Joint Transmission in Cellular Networks with CoMP - Stability and Scheduling Algorithms

TL;DR

This paper investigates joint transmission in cellular networks with CoMP, proposing algorithms for scheduling and demonstrating that effective CoMP deployment can significantly improve user throughput, especially at cell edges, with manageable backhaul requirements.

Contribution

It formulates the NP-hard joint scheduling problem for CoMP with JT, develops optimal and approximation algorithms, and proves throughput optimality of the proposed scheduling policy.

Findings

CoMP with JT improves cell-edge user throughput.

Low-capacity backhaul suffices for significant gains.

Proposed algorithms distribute resources evenly with minimal intra-cell impact.

Abstract

Due to the current trend towards smaller cells, an increasing number of users of cellular networks reside at the edge between two cells; these users typically receive poor service as a result of the relatively weak signal and strong interference. Coordinated Multi-Point (CoMP) with Joint Transmission (JT) is a cellular networking technique allowing multiple Base Stations (BSs) to jointly transmit to a single user. This improves the users' reception quality and facilitates better service to cell-edge users. We consider a CoMP-enabled network, comprised of multiple BSs interconnected via a backhaul network. We formulate the OFDMA Joint Scheduling (OJS) problem of determining a subframe schedule and deciding if and how to use JT in order to maximize some utility function. We show that the OJS problem is NP-hard. We develop optimal and approximation algorithms for specific and general topologies, respectively. We consider a time dimension and study a queueing model with packet arrivals in which the service rates for each subframe are obtained by solving the OJS problem. We prove that when the problem is formulated with a specific utility function and solved optimally in each subframe, the resulting scheduling policy is throughput-optimal. Via extensive simulations we show that the bulk of the gains from CoMP with JT can be achieved with low capacity backhaul. Moreover, our algorithms distribute the network resources evenly, increasing the inter-cell users' throughput at only a slight cost to the intra-cell users. This is the first step towards a rigorous, network-level understanding of the impact of cross-layer scheduling algorithms on CoMP networks.