Research And Method Design On Task Allocation And Reallocation Of Multi-UAV Cooperative Strike Tasks

With the development of military science and technology in various countries and the continuous improvement of the technical level in the aviation field,UAV combat will occupy an increasingly important position in future wars.Traditional single UAV cannot be compared due to the limitations of capabilities and resources.How to use efficient control strategies to achieve multi-aircraft cooperative combat to maximize the overall combat effectiveness has become one of the hot topics of modern research.As the core of multi-aircraft tactical deployment and coordinated control,the quality of the task allocation plan largely determines the combat effectiveness and effectiveness of multi-UAV missions.Due to the limitations of the drone itself and the environment,the problem of multi-UAV collaborative task assignment is essentially a multi-objective optimization problem with constraints.This paper takes the multi-aircraft cooperative attack target as the research background,takes the establishment of the task allocation model and the design of the solution method as the main line,and focuses on the two key issues of task allocation-multi-UAV cooperative static task allocation and multiple unmanned aerial vehicle cooperative dynamic task redistribution problem.Mainly carried out the following work:(1)、Establish a pre-allocation model for multi-UAV collaborative tasks.It mainly considers the constraints of multi-aircraft coordination,ammunition constraints,and combat radius.The target probability function,flight range function,ammunition cost function,and value target return are used as evaluation benchmarks,and the importance of each index is expressed through target weighting.The establishment of the task preallocation model lays the foundation for the following solution.(2)、 Propose a hybrid-discrete particle swarm algorithm.Aiming at the characteristics of the task allocation problem,the traditional particle swarm algorithm is discretized,and a coding criterion is proposed to express the task execution plan and execution order.Aiming at the shortcomings of traditional particle swarm algorithm that it is easy to fall into the local optimal value,the hybrid mechanism is introduced to expand the search range of examples to help particles jump out of the local optimal value;the inertia weight and learning factor of the algorithm are nonlinearly adjusted to accelerate the convergence of the algorithm.An example is also carried out to verify the effectiveness of the improved algorithm.(3)、Two situations that trigger task redistribution are described and analyzed,and related allocation models are established.In order to solve the problem of the inability to auction multiple tasks at the same time when the traditional contract network algorithm solves the distributed task allocation and the uneven distribution caused by the unbalanced ability of drones,a contract network algorithm with buffer pool and load balancing strategy is proposed.And the newly proposed algorithm was verified by experiments,which verified the effectiveness and rationality of the algorithm.(4)、The problem of communication loss in the process of task redistribution is described and analyzed,and the redistribution problem is solved when the distance between drones exceeds the maximum allowed communication distance.In view of the two situations,the algorithm of packet parallel contract network and the algorithm of contract network based on communication reorganization are proposed.Through simulation verification,these two algorithms can effectively deal with the task redistribution problem in different situations.