Flow-level delay optimization with traffic adaption and inter-cell interference coordination in cellular networks

In this paper we propose an algorithm to optimize the network-level delay based on flow-level model, denoted as flow-level delay, in fractional frequency reuse (FFR) cellular networks. The flow-level model can capture the impact of the dynamic traffic in the network by considering the arrival and departure of the traffic flows in all cells. In addition to this temporal distribution, the spatial distribution also affects the network-level delay when some cells are with hotspots. To achieve minimal network-level delay by adapting to the traffic in the network, we should reallocate the sub-bands among cells while coordinating the inter-cell interference. Based on flow-level model, we formulate it as a sub-band allocation problem at network-level which is a 0-1 optimization and mathematically intractable. Hence we turn it into a graph coloring problem and utilize the concept of maximum weight independent set in graph theory to solve it. Extensive simulations prove that our algorithm can reduce the network flow-level delay significantly.