Heterogeneous Machine-Type Communications in Cellular Networks: Random Access Optimization by Deep Reinforcement Learning

One of the significant challenges for managing machine-to-machine (M2M) communication in cellular networks, such as LTE-A, is the overload of the radio access network due to very many machine type communication devices (MTCDs) requesting access in burst traffic. This problem can be addressed well by applying an access class barring (ACB) mechanism to regulate the number of MTCDs simultaneously participating in random access (RA). In this regard, here we present a novel deep reinforcement learning algorithm, first for dynamically adjusting the ACB factor in a uniform priority network. The algorithm is then further enhanced to accommodate heterogeneous MTCDs with different quality of service (QoS) requirements. Simulation results show that the ACB factor controlled by the proposed algorithm coincides with the theoretical optimum in a uniform priority network, and achieves higher access probability, as well as lower delay, for each priority class when there are heterogeneous QoS requirements.