A Gossip-Based Auction Algorithm for Decentralized Task Rescheduling in Heterogeneous Drone Swarms

One of the key technologies used in scheduling for drone swarms, which execute complex tasks and have broad application areas, is task assignment. Scheduling ensures the efficient assignment and execution of tasks in drone swarms by considering time constraints and environmental variables. The dynamic nature of the environments in which drone swarms operate requires the swarm to possess a distributed replanning capability that can dynamically accommodate unassigned or newly added tasks while adhering to time constraints. The distributed approach enables rapid task allocation without a central node, allowing drones to freely join or leave the swarm, thereby enhancing both resilience and flexibility. In this article, a distributed method called Harmony drone task allocation (DTA) is proposed for solving the multidrone task assignment problem with complex time window constraints. The proposed Harmony DTA aims to minimize total system cost and task execution delay while ensuring the conflict-free assignment of all tasks within their valid time intervals in dynamic environments with communication constraints. By utilizing a consensus-based auction mechanism and integrating a gossip-based approach, Harmony DTA efficiently minimizes communication load while ensuring that tasks are assigned to the most appropriate drones, balancing task urgency with resource availability. Simulations demonstrate that the proposed method can effectively assign newly emerging time-limited tasks among drones in dynamic environments. Results from Monte Carlo simulations show that Harmony DTA provides assignments with lower total costs and reduced total message size. In addition, the developed gossip-based consensus algorithm has been demonstrated through simulations to deliver conflict-free assignments in communication-limited environments.