Particle Field Diffusion Evolutionary for Multi-Objective QoS Optimization in Swarm UAVs

This paper proposes a novel particle field diffusion evolutionary method for multi-objective QoS optimization in UAV swarm networks (MQSPFO). To address the multi-objective QoS trade-offs of energy, delay, reliability and other conflicting objectives in UAV swarm, and the limitations of existing evolutionary methods in dynamic environment, by integrating Support Vector Regression (SVR) with particle field diffusion, MQSPFO dynamically optimizes serval key metrics of QoS-aware while maintaining Pareto diversity. Experimental results and comparisons shown that the proposed method achieves higher bandwidth, lower latency, and reduced packet loss, with superior scalability, providing a computationally efficient, adaptive solution for future UAV swarm QoS optimization under complex dynamic topologies.