Hierarchical planning about goals and commitments

Abstract

We consider the problem of relating an agent's internal state (its beliefs and goals) and its social state (its commitments to and from other agents) as a way to develop a comprehensive account of decision making by agents in a multiagent system. We model this problem in terms of hierarchical task networks (HTNs) and show how HTN planning provides a natural representation and reasoning framework for goals and commitments. Our approach combines a domain-independent theory capturing the lifecycles of goals and commitments, generic patterns of reasoning, and domain models. Specifically, our approach shows how each agent may take into account its capabilities, costs, and preferences as it plans its interactions (captured as operations on commitments) with other agents to attempt to achieve its goals.