The demand for wireless services and the need for high data-rates are growing rapidly. Future generation networks are expected to provide high data-rates in the order of 1Gbps in local area and 100Mbps in wide area. It is a challenge for operators to meet this rising demand for high data-rates as the radio-spectrum is an expensive and scarce resource. In the last World Radio Communication conference (WRC'07), less than 600 MHz bandwidth has been allocated to mobile communication systems. When considering the total predicted bandwidth demand of 1200MHz - 1700MHz in 2020, it is clear that there is a spectrum scarcity for mobile communication systems. This scarcity arises due to the exclusive allocation of the spectrum among different Radio Transmission Systems (RTS). A further exclusive splitting of spectrum among different operators leads to an apparent scarcity of the resources. While doing so, it should be clear that no operator will suffice in its own to meet the rising demand for high data-rates, when the current traditional way of spectrum utilization continues. Based on the arguments mentioned, the idea of spectrum sharing was born. When the operators share their licensed spectrum bands, they simply will reach higher bandwidths, the spectrum will always be utilized when an operator does not utilize it. Spectrum sharing among cellular operators introduces a new concept of Inter-Operator Interference (Inter-OI). Interference which is a natural result of operating in the same common spectrum band limits the capacity obtained from the spectrum. Therefore, it should be mitigated in an intelligent way. As opposed to other interference generation mechanisms known in wireless-networking, Inter-OI is a problem of two independent networks with conflicting objectives on the common resource. When this conflict is not solved, the advantages may turn into disadvantages. To solve the Inter-OI in the uplink and downlink of the involved cellular networks, there are some considerations that one has to take into account. First of all, information exchange: How much information can we gather about the interfering signals? There are two extreme cases possible: When we do not know anything about the interfering signals, we can make a Gaussian Random Signal Approximation which is not a realistic model of Inter-OI as it can be more severe due to the overlapping-cells of two different cellular networks. When we could get the whole interfering signal structure, we could jointly construct the signal or pre-cancel it in a multi-cell processing-fashion. However, due to the limited information exchange capability of the operators, full information exchange is not desirable. Once the exchangeable amount of information is fixed, the solution should provide enough efficiency to satisfy the operators above their non-sharing case. Furthermore, the solution should provide fairness among the operators. Of course, all should occur within an acceptable complexity. In this thesis, to cover the considerations mentioned above, a possible solution for the uplink-problem has been proposed by the means of a receive-beamforming approach for which the basestations need the Channel State Information (CSI) of the direct channels to their intended users and crosstalk channels to their non-intended users. To capture the needed CSI in this heterogeneous environment, the mobile terminals have been given user-specific pilots which are recognized by the basestations. For this approach, registration to both operators is required in order to capture the CSI while the users get the service from their own operator. For the downlink case, a transmit-beamforming approach has been proposed. The downlink-part of the problem is different. In this case, there are two base-stations, two decision-makers with conflicting objectives. Resource sharing problems with multiple decision-makers are subjected to Game Theory of the Applied Mathematics. Game Theory provides tools to predict the results of selfish and cooperative actions in a resource sharing problem. Instead of applying their best-response strategies selfishly, this thesis has proposed to apply SLNR-based beamforming for the beamforming-dimension of the problem and to apply the power levels in a leader-follower relationship as described in the literature. The needed objective functions have been constructed for the leader operator and follower operator by the means of capacity functions and have been solved as a non-convex optimization problem. The proposed approach has been simulated in a realistic scenario with i.i.d. Rayleigh Fading. The results have been shown to be satisfactory in comparison to the non-sharing case qua efficiency and fairness.