Research on an Optimized Model for Drone Decoy Jammer Shielding Based on Particle Swarm Optimization Algorithm and Two-Stage Decomposition Algorithm

Abstract

This paper proposes a shielding optimization model for smoke decoy deployment by unmanned aerial vehicles (UAVs) and a particle swarm optimization (PSO) algorithm solution. It focuses on exploring model construction and PSO algorithm application effectiveness under different UAV and smoke decoy configurations. First, for a scenario involving a single UAV releasing one smoke grenade, a model is constructed with the objective of maximizing shielding duration. Decision variables and constraints—including velocity vectors and detonation point locations—are defined. After 25 iterations, the PSO algorithm converges to an optimal or locally optimal solution. Second, for a scenario involving a single UAV releasing three smoke grenades consecutively, the model was expanded to eight decision variables. Considering the spatiotemporal synergistic effects of smoke clouds, the PSO algorithm maximized the combined shielding duration. The objective function rapidly decreased during the initial algorithm phase before stabilizing. Finally, for a scenario involving five drones each releasing up to three smoke grenades, the high-dimensional problem was decomposed into two stages. The PSO algorithm was employed to determine the maximum total effective screening duration. This model provides a scientific strategy for drone smoke interference. Its advantages include PSO's efficient handling of nonlinear non-convex problems, while the staged approach reduces dimensionality and avoids getting stuck in local optima.