Evaluating the Efficiency of Robust Team Formation Algorithms Chad Crawford, Zenefa Rahaman, and Sandip Sen University of Tulsa {chad-crawford,zenefa-rahaman,sandip-sen}@utulsa.edu Abstract. Selecting a minimal-cost team from a set of agents, with as- sociated costs, to complete a given set of tasks is a common and impor- tant multiagent systems problem. Often, some degree of fault-tolerance in such teams is also required which enables the team to continue to com- plete all tasks even if a subset of the agents are incapacitated. A k-robust team is one that is capable of completing all assigned tasks when any k team members are not available. The corresponding decision problem of selecting a k-robust team that costs no more than a desired cost thresh- old has been shown to be NP-Complete. We present and experimentally evaluate, for varying problem sizes and characteristics, heuristic and evo- lutionary approximation approaches to find optimal-cost k-robust teams which can be used for large problems. We present a Linear Program- ming approximation algorithm that is found to produce optimal results for small problem sizes, and prove that it is a 2 ln(n + k)+ O(ln n)- factor approximation. We also present three heuristic algorithms and an evolutionary computation approach which scales up to larger problems. Another advantage of the evolutionary scheme is that can approximate the Pareto-frontier of teams trading off robustness and cost. Keywords: Robust teams, Approximation algorithms, Set multicover algorithms 1 Introduction Team formation is a critical activity in multiagent systems, particularly in coop- erative settings. In an often-studied scenario, teams are formed in a distributed setting where demands of the situation necessitates otherwise self-interested agents to join together in cooperative teams or coalitions [19]. A significant amount of work in the cooperative game theory and in the multiagent systems community on coalition formation addresses this scenario. These approaches seek fair payoff distribution schemes between team members to incentivize agents joining and staying in stable coalitions. Teams can be also formed in a centralized manner with prior knowledge of the capabilities and expertise requirement in the domain. A centralized decision maker is provided requirements for fulfilling a set of tasks in a domain, where each task has certain resource and/or capability requirements, using a set of