A resource-constrained distributed task allocation method based on a two-stage coalition formation methodology for multi-UAVs

The task allocation problem is an important research field in unmanned aerial vehicles (UAVs). However, most existing task allocation algorithms can form coalitions to address the resources constraints, but cannot support starting tasks at the same time, nor can cope with the new emerging tasks flexibly. To this end, we propose a novel resource-constrained task allocation method based on the performance impact algorithm (RCPIA) to support simultaneously starting tasks and provide more flexibility to reallocate the new tasks. More specifically, based on the proposed task allocation model, we firstly modify the task inclusion phase and conflict resolution phase of the baseline PI algorithm to preferentially allocate the tasks to the UAVs that can complete tasks individually. After that, to make full use of resources and further allocate remaining unassigned tasks, a two-stage coalition formation method is creatively proposed to form a coalition for the tasks that cannot be performed by a single UAV to provide enough resources. Especially, an idle time slot mechanism (ITSM) is investigated to shift the start times of tasks that can be performed by a single UAV to create a longer feasible time slot to insert the task. Thirdly, the reassignment application of the two-stage coalition method is introduced to cope with new emerging tasks. Finally, numerical simulations are constructed to illustrate the procedure of RCPIA and verify the superiority of RCPIA compared with other task allocation algorithms in efficiency and success allocation rate.