An efficient and adaptive resource allocation scheme for next generation cellular systems Simone Merlin, Andrea Zanella University of Padova – Department of Information Engineering Via Gradenigo 6/B, 35131 Padova, Italy {andrea.zanella, simone.merlin}@dei.unipd.it Abstract— 1 A dynamic and distributed multiuser radio resource allocation scheme for a FDMA-TDMA based mul- ticellular system is proposed. User interference measure- ments and channel estimation at the mobiles are used in or- der to exploit the multiuser diversity of the system. Basic information bearers (i.e. the subcarrier/time-slot pairs) are efficiently allocated and power loaded to accommodate the users rate-demands and dynamically adapt to the interfer- ence produced by neighboring cells. Our solution leverages on the possibility to feed the RRM algorithm with param- eters coming from different layers, following a cross-layer approach. Keywords: Cellular networks, radio resource manage- ment, dynamic distributed allocation. 1. Introduction 4G systems are expected to overcome 3G systems in terms of data rate, efficiency of the band usage, reconfig- urability and QoS provisioning. A promising approach to realize such a vision consists in designing cross-layer so- lutions, which allow the exchange of information among different protocols layers to maximize the overall system performance. Improvements are expected by feeding the radio resource management (RRM) module with physical layer information, like radio links conditions and interfer- ence levels. Such information can be used to exploit the multiuser diversity of the system, while providing the re- quested QoS. In this paper, we focus on the cellular system proposed within the framework of the P.R.I.M.O. project [9]. The radio interface is based on a OFDM modulation, which allows for a FDMA–TDMA medium access mechanism. The elementary radio resource that can be independently allocated is the subcarrier/time–slot couple, which will be called Physical Basic Unit (PBU). In the perspective of maximizing the use of the radio resources, avoiding static and a priori frequency planning, the project re- quires a complete reuse of the radio resources among ad- jacent cells. Furthermore, at this stage of the work, no direct communication is assumed among cells. On the one hand, these requirements promote the definition of 1 This work was supported by MIUR within the frame- work of the ”PRIMO” project FIRB RBNE018RFY (http://primo.ismb.it/firb/index.jsp). a flexible and reconfigurable system which can dynami- cally adapt to traffic load and interference conditions. On the other hand, effective resource allocation algorithms, able to adapt to traffic and interference dynamics, are re- quired to avoid very high interference among neighboring cells [1]. According to the scenario proposed in [9], we assume the algorithm is performed by the basestation (BS) of each cell, while no explicit signalling among cells is con- sidered. We assume that the BS has a clear vision of the channel status and interference levels perceived by each user, for each allocable resource. Such informa- tion is passed to the allocation algorithm that allocates the PBUs, setting power and bit loads, to the mobiles in order to satisfy the rate-demands of the users. So far, only downlink allocation is considered. The rest of the paper is organized as follows. Section 2 overviews the related works. Section 3 describes the sys- tem model considered in the study. Section 4 presents the allocation algorithm and the related heuristic. Results and comments are reported in Section 5. Finally, Section 6 concludes the paper and describes possible evolution of the work. 2. Related works The definition of the allocation algorithm requires the identification of the objective to be optimized and the constraints to be guaranteed. Many algorithms have been proposed in literature. In [2] the objective is to maximize the minimum rate allocated to each user, subjected to a limit in the total power that can be allocated. A different approach is followed in [3], [6] and [4], where the opti- mization problem aims at minimizing the total transmit- ted power, while guaranteeing a minimum rate for each user. Another approach is proposed in [5], where the ob- jective is the maximization of the allocated capacity un- der the constraint of satisfying the rate-demand of each user. Since the problem turns out to be nonlinear, a sim- plified approach is proposed, where the power is allocated in advance, leading to a linear problem. Starting from this approach, we propose a new allo- cation algorithm that aims at maximizing an efficiency function in the PBU allocation, while guaranteeing the IWS 2005 / WPMC’05 - Aalborg, Denmark Copyright  2005 WPMC 1 18-22 September 2005