Driven by data hungry applications, the architecture of mobile networks is moving towards that of densely deployed cells where each cell may use a different access technology as well as a different frequency band. Next generation networks (NGNs) are essentially identified by their dramatically increased data rates and their sustainable deployment. Motivated by these requirements, in this thesis we focus on (i) capacity maximisation, (ii) energy efficient configuration of different classes of radio access networks (RANs). To fairly allocate the available resources, we consider proportional fair rate allocations. We first consider capacity maximisation in co-channel 4G (LTE) networks, then we proceed to capacity maximisation in mixed LTE (including licensed LTE small cells) and 802.11 (WiFi) networks. And finally we study energy efficient capacity maximisation of dense 3G/4G co-channel small cell networks. In each chapter we provide a network model and a scalable resource allocation approach which may be implemented in a centralised or distributed manner depending on the objective and network constraints.