Heterogeneous networks with dense deployment of femto cells has provided the promising solution to enhance the system throughputs for the next generation wireless communication. When the large number of heterogeneous networks are overlapped, then traditional intercell interface technique failed to mitigate the interference in between the cells. So, to mitigate the interference, it requires advanced approach for improving the cell edge throughputs and spectral efficiency. For this, the paper presents a frame work to allocate the efficient resource among the users in dense networks. We proposed affinity propagation unsupervised learning to form the cluster with center and then regularized the cluster for effectively allocated the resource. Users on the cluster edge has suffering the inter cluster interface, a victim aware and coordination multipoint mechanism is further proposed to allocated the required resources for these victimized users. We analyzed the performance of our proposed framework with proportional fair based criteria. The total throughputs, edge throughput and spectral efficiency of the system are significantly enhanced in our simulation results through this proposed framework.

毫微微蜂窝密集部署的异构网络为提高下一代无线通信的系统吞吐量提供了有希望的解决方案。当大量异构网络重叠时，传统的小区间接口技术无法缓解小区间干扰。因此，为了减轻干扰，需要先进的方法来提高小区边缘吞吐量和频谱效率。为此，本文提出了在密集网络中的用户之间分配有效资源的框架。我们提出了亲和力传播无监督学习来形成具有中心的集群，然后对集群进行规范化以有效地分配资源。群集边缘的用户遇到了群集间接口，进一步提出了受害者感知和协调多点机制，为这些受害用户分配所需的资源。我们使用基于公平比例的标准分析了我们提出的框架的绩效。通过此提议的框架，我们的仿真结果显着提高了系统的总吞吐量，边缘吞吐量和频谱效率。