Frequency Assignment for Aviation Navigation Stations Based on an Improved Multi-Objective Genetic Local Search Algorithm

Abstract

With the rapid development of both commercial and general aviation, the frequency assignment problem for aviation navigation stations has become increasingly important. This paper presents a general algorithm for frequency assignment at individual aviation navigation stations. Subsequently, a frequency assignment model for multiple civil aviation navigation stations is established to address large-scale frequency allocation challenges. To overcome the limitations of traditional multi-objective genetic algorithms, such as slow convergence speed and susceptibility to local optima, this study proposes several improved algorithms, including the multi-objective genetic algorithm with randomly assigned weights, the multi-objective genetic local search algorithm, and an improved multi-objective genetic local search algorithm, while optimizing key algorithm parameters. The problem involves multiple objectives, including minimizing interference in frequency assignment and reducing the total number of assigned frequencies. Experimental results demonstrate that the proposed improved multi-objective genetic algorithms—especially IMOGLSA-II—effectively address the frequency assignment problem for aviation navigation stations, achieving notable improvements in solution quality, convergence speed, and stability compared with other multi-objective genetic algorithms. In particular, although the time complexity of the proposed algorithm is slightly higher due to the incorporation of local search mechanisms, it exhibits clear advantages in reducing parameter sensitivity, simplifying algorithm structure, and enhancing engineering applicability. These characteristics make the proposed method not only well-suited to the static and constrained nature of aviation frequency assignment, but also more practical and effective than other mainstream multi-objective optimization algorithms in similar engineering scenarios. Furthermore, the proposed method offers a reliable approach that can be extended to other static frequency assignment problems and broader classes of multi-objective optimization tasks.