Intent Recognition in Multi-Agent Systems: Collective Box Pushing and Cow Herding

Abstract

In a multi-agent system, an idle agent may be available to assist other agents in the system. An agent architecture called intent recognition is proposed to accomplish this with minimal communication. In order to assist other agents in the system, an agent performing recognition observes the tasks other agents are performing. Unlike the much studied field of plan recognition, the overall intent of an agent is recognized instead of a specific plan. The observing agent may use capabilities that it has not observed. This study focuses on the key research questions of: (1) What are intent recognition systems? (2) How can these be used in order to have agents autonomously assist each other effectively and efficiently? A conceptual framework is proposed for intent recognition systems. An implementation of the conceptual framework is tested and evaluated. We hypothesize that using intent recognition in a multi-agent system increases utility (where utility is domain specific) and decreases the amount of communication. We test our hypotheses using two experimental series in the domains of Box Pushing, where agents attempt to push boxes to specified locations; and Cow Herding, where agents attempt to herd cow agents into team corrals. A set of metrics, including task time and number of communications, is used to compare the performance of plan recognition and intent recognition. In both sets of experimental series, intent recognition agents communicate fewer times than plan recognition agents. In addition, unlike plan recognition, when agents use the novel approach of intent recognition, they select unobserved actions to perform, which was seen in both experimental series. Intent recognition agents were also able to outperform plan recognition agents by sometimes reducing task completion time in the Box Pushing domain and consistently scoring more points in the Cow Herding domain. This research shows that under certain conditions, an intent recognition system is more efficient than a plan recognition system. The advantage of intent recognition over plan recognition becomes more apparent in complex domains.