Improved AGNES multi-agent dynamic alliance system in complex environment

With the increasing proportion of feedback input in current environment, an Agnes hierarchical clustering algorithm is proposed in unsupervised learning to improve the traditional multi-source agent alliance theoretical framework. The paper comprehensively considers the complexity and ambiguity factors in the environment and the sensor's different individual performance first, describe the multi-agent model of the alliance and design the specific prompt steps of the dynamic alliance detection system under multi-agent cross cue. In addition, the targeted improvement of the unsupervised learning Agnes clustering algorithm combined with the recognition consistency function of the target detection information, Make the system build a dynamic model directly from the changing environment, in order to achieve the accumulation of long-term benefits of the alliance optimization task, a more favorable decision is made on the current short-term benefits. The simulation shows that the traditional overall mobilization strategy such as the relative swarm intelligence algorithm, the improved algorithm is an intelligent method that is more in line with actual needs. Simulation results show that the relative group of traditional overall strategy to mobilize intelligent algorithm, the improved algorithm is more in line with the actual needs of intelligent methods.