Two Decades of Multiagent Teamwork Research: Past, Present, and Future

Abstract

Abstract. This paper discusses some of the recent cooperative multiagent systems work in the TEAMCORE lab at the University of Southern California. Based in part on an invited talk at the CARE 2010 workshop, we highlight how and why execution-time reasoning has been supplementing, or replacing, planning-time reasoning in such systems.

Multiagent Coordination and Teamwork: A Case Study for Large-Scale Dynamic Ready-Mixed Concrete Delivery Problem

“…Research in teamwork has two dimensions: the teams related to cooperative MAS [52] and the coalition from game theory. Our work is related to teamwork and MAS.…”

An interaction protocol for bidirectional deliberation on direct help in agent teamwork.

“…The goal of optimization is to change the variable values so th a t the sum over a set of binary constraints and associated payoff or reward functions is maximized. DCOPs have been applied to meeting scheduling problems, allocating tasks, and coordinating teams of agents [Taylor et al, 2011].…”

“…They have been applied to real world domains including robot navigation and ma chine maintenance [Taylor et al, 2011]. Decentralized POMDPs (DEC-POMDPs) are frameworks to model this behavior in multi-agents systems.…”

“…The first one tries to find approximate solutions using efficient algorithms [Nair et al, 2003, Bernstein et al, 2005, which gives no guarantee on the solution quality. The second category tries to find the global optimal solution by applying useful DEC-POM DPs' subclasses [Becker et al, 2004, Nair et al, 2005, which lacks expressiveness [Taylor et al, 2011].…”

“…A recent work by Taylor et al [2011] argues th at execution-time reasoning is a critical component to multi-agent systems and shows th a t robustly combining execution-time and planning time reasoning in DCOP and POM DP frameworks results in better performance in terms of achieved award, runtime and scalability. It presents new algorithms that integrate execution-time and planning-time reasoning and suggests th a t multi-agent community should focus on incorporating more execution-time reasoning.…”