Decentralized Heterogeneous Statistical QoS Provisioning for Uplinks over 5G Wireless Networks

Abstract

The newly imposed heterogeneous statistical delay- bounded quality of service (QoS) provisioning, which refers to the different/variable delay-bounded QoS guarantees among different wireless links, for the fifth-generation (5G) mobile multimedia wireless networks has received much research attention recently. The heterogeneous statistical delay- bounded QoS provisioning can be classified into two categories: centralized and decentralized heterogeneous statistical delay-bounded QoS provisioning where the delay-bounded QoS requirements are centralized at the base station (BS) and distributed at different mobile user equipments (UEs), respectively. The decentralized heterogeneous statistical delay-bounded QoS provisioning is more challenging than the centralized heterogeneous statistical delay-bounded QoS provisioning. In this paper, we build up the system model for decentralized heterogeneous statistical delay-bounded QoS provisioning in terms of aggregate effective capacity and heterogeneous QoS exponents. Based on this model, we develop the optimal joint bandwidth and power allocations scheme to maximize the aggregate effective capacity of uplink transmissions supporting the heterogeneous statistical delay-bounded QoS provisionings over 5G mobile multimedia wireless networks. The BS allocates the bandwidth based on the QoS exponents of all uplinks while the mobile UEs dynamically allocate the power based on the corresponding instantaneous channel state information (CSI), the corresponding QoS exponent, and the bandwidth allocated by the BS. We conduct the extensive simulations to validate and evaluate our developed optimal joint bandwidth and power allocations schemes, showing that our proposed heterogeneous statistical delay-bounded QoS provisioning scheme can significantly increase the aggregate effective capacity as compared with the homogeneous statistical delay-bounded QoS provisioning schemes.