Abstract: In view of the situation of multiple fires in the same area, the efficiency of multi-target rescue by UAV group is low due to the time change of fire conditions, so a fire rescue route planning method of UAV group against multi-fire targets is proposed. First of all, considering the timeliness of fire rescue mission and limited rescue resources, a rescue path planning model of UAV group based on the change of fire rescue value over time was established, which realized the dynamic decision making more in line with the actual rescue requirements for rescue commanders. Secondly, in order to effectively solve the model, an improved bee colony algorithm based on Metropolis acceptance criterion is proposed, which effectively improves the solving accuracy of the algorithm. Through the simulation experiment, the results show that the overall planning ability of rescue task is improved by 11.65% when the total value of multi-objective rescue is increased. The research results can play an important role in actual fire fighting and rescue activities.

Key words: unmanned aerial vehicle, fire fighting and rescue, task allocation, path planning, artificial bee colony