IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 30, NO. 7, SEPTEMBER 2012 1 Relay Deployment in Cellular Networks: Planning and Optimization Weisi Guo, Tim O’Farrell Department of Electrical and Electronic Engineering University of Sheffield, United Kingdom Email: {w.guo, t.ofarrell}@sheffield.ac.uk Abstract— This paper presents closed-form capacity expres- sions for interfere-limited relay channels. Existing theoretical analysis has primarily focused on Gaussian relay channels, and the analysis of interference-limited relay deployment has been confined to simulation based approaches. The novel contribution of this paper is to consolidate on these approaches by proposing a theoretical analysis that includes the effects of interference and capacity saturation of realistic transmission schemes. The performance and optimization results are reinforced by matching simulation results. The benefit of this approach is that given a small set of network parameters, the researcher can use the closed-form expressions to determine the capacity of the network, as well as the deployment parameters that maximize capacity without committing to pro- tracted system simulation studies. The deployment parameters considered in this paper include the optimal location and number of relays, and resource sharing between relay and base-stations. The paper shows that the optimal deployment parameters are pre-dominantly a function of the saturation capacity, pathloss exponent and transmit powers. Furthermore, to demonstrate the wider applicability of the theoretical framework, the analysis is extended to a multi- room indoor building. The capacity improvements demonstrated in this paper show that deployment optimization can improve capacity by up to 60% for outdoor and 38% for indoor users. The proposed closed-form expressions on interference-limited relay capacity are useful as a framework to examine how key propagation and network parameters affect relay performance and can yield insight into future research directions. I. I NTRODUCTION Relays have been proposed as a solution to solving the challenge of improving local capacity in Long-Term-Evolution Advanced (LTE-A) and 802.16 j/m standards. Its primary purpose is to either increase the capacity of an existing area or to extend the coverage area of the parent cell-site. The Quality- of-Service (QoS) provided by an operator is not necessarily determined by the average performance, but by that achieved by a certain bottom percentage of customers. This is generally customers operating either on the interference limited cell- edge or indoors. Statistically, over 70% of the mobile traffic is carried to indoor users, therefore there is an urgency in addressing how to enhance capacity for users in both of these scenarios. This paper presents a novel closed-form capacity expression for an interference-limited relay-assisted cellular network, with consideration to the capacity saturation of real- istic transmission schemes, as well as both outdoor and indoor users. The benefit of this approach is that given a small set of network parameters, the researcher can determine the capacity of the network, as well as the deployment parameters that maximize capacity without committing to protracted system level simulation studies. A. Review The topic of relays in cellular networks has been well studied in the past [1] [2]. In terms of analysis, existing theory has largely focused on extending the original work on Gaussian relay channels. Closed-form expressions on optimal relay deployment in Gaussian channels was proven in [3] for location and in [4] for resource allocation. In a realistic cellular system, the effects of inter-cell interference [5] and capacity saturation of realistic modulation schemes have a significant impact on both the capacity of the system and the optimal solutions, as shown in [6]. For relay deployment in a multi-cell interference-limited network, the characterization of capacity and outage performance has been limited to simulation based studies [5] [7]. Interference-limited stochastic geometric theoretical methods have not yet been extended to relay channels [8]. The optimization of relay deployment location [9], resource allocation [1] [5], and cost efficiency [10] is conducted using iterative numerical approaches on simulation results. From a system designer perspective, there is a dichotomy in the theoretical and simulation approaches. The lack of tractable interference-limited relay capacity expressions means that one either has to rely on closed-form Gaussian channel expressions or extensive multi-cell simulation results. This has restricted the insight into how and why key network and channel parameters affect the relay performance and optimal deployment solutions. B. Contribution The novelty of this paper is to consolidate the theoretical and simulation based approaches by proposing an interfere- limited theoretical framework that considers capacity satu- ration. The paper presents closed-form capacity expressions for a relay-assisted base-station and maximizes the capacity with respect to deployment parameters. The benefit of this approach is that, for any set of network parameters, the system designer is able to characterize and optimize the multi-cell network performance without resorting to extensive multi- cell simulations. The proposed analysis is also validated by a multi-cell simulator. Furthermore, to demonstrate the wider