Adaptive Pigeon-inspired Optimization and Dynamic Gradient Updating MADDPG Decision Method for UAV Swarm Confrontation

In this paper, the Unmanned Aerial Vehicles (UAV) swarm confrontation scenario in three-dimensional environments is deeply studied, with emphasis on the situation assessment method and target assignment technology. In order to address the challenges of inaccurate battlefield situation assessment and low target assignment efficiency, we have innovatively established an expert system-based battlefield situation model and a distributed target assignment model based on the physical model of UAVs. The models proposed in this paper realize fast and accurate decision-making and reasonable resource allocation in complex combat scenarios, providing important support for intelligent decision-making in dynamic combat environments. Additionally, to enhance confrontation stability, an adaptive Pigeon-inspired Optimization (PIO) algorithm is introduced. We integrate PIO into our neural network algorithm, which incorporates convolutional neural networks, generative confrontation networks and gradient-based policies. Through extensive simulations and testing, the effectiveness of our scenario evaluation and intelligent decision-making methods is validated, significantly improving the success rate of UAV swarm confrontations in complex three-dimensional environments.