynamics. Our base framework for modeling goal change is the situation calculus as formalized in [4]. We model knowledge using a possible worlds account adapted to the situation calculus [5]. To support modeling temporally extended goals, we introduce a new sort of paths, which are essentially infinite sequences of situations.