دانلود رایگان مقاله مدل به اشتراک گذاری منابع بر اساس خوشه جدید برای شبکه فمتوسل

Abstract

Femtocells have been deployed to enhance indoor coverage, improve the system capacity of cellular networks, and increase the spectrum efficiency by means of full subcarrier reuse among macrocell and femtocells. Nevertheless, the introduction of hybrid access mode imposes new challenges for the resource sharing model in the femtocell networks such as: (1) granting access to public users while guaranteeing QoS of subscriber transmissions, (2) trade-off between level of offloaded traffic from macrocell and bandwidth allocated to femto-tier and (3) appropriate power settings that finds a compromise between the overall system performance and the bandwidth allocated to femtocells. In this paper, we propose a novel cluster formation technique together with a resource sharing model based on Particle swarm optimization technique. Our algorithm aims at maximization of the network throughput and determines the serving base station and the amount of resources per user taking into account user locations, demands, femtocell proximity and traffic load in existing clusters. Simulations are conducted to show the performance of the proposed model contrasted with a benchmark model based on known Weighted Water Filling resource allocation algorithm and known cluster formation technique.

7. Conclusion

In this paper we proposed a novel cluster based resource sharing model for OFDMA femtocell networks. The model consists of the Particle swarm optimization based resource allocation algorithm and the load balanced clustering scheme. The proposed model is able to determine the best serving BS and the bandwidth and power allocation for each user taking into account its demand, location, FC proximity and current cluster configuration. Our solution was tested under the incremental public user number scenario and compared with the benchmark model based on Weighted Water Filling resource allocation algorithm and the interference mitigation and bandwidth reduction based clustering scheme. We demonstrated that the proposed approach indeed improves the overall network throughput without depriving subscribers satisfaction by means of rewarding the femtocells with extra resources for their own transmission. Moreover, in the tested scenarios, the macrocell power consumption is reduced by 3 dB since the femtocells grant access to public users. By means of the femtocell power control, the proposed solution reduces the inter-cluster interference and allows the efficient bandwidth usage. The main disadvantage of the benchmark resource allocation algorithm lies in the lack of femtocell power control which increases the inter-cluster interference level and therefore degrades the QoS of femtocell subscribers transmissions. The proposed model has the drawback of high complexity and therefore long running times. In future works, we will investigate the PSO variants for the resource allocation, study further the cluster formation mechanism and its possibility to convert our model into a distributed model, and analyze the performance of the proposed framework using other channel propagation models that include the shadowing and fading effects.