Multi-objective algorithm speeds up UAV search-and-rescue

In a disaster, every second is critical. Large-scale disasters, such as earthquakes, hurricanes, and landslides, often involve trapped or stranded survivors in need of rescue, and locating them as quickly as possible can be the difference between life and death. Increasingly, rescuers are turning to unmanned aerial vehicles (UAVs) to quickly search large areas.

Sun et al. developed an algorithm for UAV task allocation in post-disaster scenarios. Their goal was to build upon existing algorithms by including confidence-driven multi-objective particle swarm optimization to minimize failure rates and reduce energy consumption.

The researchers incorporated several advances into their model, including a more effective particle decoding method for task allocations and a refined algorithm to identify optimal solutions. Key to their approach was including multiple constraints — task completion time, UAV payload capacity, and flight range — as part of the evaluation criteria.

“Traditional methods often optimize single objectives or fail to account comprehensively for timing, payload, and flight-range constraints, risking infeasible or suboptimal missions,” said author Na Geng. “By integrating a confidence-driven mechanism with carefully designed constraint handling, the work aims to produce higher-quality, executable allocations that adapt to local neighborhood states and evolving mission fitness.”

Their algorithm outperformed conventional approaches in both simulations and real-world tests.

The authors plan to continue their work by incorporating larger UAV fleets and additional constraints, such as sensor coverage and risk zones.

“Future work will address dynamic task arrivals and communication disruptions typical of disaster environments, aiming to improve robustness and adaptability under uncertain conditions,” said Geng. “Enhancements to handle GPS uncertainties in urban or obstructed environments will be prioritized to further bridge simulation and real-world deployment.”

Source: “Confidence-driven multi-objective PSO for post-disaster UAV task allocation considering rescue deadline and energy consumption,” by Qing Sun, Qihai Chen, Xiaohai Ren, Shuaiqi Pang, Chengkun Zhu, and Na Geng, Journal of Renewable and Sustainable Energy (2026). The article can be accessed at https://doi.org/10.1063/5.0306876 .

This paper is part of the V2G/X Integration in Smart Cities: Advancing Sustainable Urban Energy Systems Collection, learn more here .