Coordinating Busy Agents Using a Hybrid Clustering-Auction Approach

Most auction approaches assume that bidding agents must be available to take on a new task when they submit or at least commit to a bid. This works well for pre-planning, for example, before a group of robots takes on a mission with multiple tasks, as the robots have not yet been assigned to any tasks. However, once a mission has begun, it is difficult to adapt to new situations that arise using the auction approach because most of the robots may already be tasked. This reduces the pool of robots available to take on new tasks. We demonstrate a novel hybrid approach that uses negotiation methods similar to a combinatorial auction, but extends winner determination with a polynomial time constrained clustering algorithm called CLUS-STAR (CLUstering for Self-Synchronizing Tasked Agent Reallocation). CLUS-STAR is able to reassign agents to accommodate new tasks that come up without dropping existing tasks. We show that CLUS-STAR can fulfill all the needs for new and existing tasks significantly more often than a combinatorial auction approach when many of the agents are already tasked, while also decreasing the cost of the tasks. CLUS-STAR can also be used for team or coalition formation problems.