What can agents achieve? A logical theory of joint ability in teamwork

Speaker: Hojjat Ghaderi, University of Toronto

The coordination of cooperating but autonomous agents is a core problem in multiagent systems research. A team of agents is said to have joint ability to achieve a goal if despite any incomplete knowledge or even false beliefs that they may have about the world or each others, they still know enough to get to a goal state, should they choose to do so. Unlike in the single-agent case, the mere existence of a working plan is not sufficient to achieve the given goal since there may be several incompatible working plans and the agents may be unable to choose a share that coordinates with those of the rest of team members. Some formalizations of joint ability (e.g. Coalition Logic and ATEL) ignore this issue of coordination within a coalition. Others, including those based on game theory, deal with coordination, but require a complete specification of what agents believe. Such a complete specification is often not available.

In this talk, I present a new formalization of joint ability based on logical entailment in the situation calculus that avoids both of these pitfalls. Inspired by concepts from game theory, agents in the proposed framework try to coordinate their actions through discarding dominated plans. Unlike mainstream game theory, however, the framework deals with strict uncertainty and handles incomplete specifications where some aspects of the world or agents including belief/disbelief may be left unspecified.

This is joint work with Hector Levesque and Yves Lesperance.