Multi-UAV collaborative target allocation Based on Two-Way Knowledge Transfer in complex environments

Multi-UAV collaborative target allocation (MUCTA) refers to the reasonable allocation of limited UAV resources according to the location of each UAV and target point in a complex simulation environment, so that the UAV population can avoid the threat source and effectively attack target points in different locations at the same time. During the planning process, large-scale UAVs and target points greatly increase allocation conflicts and the iterative calculations of the planning system lead to a less efficient planning system. To improve the efficiency of multi-UAV collaborative target allocation, this work proposes a Differential Evolutionary Algorithm Based on Two-Way Knowledge Transfer (TWKT- DE). Firstly, the constraints of UAVs are linearly weighted and transformed into a new objective function. The value of violated constraints is an important evaluation index of UAV allocation conflict. This method can effectively reduce the allocation conflict between UAVs. Secondly, the initial UAV allocation scheme is obtained by the DE algorithm and defined as reference knowledge, and the selection probability between tasks updates the UAV allocation scheme. Finally, the method effectively improves the efficiency of the planning system search by guiding the search direction of the planning system through the two-way transfer mechanism and reducing the redundancy of the allocation scheme. Experiments 1 and 2 validate the proposed algorithm, respectively, and the results verify the outstanding performance of TWKT-DE in dealing with multi-UAV collaborative target allocation with low allocation conflicts, high collaborative capability, and high planning efficiency.