J Shanghai Jiaotong Univ Sci ›› 2021, Vol. 26 ›› Issue (4): 431-445.doi: 10.1007/s12204-021-2269-5

• •    下一篇

UAV Task Allocation for Hierarchical Multiobjective Optimization in Complex Conditions Using Modified NSGA-III with Segmented Encoding

 JIN Yudong (靳宇栋), FENG Jiabo (冯家波), ZHANG Weijun *(张伟军)   

  1. (School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China)
  • 出版日期:2021-08-28 发布日期:2021-06-06
  • 通讯作者: ZHANG Weijun *(张伟军) E-mail:zhangweijun@sjtu.edu.cn

UAV Task Allocation for Hierarchical Multiobjective Optimization in Complex Conditions Using Modified NSGA-III with Segmented Encoding

 JIN Yudong (靳宇栋), FENG Jiabo (冯家波), ZHANG Weijun *(张伟军)   

  1. (School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China)
  • Online:2021-08-28 Published:2021-06-06
  • Contact: ZHANG Weijun *(张伟军) E-mail:zhangweijun@sjtu.edu.cn

摘要: With the recent boom in unmanned aerial vehicle (UAV) technology, many UAV applications involving complex and risky tasks in military and civilian fields have emerged, such as military strikes and disaster monitoring. Task allocation for UAVs is the process of planning the division of work among UAVs, controlled from ground stations by human operators. This study formulates the UAV task-allocation problem as an extended traveling salesman problem and presents a novel UAV task-allocation model for complex air concentration monitoring tasks. Then, an optimized non-dominated sorting genetic algorithm III (NSGA-III) based on a twin-exclusion mechanism, hierarchical objective-domination operator, and segmented gene encoding (i.e., NSGA-III-TEHOD) is developed to solve complex task-allocation problems involving multiple UAVs, hierarchical objectives, obstacles, and ambient wind. The algorithm is tested in several simulations, and the results demonstrate that the new algorithm outperforms NSGA-III, non-dominated sorting genetic algorithm II (NSGA-II), and genetic algorithm (GA) in terms of efficiency of global convergence and early maturation prevention and is available for the hierarchical objective-optimization problems.


关键词: unmanned aerial vehicle (UAV), task allocation, non-dominated sorting genetic algorithm (NSGA), multiobjective optimization

Abstract: With the recent boom in unmanned aerial vehicle (UAV) technology, many UAV applications involving complex and risky tasks in military and civilian fields have emerged, such as military strikes and disaster monitoring. Task allocation for UAVs is the process of planning the division of work among UAVs, controlled from ground stations by human operators. This study formulates the UAV task-allocation problem as an extended traveling salesman problem and presents a novel UAV task-allocation model for complex air concentration monitoring tasks. Then, an optimized non-dominated sorting genetic algorithm III (NSGA-III) based on a twin-exclusion mechanism, hierarchical objective-domination operator, and segmented gene encoding (i.e., NSGA-III-TEHOD) is developed to solve complex task-allocation problems involving multiple UAVs, hierarchical objectives, obstacles, and ambient wind. The algorithm is tested in several simulations, and the results demonstrate that the new algorithm outperforms NSGA-III, non-dominated sorting genetic algorithm II (NSGA-II), and genetic algorithm (GA) in terms of efficiency of global convergence and early maturation prevention and is available for the hierarchical objective-optimization problems.


Key words: unmanned aerial vehicle (UAV), task allocation, non-dominated sorting genetic algorithm (NSGA), multiobjective optimization

中图分类号: