Optimizing the dynamic task allocation for multi UAV systems in a search and rescue scenario

Abstract Unmanned aerial vehicles (UAVs) are increasingly popular for their versatility and cost-effectiveness in responding to natural disasters and conducting search-and-rescue (SAR) missions. During the execution of the SAR phase, dynamic events may occur that demand changes in task assignments or the number of UAVs to address situations requiring higher priority. This paper extends the compromised dynamic performance impact (CDPI) algorithm and introduces an enhanced CDPI (ECDPI) that can effectively handle dynamic events during task execution. The ECDPI algorithm enhances the detection of dynamic activity, allowing for the identification and management of multiple dynamic events, such as UAV failures, task variations, and system scalability, during execution. The ECDPI algorithm is evaluated against five dynamic events: a UAV fault, variations in task duration, task deadlines, task deletions, and the addition of a new UAV. Simulation results show that the ECDPI algorithm effectively manages multiple dynamic events during task execution. It enables successful task reassignment and ensures tasks are completed on time, even under varying operational conditions..