A Heuristic Task Allocation Method Based on Overlapping Coalition Formation Game for Heterogeneous UAVs
Yibing Li,Zitang Zhang,Zongyu He,Qian Sun
DOI: https://doi.org/10.1109/jiot.2024.3406336
IF: 10.6
2024-08-25
IEEE Internet of Things Journal
Abstract:The demand for heterogeneous unmanned aerial vehicles (UAVs) equipped with various types and complementary resources to perform complex tasks is increasing. Existing researches on game theory-based resource allocation primarily focus on continuous resources, by establishing models and transforming them into convex optimization problems. However, the discrete task resource allocation problem considered in this article does not meet the assumptions of continuous optimization models, therefore, a reasonable allocation method is needed. To enhance the allocation efficiency for multi-UAVs operating under discrete resource constraints, we design a heuristic task allocation method based on game theory. Our method integrates multiple critical factors, including task priority and dynamic task benefits, into the resource matching process, offering a robust solution tailored to the diverse demands. Specifically, we first establish an optimization model encompassing UAVs, requirements and revenue, then we adopt a utility function to depict these relationships. Furthermore, considering the distributed cooperative characteristics of UAVs during task execution, we incorporate the correspondence between UAVs and tasks into the framework of a coalition formation game (CFG). Building on this foundation, we develop a heuristic allocation method based on task validity matching and an ineffective resource exit mechanism. We also introduce tabu lists to avoid ineffective searches, aiming to enhance resource utilization while ensuring task utility. Finally, we analyze the convergence of our proposed algorithm and conduct simulation validations across various scenarios. The results indicate that our algorithm can adapt to dynamic changes in task demands, exhibits good scalability, and is capable of enhancing task utility.
computer science, information systems,telecommunications,engineering, electrical & electronic