In this paper, we consider a Long Term Evolution (LTE) network with multiple Packet Gateways (P-GWs), multiple User Equipments (UEs) and a single Mobility Management Entity (MME). The P-GW connects the UE to the external networks. The sessions are created between the User Equipment (UE) and the P-GW via Serving Gateway (S-GW). The MME selects the P-GW and the S-GW for the sessions and handles the control signaling. Multiple P-GWs handle these packets and transfer them to the external packet data networks. Thus, based on how the sessions are established, different gateways may have to handle different amount of load. If proper load balancing algorithms are not used, the load on some gateways may be higher than others leading to Quality of Service (QoS) degradation for the users. We propose a Linear Programming formulation of the problem and four different heuristics: Load Balancing Threshold (LBT); Static APN Associated Weights (SAAW); Dynamic Weights (DW); and Entropy Based Assignment (EBA). The entropy function provides an accurate measure of the difference in the load on the gateways. It measures the uncertainty or randomness in the outcome of an experiment. The performance studies show that the EBA algorithm provides 18% performance improvement for networks consisting of 1,00,000 users, 12 APNs and 50 P-GWs.
