.png){kind=icon}
.png){kind=icon}
Return
Experience playback-based DBO algorithm for cooperative multi-type task allocation in sea-air heterogeneous unmanned systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.oceaneng.2025.121464.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper studies the multi-type task allocation problem for sea-air heterogeneous unmanned systems. First, a novel task allocation model is proposed, which integrates multi-type task constraints and simultaneous arrival constraints. Due to the interdependence of task sequences under time constraints and the requirements of diverse types of tasks performed by unmanned vehicles with varying capabilities, the established model exhibits high complexity, large solution space, and limited feasible solutions. This makes the existing optimization methods less efficient in solving the model. Therefore, an experience playback-based DBO algorithm (EPDBO) was proposed, which designs an independent/dependent iterative evolutionary strategy through individual cognitive differences in order to enhance the local optimal escape ability of the traditional DBO algorithm. Meanwhile, the introspection mechanism of individual learning of global and local optimums is designed to mitigate the uncontrollable effect of stochastic evolution on the direction. In addition, the proposed EPDBO algorithm indicates the type of tasks the unmanned vehicle performs with a specific capability and its sequence of tasks through tailored coding and decoding operations. Finally, numerical simulation and hardware-in-the-loop (HIL) experiments demonstrate the effectiveness and superiority of the proposed method.
Keywords:
Heterogeneous unmanned system
Multi-type task allocation
Experience playback
Dung beetle optimization algorithm