With the wide application of UAVs in modern operations, efficient cooperative task assignment of heterogeneous UAVs under complex constraints has become crucial for enhancing mission success rates. Recently, a team led by Professor Mou Chen from Nanjing University of Aeronautics and Astronautics published a notable research outcome in Defence Technology: an energy learning hyper-heuristic (EL-HH) algorithm for solving this key problem.
"Existing algorithms often face issues like being trapped in local optima and slow convergence when dealing with complex constraints," Chen explains. "We designed a comprehensive mathematical model covering task types, time windows, and UAV payloads, and proposed a three-layer encoding scheme (task sequence, UAV sequence, waiting time) to accurately describe assignment schemes."
The solution lies in the EL-HH strategy, which adaptively adjusts operator selection probabilities through energy learning, combined with multiple optimization operators and directed graph-based task order and time adjustment strategies. "This framework fully explores the solution space while ensuring compliance with various constraints," shares Chen.
Validated through simple/complex simulation scenes and real indoor experiments, the results show that the EL-HH algorithm outperforms PSO, GWO, and other traditional algorithms in convergence speed and solution quality, enabling heterogeneous UAVs to complete tasks efficiently while avoiding obstacles.
"This study provides robust technical support for the cooperative operation of UAV swarms in complex scenarios," adds Chen. "Future research should focus on optimizing the hyper-heuristic strategy to further improve the algorithm's efficiency and adaptability to more dynamic battlefield environments."
