Multi-Traveling Salesman Algorithm for UAV Swarms Based on Improved Genetic Algorithm

The aim of this paper is to solve the path design problem for UAV swarms traversing multiple targets, which can be approximated as a multi-travelling salesman problem. In this paper, an improved genetic algorithm is first introduced. By introducing K-means cluster analysis and combining it with the greedy algorithm, a good initial solution is provided for the genetic algorithm. In this paper, suitable crossover and mutation operators are specially designed for the characteristics of the MTSP problem, and two new parameters are introduced to assist in controlling the selection, crossover and mutation of the algorithm. Simulation experiments prove that the new algorithm has good optimization results within a certain range of data sizes, and further analyze the influence of each parameter on the experimental results, and finally obtain a set of excellent parameter values. The algorithm can calculate the flight routes of UAV swarms in a short period of time, achieving savings in energy and time.