.png){kind=icon}
.png){kind=icon}
Return
An effective knowledge-based evolutionary algorithm for task assignment problem of pollination robots and spraying drones in multi-orchard scenarios
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.127408.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Collaborative task assignment between intelligent agricultural robots and drones has significantly enhanced operational efficiency in agriculture. However, research on robot-drone collaboration in distributed agricultural scenarios remains highly limited. To address this gap, this paper investigates the task assignment problem of pollination robots and spraying drones in multi-orchard scenarios (MORDTA), aiming to minimize the maximum completion time across all orchards. A mathematical model is formulated, and an improved evolutionary algorithm with problem-specific knowledge (IEAPK) is proposed to solve the problem. The IEAPK algorithm incorporates two heuristic strategies in the initialization phase, ensuring efficient assignment of robots and drones to orchards while optimizing task distribution. Four novel reallocation-based evolutionary operators are designed to redistribute tasks, making task assignment more reasonable. An evolutionary stage based on multiorchard equilibrium is then presented to guarantee a balanced distribution of robots and drones across orchards. An adaptive local search mechanism is also developed, allowing each orchard to dynamically select the optimal neighborhood search operator, further optimizing task assignments. Additionally, extra evolutionary operations are applied to the critical orchards, facilitating targeted optimization. Comprehensive experiments show that the proposed IEAPK algorithm outperforms existing state-of-the-art algorithms by a substantial margin in both efficiency and effectiveness.
Keywords:
Multi-orchard scenario
Task assignment problem
Robot-drone collaboration
Improved evolutionary algorithm
Adaptive local search
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W
Organization
No organization information available