文献类型：期刊文献

英文题名：A simplified task-assisted constrained multi objective genetic algorithm for USV-UAV collaborative maritime search and rescue

作者：Deng, Pengyu[1]; Sun, Bing[2]; Jiang, Qingchao[3]; Fan, Qinqin[1]

机构：[1] Shanghai Maritime University, Logistics Research Center, Shanghai, 201306, China; [2] Shanghai Maritime University, Shanghai Engineering Research Center of Intelligent Maritime Search & Rescue and Underwater Vehicles, China; [3] East China University of Science and Technology, Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, Shanghai, 200237, China

年份：2025

外文期刊名：IEEE Transactions on Vehicular Technology

收录：EI(收录号：20254519446780)

语种：英文

外文关键词：Accidents - Antennas - Clustering algorithms - Collaborative learning - Conformal mapping - Deep learning - Economics - Genetic algorithms - Unmanned aerial vehicles (UAV) - Waterway transportation

摘要：The maritime transportation is integral to global economic activity. However, the increased frequency of maritime operations heightens the risk of accidents, exacerbated by unpredictable weather conditions and high-risk task environments. Therefore, devising efficient and reliable maritime search and rescue (MSR) systems using unmanned technologies is essential to mitigating these risks and enhancing safety. To implement the above objective, the unmanned surface vehicle (USV) - unmanned aerial vehicle (UAV) collaborative MSR problem with time window constraints is proposed in the present study. To do this, we propose a simplified task-assisted constrained multi-objective genetic algorithm based on a spatio temporal clustering method, called the CMOGE-STC. Firstly, task points are clustered using the self-organizing maps (SOM) method based on their spatiotemporal characteristics, reducing the search space. The clustering centers are then used to determine the positions of USVs. Subsequently, a simplified task assisted multi-objective genetic algorithm is developed to solve multi-UAVs MSR problems with time window constraints. In the CMOGE-STC, the main task aims to address the original MSR problem, while the other two simplified tasks focus on providing useful evolutionary knowledge for the main task. To verify the effectiveness of the CMOGE-STC, 16 artificial MSR scenarios and three famous multi-objective evolutionary algorithms are used in experiments. The experimental results demonstrate that the CMOGE-STC can consistently outperform compared algorithms in terms of hypervolume (HV) of feasible solutions. Additionally, the results of all compared algorithms in a real word scenario indicate that the proposed algorithm is a competitive and effective approach for solving complex MSR tasks. ? 1967-2012 IEEE.