Configuring dynamic heterogeneous wireless communications networks using a customised genetic algorithm

Abstract

Wireless traffic is surging due to the prevalence of smart devices, rising demand for multimedia content and the advent of the “Internet of Things”. Network operators are deploying Small Cells alongside existing Macro Cells in order to satisfy demand during this era of exponential growth. Such Heterogeneous Networks (HetNets) are highly spectrally efficient because both cell tiers transmit using the same scarce and expensive bandwidth. However, load balancing and cross-tier interference issues constrain cell-edge rates in co-channel operation. Capacity can be increased by intelligently configuring Small Cell powers and biases, and the muting cycles of Macro Cells. This paper presents a customised Genetic Algorithm (GA) for reconfiguring HetNets. The GA converges within minutes so tailored settings can be pushed to cells in real time. The proposed GA lifts cell-edge (2.5th percentile) rates by 32% over a non-adaptive baseline that is used in practice. HetNets are highly dynamic environments. However, customers tend to cluster in hotspots which arise at predictable locations over the course of a typical day. An explicit memory of previously evolved solutions is maintained and used to seed fresh runs. System level simulations show that the 2.5th percentile rates are boosted to 36% over baseline when prior knowledge is utilised.