Multi-heterogeneous Unmanned Surface Vehicles Task Assignment Method Based on Improved Genetic Algorithm

To address the task assignment problem of multi-heterogeneous unmanned surface vehicles (USVs) cooperatively visiting multiple targets, we establish a bi-objective mathematical model. The dual optimization objectives are to minimize the total travel distance of the USV fleet and to maximize the total reward from target visits. Furthermore, an improved non-dominated sorting genetic algorithm II (INSGA-II) is proposed. First, a heuristic algorithm tailored to the multiple objectives is designed to construct the initial solutions. Second, a fast non-dominated sorting mechanism is utilized to classify and select the solution set. Finally, an adaptive acceptance probability mechanism is integrated with a local search strategy to strike a balance between global exploration and local exploitation. Comparative analyses indicate that INSGA-II significantly enhances solution quality. Compared with the improved CMGA (coevolutionary multi-population genetic algorithm) and the classic NSGA-II algorithms, INSGA-II reduces the total travel distance by 15.6% and 16.9%, respectively, while achieving an average increase of 7.9% in the total reward of target visits, thereby realizing a superior Pareto balance among the multiple objectives. In conclusion, the proposed algorithm not only demonstrates strong robustness across task scenarios of varying scales, but also provides a highly competitive theoretical foundation and algorithmic reference for the intelligent cooperative control of large-scale USV swarms in the future.