Role-Based Ad Hoc Teamwork (2013)
Katie Genter, Noa Agmon, and Peter Stone
An ad hoc team setting is one in which teammates work together to obtain a common goal, but without any prior agreement regarding how to work together. We introduce a role-based approach for ad hoc teamwork, in which each teammate is inferred to be following a specialized role. In such cases, the role an ad hoc agent should select depends on its own capabilities and on the roles selected by its teammates. In this chapter we formally define methods for evaluating the influence of an ad hoc agent's role selection on the team's utility and show that use of these methods facilitates efficient calculation of the role yielding maximal team utility. We examine empirically how to choose the best suited method for role assignment and show that once an appropriate assignment method is determined for a domain, it can be used successfully in new tasks that the team has not encountered before. Unlike much of the rest of the book, this chapter does not focus on methods for recognizing the roles of the other agents. Rather, it examines the question of how to use successful role recognition towards successful multiagent decision-making.
In Plan, Activity, and Intent Recognition, Gita Sukthankar (Eds.), Philadelphia, PA, USA 2013. Elsevier.

Katie Genter Ph.D. Alumni katie [at] cs utexas edu
Peter Stone Faculty pstone [at] cs utexas edu