Radio resource allocation algorithms for multi-service OFDMA networks: the uniform power loading scenario

Abstract

Adaptive Radio Resource Allocation is essential for guaranteeing high bandwidth and power utilization as well as satisfying heterogeneous Quality-of-Service requests regarding next generation broadband multicarrier wireless access networks like LTE and Mobile WiMAX. A downlink OFDMA single-cell scenario is considered where heterogeneous Constant-Bit-Rate and Best-Effort QoS profiles coexist and the power is uniformly spread over the system bandwidth utilizing a Uniform Power Loading (UPL) scenario. We express this particular QoS provision scenario in mathematical terms, as a variation of the well-known generalized assignment problem answered in the combinatorial optimization field. Based on this concept, we propose two heuristic search algorithms for dynamically allocating subchannels to the competing QoS classes and users which are executed under polynomially-bounded cost. We also propose an Integer Linear Programming model for optimally solving and acquiring a performance upper bound for the same problem at reasonable yet high execution times. Through extensive simulation results we show that the proposed algorithms exhibit high closeto-optimal performance, thus comprising attractive candidates for implementation in modern OFDMA-based systems.Define the following subchannels and users subsets: J CBR = {k ∈ U CBR : lhs k b k } ⊆ U CBR : the subset of CBR users for which minimum QoS is met.J ′ CBR = U CBR \J CBR : the complementary subset of undersatisfied CBR users. J BE = {k ∈ U BE : lhs k > 0} ⊆ U BE : the subset of BE users allocated at least one subchannel or equivalently posses non-zero data rate. J = J CBR ∪ J BE : the subset of users from which we can remove subchannels J ′ = J ′ CBR : the subset of users to which we must add subchannels in order to meet the minimum QoS requirements. N = n ∈ S, k ∈ J : ρ n,k = 1 : the subchannels pool