PCI Planning Based on Binary Quadratic Programming in LTE/LTE-A Networks

In recent years, interference has played an increasingly significant part in bulkier and denser Long Term Evolution (LTE/LTE-Advanced) networks. Though intra-cell interference is successfully improved by Orthogonal Frequency Division Multiple Access (OFDMA), inter-cell interference (ICI) could cause a degradation of throughput and significantly impact Signal-to-Noise-Ratio (SINR) in the downlink (DL) network. Physical Cell ID (PCI) planning, an effective approach to eliminate ICI, is required to reduce collision, confusion and mod $q$ interference, where $q=3$ for Single-Input Single-Output (SISO) system, and $q=6$ for Multiple-Input Multiple-Output (MIMO) system. In this study, a new definition of neighborhood relations was proposed based on the measurement report (MR) data in the actual network. Binary quadratic programming (BQP) model was built for PCI planning through a series of model deductions and mathematical proofs. Since BQP is known as NP-hard, a heuristic Greedy algorithm was proposed and its low complexity both in time and space can ensure large-scale computing. Finally, based on the raw data extracted from the actual SISO system network and the simulation calculation of MATLAB, the experimental results demonstrated that Greedy algorithm not only eliminates conflict and confusion completely, but also reduces the mod 3 interference of 26.213% more than the baseline scheme and far more than the improvement ratio of 4.436% given by the classical graph coloring algorithm.