Novel throughput quantile averaging methods for the proportional fair algorithm in indoor mobile networks

Smart optimisation of radio resource distribution techniques is crucial in improving the overall performance of Next-Generation Networks (NGNs). Since higher data rates, lower latency and increased fairness for users are the main goals, an efficient scheduling algorithm has to be used. In this paper an evaluation of the performance of an indoor network is realised using the Realistic Indoor Environment Generator (RIEG), when 20 different Throughput Quantile Averaging Methods (TQAMs) for the downlink Resource Allocation Algorithm (RAA) called Proportional Fair (PF) are used. The TQAMs are compared by using a trustworthy metric called Comparative Factor (CF), which evaluates the overall performance of the mobile network, simultaneously taking into account the average throughput of the indoor users, the fairness of resource distribution and the outage ratio. Each type of the TQAMs takes a different part of the previous throughputs to evaluate the average throughput used by the PF scheduling algorithm in order to allocate network resources to the users. The experimental results show that the TQAMs which incorporate bigger groups of previous throughputs are more stable when the number of users increases. When the min-max groups are taken into account an increase of overall throughput is observed, while using the min or max groups leads to an increase in fairness. The performance of the cell-edge users is improved the most when the max or middle part of the previous throughputs is taken into account.