Cell based fair resource allocation in fixed clustered cellular systems using a genetic algorithm

In this paper we consider the uplink of a cellular network partitioned into localized jointly decoded cells. These jointly decoded cells are implemented as fixed size clusters. Such networks have a potential for real world deployments with improved spectral efficiency and user experience. Similar to conventional cellular networks, frequency planning can be considered as an efficient method to control interference between cells belonging to different clusters. Here we consider frequency planning in the form of allocating set of frequency bins to cells within clusters. As service providers are more interested in providing QoS, the considered bin allocation apart from maximizing system throughput, should also allocate cell resources in a fair manner. We propose a new cell based QoS balancing function which helps to maximize sum rate as well as achieve cell based fairness using both coupled and decoupled power allocation schemes. To implement this function, we use SIC in order to derive cell based sum rate. The derived formulation is conditioned for both hard and soft fair1ness constraints. This is then applied as input to a Genetic Algorithm in order to optimize the derived network wide QoS balancing function. Numerical results indicate that under wide range of bandwidth conditions, and in densely located cells employing decoupled power allocation, resources are more fairly allocated than in cells employing coupled power allocation.