Concurrent mobility load balancing in LTE self-organized networks

Mobility load balancing (MLB) is an important use case in the long term evolution (LTE) self-optimized networks (SON). To combat the potential Ping-Pong load transfer and low convergence issues, we propose a concurrent MLB (CLB) scheme to solve out the asymmetry traffic distribution among multiple cells and the potential hidden-cell problem. CLB is implemented among multiple source and target cells, which can be regarded as a load redistribution game among them. Here, source cells are referred to as the sellers since they have excess load, and target cells as the buyers since they can absorb more load from the sellers. However, the gaming implementation complexity of such CLB is very high and the overhead is too heavy due to onetime trigger of CLB always involving multiple cells. Also, this situation is further challenged by the large-scale deployment of multiple cells in a specific region. Therefore, we further propose an irregular CLB (termed as ICLB) scheme, which overcomes the shortage of CLB of all associated cells who taking part. For ICLB, we denote the cell selection mechanism in detail. Simulation results show that both CLB and ICLB algorithms can overcome Ping-Pong load balancing problem. Meanwhile, the load distribution is more balancing than that of conventional MLB, and the performance of average blocking probability and the number of unsatisfied users are improved significantly.