A new hybrid algorithm based on golden eagle optimizer and grey wolf optimizer for 3D path planning of multiple UAVs in power inspection

Ji-Xiang Lv,Li-Jun Yan,Shu-Chuan Chu,Zhi-Ming Cai,Jeng-Shyang Pan,Xian-Kang He,Jian-Kai Xue
DOI: https://doi.org/10.1007/s00521-022-07080-0
2022-03-12
Neural Computing and Applications
Abstract:As an emerging power inspection method, unmanned aerial vehicle (UAV) inspection has the advantages of high safety, high efficiency, and low cost. In the process of power inspection, UAVs need to inspect multiple task points in a complex environment and plan an efficient and feasible path. In this research, the multiple UAVs inspection in the two cases of initial task points and newly added task points is considered. Aiming at these two cases, a hybrid algorithm is proposed in this paper. Firstly, the personal example learning strategy is applied to the golden eagle optimizer (GEO) to get a personal example learning GEO called PELGEO to improve the search ability of the GEO and reduce the possibility of GEO falling into a local optimum. Secondly, the grey wolf optimizer (GWO) is simplified and the differential mutation strategy is introduced to create the simplified GWO with differential mutation called DMSGWO. Finally, to give full play to the advantages of the PELGEO and the DMSGWO, an adaptive hybridization strategy is used to hybridize PELGEO and DMSGWO. The new hybrid algorithm based on GEO and GWO named HGEOGWO is proposed. The HGEOGWO and other algorithms are tested under the CEC2013 test suite. The experimental results show that the HGEOGWO has better optimization performance and stability than some popular algorithms. For the 3D path planning problem of multiple UAVs in power inspection, the proposed algorithm also has obvious advantages compared with some popular algorithms. The code of HGEOGWO can be publicly available at https://www.mathworks.com/matlabcentral/fileexchange/97807-a-new-hybrid-algorithm-based-on-geo-and-gwo.
computer science, artificial intelligence
What problem does this paper attempt to address?