Path planning of multiple UAVs low-altitude penetration based on improved Multi-agent Coevolutionary Algorithm

In order to solve the multi-UAV cooperative path planning problem of low-altitude penetration, the paper proposes an improved Multi-agent Coevolutionary Algorithm (IMACEA), which introduces co-evolution mechanism based on Multi-agent Genetic Algorithm (MAGA) to find the optimal solution of multi-objective optimization problem by combing the agents´ perception and response capabilities of environment, information sharing capacity among multi-agent systems and heuristic search ability of heuristic search. At the same time, we use absolute Cartesian coordinates and relative polar coordinates coding method to reduce the search space and speed up the convergence rate. To adapt to multi-path planning problems of UAVs, new multi-agent co-evolution operators are designed. Finally, the proposed algorithm is used for multiple UAVs offline and online route planning, simulation results show the effectiveness of the algorithm.