Interference graph based knapsack clustering algorithm in ultra-dense networks

To solve the problem of high inter-cell interference in ultra-dense networks, this paper proposes an interfer-ence graph based knapsack clustering algorithm. Firstly, the weighted undirected interference graph is drawn by measuring the reference signal receiving power (RSRP) of adjacent base stations (BSs) through user equipment (UE). On this basis, the clustering result matches the distribution of user service through dynamic programming, in which avoiding intra-cluster interference and minimizing inter-cluster interference serve as a criterion. The simulation results show that compared with traditional clustering algorithm, the proposed algorithm effectively improves the satisfaction rate of user service as well as system throughput.