Reducing Data Traffic Congestion Using Device-to-Device (D2D) Communication and a Greedy Algorithm in Heterogeneous Netw
The rapid growth of cellular users (CU) has significantly increased the demand for wireless network capacity, causing data traffic congestion, especially at the Base Station (BS). As mobile devices proliferate and user activities shift increasingly toward data-intensive services, ensuring stable and efficient network performance has become a central concern for telecommunication providers. One promising solution to reduce the load on BS infrastructure is Device-to-Device (D2D) communication.
D2D communication enables two mobile devices to communicate directly without routing their data through a central BS. This direct mode of communication can help offload traffic from the BS and reduce latency, making it highly suitable for applications in 5G networks and beyond. However, the implementation of D2D communication also introduces new challenges, particularly interference. Since D2D and CU users share the same Resource Blocks (RBs), overlapping signals can degrade communication quality for both parties.
To mitigate this interference, this study adopts a downlink Heterogeneous Network (HetNet) model. A HetNet combines multiple layers of different types of base stations—primarily macro base stations (MBs) and small cell base stations (SBs)—to improve coverage, capacity, and overall performance. By incorporating small cells with limited coverage areas, HetNets offer localized data handling that helps relieve the load on larger MBs.
Resource allocation in this study is handled using a greedy algorithm. Greedy algorithms are widely used in wireless networks due to their simplicity and speed, as they make locally optimal choices at each step with the hope of reaching a globally optimal solution. To further enhance system performance, two small cell base stations—SB1 and SB2—were integrated into the network architecture, resulting in an enhanced greedy algorithm known as the greedy algorithm with SB1SB2.
The study compares the performance of this enhanced greedy algorithm with the traditional greedy algorithm that does not include SB1 and SB2. Several key performance indicators were used for this comparative analysis: sum data rate, spectral efficiency, fairness between CU and D2D users, and power efficiency. Additionally, the researchers adjusted the radius of the MB to evaluate how cell size affects network performance under both algorithmic approaches.
Interestingly, the simulation results revealed that the greedy algorithm with SB1SB2 produced lower values for several critical metrics. The total data rate achieved using this method was 1.62 × 10⁸ bps, which was lower than that of the traditional greedy algorithm. Similarly, the spectral efficiency was recorded at 9.02 bps/Hz, and total fairness between CU and D2D users was just 0.4095. In this context, fairness refers to how equitably resources are distributed among users, which is vital in maintaining user satisfaction and quality of service.
On the other hand, the greedy algorithm with SB1SB2 demonstrated superior performance in terms of power efficiency, achieving a 10.40% improvement compared to the conventional method. This indicates that although some key performance parameters dropped, the modified algorithm offers energy savings, an increasingly important factor in the design of sustainable and cost-effective networks.
The study also explored the impact of varying the radius of the MB. Increasing the MB radius typically extends coverage but may also increase the risk of interference if not managed properly. By experimenting with cell radius adjustments, the researchers gained insights into the trade-offs between coverage area and interference management, contributing to a more nuanced understanding of network performance optimization.
Despite the advantages of incorporating SB1 and SB2, the study concludes that neither of the two greedy algorithm implementations provides an optimal resource allocation solution. This limitation is mainly due to the greedy algorithm’s inherent nature of focusing on local optimization without considering the broader, long-term impact of each decision.
As a result, future work should explore more advanced optimization techniques. Algorithms such as Genetic Algorithms (GA), Particle Swarm Optimization (PSO), or other metaheuristic approaches could be utilized to explore a larger solution space and achieve a more globally optimized resource allocation strategy. These algorithms are capable of balancing multiple constraints and objectives, potentially improving data rate, fairness, and power consumption link simultaneously.
In conclusion, D2D communication offers a promising strategy to alleviate data traffic congestion in modern cellular networks. However, effective interference mitigation and intelligent resource allocation remain key challenges that must be addressed to fully realize its potential. The addition of small cell base stations like SB1 and SB2 introduces some advantages, particularly in energy efficiency, but also highlights the complexity of achieving balanced performance across all system parameters.
This study provides valuable insights into the practical implementation of D2D communication in HetNets, especially in relation to resource allocation strategies. As wireless networks continue to evolve toward 6G and beyond, such research will play a crucial role in shaping efficient, scalable, and sustainable communication systems.