Knowledge-based memetic algorithm for joint task planning of multi-platform earth observation system. (October 2021)

• Record Type: Journal Article
• Authors: Xiang, Shang; Wang, Ling; Xing, Lining; Du, Yonghao; Zhang, Zhongqingyang
• Is Part Of: Computers & industrial engineering. Volume 160(2021)
• Publication Date: 2021-10
• Subjects: 00-01 -- 99-00; Joint task planning -- Multi-platform earth observation system -- Memetic algorithm -- Knowledge model; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2021.107559
• Languages: English
• ISSNs: 0360-8352

Abstract

Abstract: With the increasing complexity of earth observation tasks, the single observation mode can no longer meet the current requirements for timeliness and accuracy. The joint use of multi-platform observation resources can give full play to the advantages of each platform and improve the efficiency of task execution. The research on joint task planning of multi-platform is the key to achieve the collaborative effect of multi-platform earth observation system. In this paper, a unified observation path model is proposed. The objective functions of tracking rate, coverage rate, recognition rate and surveillance rate are proposed respectively for moving target tracking, regional target search, multi-point target recognition and continuous target surveillance. To address the problem, a problem-specific knowledge-based memetic algorithm is proposed. The mechanism of combining swarm intelligence algorithm and local search algorithm in the framework of memetic algorithm is used to enhance the ability of exploration and exploitation. Problem-specific knowledge operators are introduced into global search and local search. It can increase the learning ability of the algorithm in the iterative process, and guide the algorithm to adopt reasonable evolutionary strategy. Finally, speed up convergence while avoiding premature maturity. Simulation experiments verify the superiority of the algorithm. Highlights: A multi-platform joint task planning problem is introduced. A unified multi-platform observation path decision model is proposed. The objective function and constraint can be flexibly assembled. A knowledge-based memetic algorithm is presented to address the issue. The problem-specific knowledge of elite, performance and component are designed.