A Successive Coordinated Power Allocation Water- Filling Algorithm for Heterogeneous OFDMA Cellular Systems Athanasios Vasileiadis Dept. of Electronics Alexander Technological Educational Institute (ATEI) of Thessaloniki, Greece vasathanasios@gmail.com Athanassios C. Iossifides Dept. of Electronics Alexander Technological Educational Institute (ATEI) of Thessaloniki, Greece aiosifidis@el.teithe.gr Abstract—This paper proposes a successive coordinated power allocation (SCPA) scheme for orthogonal frequency division multiple access (OFDMA) systems based on the water-filling algorithm. The model assumes two transmitters belonging to heterogeneous cells, i.e. a macro cell and a micro/pico cell with predefined different power capabilities that cover a common geographical area over a channel that includes path loss and shadowing. The main idea is the reduction of the computation complexity and the minimization of information exchange between the two transmitters while preserving similar performance with the optimal cooperative power allocation (OCPA) joint transmission algorithm, recently proposed in the literature. The algorithms are evaluated via extensive Monte- Carlo simulations. The results show that the proposed algorithm provides similar sum rate than OCPA and presents a higher mean number of serving mobile units while reducing, at the same time, the complexity by roughly to orders of magnitude. Index Terms—transmission coordination; power allocation; watre-filling; cooperative transmission; heterogeneous cells; I. INTRODUCTION Coordinated multipoint (CoMP) transmission has received significant attention in academic literature and is being studied as a technique to increase performance, especially at cell edge, within numerous wireless cellular systems including LTE, IEEE 802.16, IEEE 802.22 etc. [1-2]. In general, there are three different types of coordination which can be applied, i.e. (i) coordinated scheduling, where the transmitters are cooperating to silence each other’s subchannels that are already being used to transfer information, (ii) coordinated beamforming, in which each transmitter creates multiple beams in the covered area, that can be used accordingly to reduce interference among the transmitters, and (iii), joint transmission, where two or more transmitters can send information simultaneously to one or more receivers [3]. This paper focuses on the downlink power allocation of cooperating transmitters over a predefined number of subcarriers of an orthogonal frequency division multiple access (OFDMA) system. This topic has gained research interest lately. Specifically, Luo et al. [4]-[5] analyzed a two- transmitter model and proposed a joint transmission algorithm, referred to as optimal joint water-filling (OJWF). They proved that the sum rate is maximized when only one (among the available) subchannel is jointly used by the two transmitters for transmission while the rest of the subchannels are appointed to the transmitter with the better signal-to-noise ratio (SNR) following an independent (per transmitter) water-filling process. In [6] the same authors proved that the same algorithm can achieve almost similar results with constant power allocation. In this case, instead of the final step of power allocation according to joint water-filling process, all the subchannels approved by the water-filling process are assigned with equal power. Furthermore, H. Wang et al. [7] proposed an improved version of the aforementioned joint transmission algorithm, namely, the optimal cooperative power allocation (OCPA) algorithm. OCPA applies the same water-filling process with OJWF by adding a quality criterion about the water level and the common subchannel use. When the criterion is met, the algorithm proceeds normally (by finding the best subchannel for joint transmission), otherwise the algorithm reverts to a standard (non-joint) independent water- filling power allocation where the power of the subchannels is decided according to their SNR, independently by each transmitter. From a practical point of view, all the above mentioned algorithms belong to the third type of coordinated multipoint transmission, i.e. joint transmission, and thus suffer from three drawbacks: (i) they run solely at one of the transmitters which informs the other about the power allocation decision (in a master-slave fashion) (ii) the complexity of the algorithms is significant, growing exponentially with the number of subchannels and (iii) joint transmission requires very good synchronization between the transmitters. In this paper we propose a simple algorithm that belongs to the first type of coordinated multipoint transmission above mentioned, i.e. coordinated scheduling which we will refer to as successive coordinated power allocation algorithm (SCPA). In the proposed algorithm the transmitters apply a water-filling process independently and successively, based on the available SNR subchannel information. The algorithms are evaluated extensively over a heterogeneous simulation system comprising a macro cell and a pico cell [8] that follow different 978-1-4673-5828-6/13/$31.00 ©2013 IEEE