Between collaboration and competition: An initial Formalization using Distributed POMDPs

This paper presents an initial formalization of teamwork in multi-agent domains. Although analyses of teamwork already exist in the literature of multi-agent systems, almost no work has dealt with the problem of teams that comprise self-interested agents. The main contribution of this work is that it concentrates specifcally on such teams of self interested agents. Teams of this kind are common in multi-agent systems as they model the implicit competition between team members that often arises within a team. Our work models the internal struggle of agents that are acting in a team as they try to maximise their individual payo while at the same time acting in a manner that is benefcial to the entire team. This dilemma of self interest versus team interest is a problem that has been studied in game and decision making theory, although no clear-cut solution that applies to agent systems has been proposed. Our formalisation is based on the theory of Partially Observable Markov Decision Processes (POMDPs). In this work, we reintroduce and extend the Electric Elves (EElves), an application of personal assistant agents that displays all the characteristics of competition within a cooperative setting. Using E-Elves we show how competition arises out of a collaborative scenario and analyse the shortcomings of previous approaches in handling this competition. Finally, we provide some initial thoughts on how to cope with these problems based on our previous experience with E-Elves.