A Microeconomic approach to multi-robot team formation

The aggregation of robots into teams is necessitated due to the limited power and communication capabilities in emergency environments. The formation of robot teams significantly enhances the performance and efficiency of search and rescue missions in such environments. As opposed to the classical partitioning application domains, the robot aggregation requires multiple conflicting objectives to be optimized. We propose a novel microeconomic methodology for simultaneous multi-objective partitioning of robots. The method utilizes the strengths of K-Means algorithm, game theoretic modeling, and Nash equilibrium methodology for fast and socially fair partitioning. In this work, partitions are created on the basis of compaction and equipartitioning objectives to identify decentralized robot teams with each robot in a team closest to its communication gateway, as well as each team equally represented in terms of strength. Rigorous simulations were performed to evaluate the performance of the method, and the results indicate that the proposed method performs significantly better than the K-Means methodology, and identifies good solution points.