A Proposed Network Balance Index for Heterogeneous Networks

Abstract

Load balancing and fairness are used in the heterogeneous network literature to describe how the users or user rates are distributed across the network. While quantitative metrics of fairness exist, there is no formal metric for quantifying load balance. In this letter, we demonstrate that a “fair” network may not be a balanced one, since fairness is from a user perspective, while load balance is from a network perspective. We propose a new network balance index (NBI) metric to measure the load balance across the network, which accounts for transmit powers, bias values for cell range expansion, and multiplicatively weighted Voronoi coverage areas of heterogeneous base stations. We show an application of the proposed metric by implementing a user association refinement algorithm, which aims to improve the NBI metric. Our mathematical analysis and simulations establish that when users are heavily clustered, increasing the NBI metric using the proposed algorithm also increases the sum rate despite decreasing fairness. This highlights the usefulness of the proposed NBI metric in network planning. Show more