An Efficiency Framework for Task Allocation Based on Reinforcement Learning
Jun Xu, Weiwei Li, Jun Ping Yao · 2023
In response to the challenges of task allocation for Unmanned Aerial Vehicles (UAVs) in large-scale scenarios, this paper proposes an improved task allocation method based on Particle Swarm Optimization (PSO) algorithm and reinforcement learning, named PSO-DQN. We first design an innovational clustering method based on the K-Means algorithm, which combines the target importance with distance to reduce the computation complexity. The PSO-DQN method we formulated serves to allocate targets from UAV swarms to specific target clusters while simultaneously mapping individual targets within a cluster to UAVs. The experimental results demonstrate that the superior task allocation accuracy of the PSO-DQN method when compared to traditional approaches. The framework we proposed also proves that it can significantly reduce the complexity of computation of task allocation for UAV swarms in large-scale scenarios.