1 Game Theoretic Infrastructure Sharing in Multi-Operator Cellular Networks Alexandra Bousia, Student Member, IEEE, Elli Kartsakli, Member, IEEE, Angelos Antonopoulos, Member, IEEE, Luis Alonso, Senior Member, IEEE, and Christos Verikoukis, Senior Member, IEEE Abstract—The introduction of 4th Generation (4G) wireless technologies has fueled the rapid development of cellular net- works, significantly increasing the energy consumption and the expenditures of the Mobile Network Operators (MNOs). In addition, the network underutilization during low traffic periods (e.g., night zone) has motivated a new business model, namely infrastructure sharing, that allows the MNOs to have their traffic served by other MNOs in the same geographic area, thus being able to switch off part of their network. In this paper, we propose a novel infrastructure sharing algorithm for multi-operator environments, which enables the deactivation of underutilized Base Stations (BSs) during low traffic periods. Motivated by the conflicting interests of the MNOs and the necessity for effective solutions, we introduce a game theoretic framework that enables the MNOs to individually estimate the switching off probabilities that reduce their expected financial cost. Our approach reaches a Dominant Strategy Equilibrium (DSE), which is the strategy that minimizes the cost of each player. Finally, we provide extensive analytical and experimental results to estimate the potential energy and cost savings that can be achieved in multi-operator environments, incentivizing the MNOs to apply the proposed scheme. Index Terms—Infrastructure sharing, Switching off, Base sta- tion sleep, Game theory, Energy efficiency I. I NTRODUCTION The rapid expansion of mobile services, along with the emerging demand for multimedia applications, driven by the widespread use of laptops, tablets and smart devices, has led to an impressive growth of the data traffic volume during the last few years. According to recent market predictions [1], global mobile data traffic is expected to increase nearly 11- fold in the next five years, reaching 15.9 exabytes per month by 2018. Hence, Mobile Network Operators (MNOs 1 ) seek to extend their infrastructure by installing more Base Stations (BSs), in an effort to increase the capacity of their network and meet these pressing traffic demands. The additional infrastructure not only implies a rise in the Capital Expenditures (CapEx), but also has a direct impact A. Bousia, E. Kartsakli and L. Alonso are with the Department of Signal Theory and Communications (TSC) of the Technical University of Catalonia (UPC), Barcelona, Spain, e-mail: {alexandra.bousia, ellik, luisg}@tsc.upc.edu. A. Antonopoulos is with the Department of Signal Theory and Communica- tions (TSC) of the Technical University of Catalonia (UPC), Barcelona, Spain and with Telecommunications Technological Center of Catalonia (CTTC), Barcelona, Spain, e-mail: aantonopoulos@cttc.es. C. Verikoukis is with Telecommunications Technological Center of Cat- alonia (CTTC), Barcelona, Spain, e-mail: cveri@cttc.es. 1 Hereafter, the terms “MNO” and “operator” will be used interchangeably. on the network energy consumption, thus resulting in higher Operational Expenditures (OpEx) [2]. In particular, the use of information and communication technology across a wide range of applications accounts for 5.7% of the world’s elec- tricity consumption and 1.8% of global carbon emissions [3], something that translates into electricity bills in the order of $10 billion for the MNOs worldwide [4]. Hence, there is a strong motivation to investigate solutions to bring down the energy consumption and the cost of cellular networks, thus yielding both environmental and financial gains. Given that cellular networks are dimensioned according to peak-hour traffic demands, an effective approach towards this direction is to temporarily switch off part of the BS infrastructure that remains underutilized when the network traffic is low. The coexistence of multiple MNOs in the same geographical area [5], due to legal regulations that obligate them to install their antennas on the same buildings, has motivated a new business model, known as infrastructure sharing [6], [7]. This new paradigm embraces a set of strategies that enable the MNOs to use their resources jointly to reach their common goal, which is to guarantee user service while achieving energy and cost reduction. Infrastructure sharing is classified into three categories [8]: i) passive sharing of sites, masts and building premises, ii) active sharing of the active network com- ponents such as antennas, switches and backhaul equipment, and iii) roaming-based sharing, where the MNOs share the cell coverage for a pre-negotiated time period. In this paper, motivated by the aforementioned issues, we propose a roaming-based infrastructure sharing scheme, applicable in multi-operator environments during low traffic periods. Taking into account the rationality of the MNOs and their conflicting interests, we introduce a game theo- retic framework that enables the MNOs to make individual switching off decisions for their own BSs, thus bypassing potential complicated service level agreements among them. Besides the expected energy efficiency benefits, the proposed scheme allows the MNOs to significantly reduce their financial costs independently of the strategies of the coexisting MNOs, providing them with the incentives to participate in the game. Our contribution is summarized as follows: 1) As a part of an integrated roaming-based infrastructure sharing scheme for multi-operator environments, we in- troduce a game theoretic switching off algorithm that aims at minimizing the individual MNO cost in a dis- tributed manner. We define a realistic cost function that explicitly considers actual roaming and operational costs for the MNOs. We show that, in the proposed game, a