Abstract—In Ultra-Dense Networks (UDNs), the load across the small cells is not equally distributed due to the random deployment of small cells, the mobility of user equipments (UEs) and the preference of small cells during the selection and reselection. This results in performance degradation concerning the throughput and successful handovers. To address this problem, this paper proposes proactive algorithms for balancing the load across the small-cell clusters and compares their balancing results to the previous reactive algorithms. The proactive algorithms distribute the new UEs, one by one, to the small cells, while the reactive algorithms are only triggered when the load of the chosen cluster reaches a predefined threshold. In addition, this paper employs the design structure matrix (DSM) method in order to balance the load across the small cells and to reduce the inter-communications between the access points (APs) as well. The numerical analysis indicates that the load distribution and the balance efficiency using the proactive algorithm with user rejection are better than those in the reactive algorithms by 34.97% and 9.09%, respectively. Moreover, the proactive algorithm without user rejection with the DSM method achieves the best balance efficiency and reduces the inter-communications between the APs in some cases by 60.60%.

Index Terms—UDN, load balancing, proactive algorithm with or without user rejection, reactive algorithms, DSM method.

Mohamad Salhani is with Aalto University, Espoo, Finland.

Cite:Mohamad Salhani, "Employing the Proactive Algorithms and the Design Structure Matrix Method for Load Balancing in UND Networks," International Journal of Computer and Communication Engineering vol. 8, no. 4, pp. 138-154, 2019.