Effects of load dependent dynamic biasing and association order for cell range expansion

Cell range expansion is a popular technique to encourage more user association towards small cells for heterogeneous networks to improve overall system performance. However, a constant bias value may be unsuitable for higher loads as it may lead to overloading. We propose a load dependent dynamic bias function which decreases the bias value with increasing load, acting as a natural prevention of overloading. Linking user association and biasing, we also investigate the effects of two different association orders and show that associating closest users first produces higher sum rate and average pico user rate than associating farthest users first, yet associates fewer users. Mathematical derivation of equivalent static biases that associate the same number of users is also presented to aid bias configuration. Our results indicate that dynamic biasing can bring improvements in rate performance, and is most beneficial when a minimum rate constraint is not imposed by the small cells.