Research on Throughput Optimization Techniques for UAV-Assisted Communication Scenarios

In UAV communication networks, there exists a conflict between the limited bandwidth allocation and the non - uniform bandwidth demands of terminal devices. To address this issue, an enhanced Kuhn - Munkres (KM) algorithm is proposed to achieve optimal bandwidth allocation. This problem is formulated as a nonlinear and non-convex optimization problem, which is approximated as an optimal matching problem in a bipartite graph by treating users and UAVs as two distinct sets. The Air-to-Ground (A2G) model, which is separated into line-of-sight (LoS) and non-line-of-sight (NLoS) propagation, is used to describe the communication links between UAVs and users. Under the constraints that the total number of users served by UAVs and the total allocated bandwidth do not exceed their maximum limits, the node weights are initialized, and an unweighted subgraph that meets the matching conditions is determined. Subsequently, the adjacency matrix of this unweighted subgraph is input into the improved KM algorithm. The algorithm assesses whether a single match is completed or if further execution is necessary, and finally, it generates the allocation scheme with the maximum throughput. Simulations in different environments are carried out to compare the performance of various algorithms. The results show that, in terms of efficiency, the KM algorithm outperforms the greedy and genetic algorithms by 26.67% and 171.43% respectively. In terms of throughput, the improvements of the KM algorithm achieves enhancements of 29.84% and 167.80% over greedy and genetic algorithms, respectively.