Joint RAN Slicing and Computation Offloading for Autonomous Vehicular Networks: A Learning-Assisted Hierarchical Approach

In this paper, a two-timescale radio access network (RAN) slicing and computing task offloading problem is investigated for cloud-enabled autonomous vehicular (C-AVN). We aim at jointly maximizing the communication resource utilization with diverse quality-of-service (QoS) guarantee driving tasks. Specifically, to capture small-timescale dynamics, scheduling formulated as stochastic optimization program, long-term network-wide computation load balancing minimum variations. Due large size unavailable state transition probabilities, we employ cooperative multi-agent deep Q -learning (MA-DQL) fingerprint solve by learning set of stationary policies stabilized convergence. Given decisions, further study RAN in timescale, which convex program. focus on optimizing ratios among base stations, maximize aggregate utility statistical QoS provisioning Based impact balancing, propose hierarchical framework both utilization. Extensive simulation results are provided demonstrate effectiveness proposed comparison state-of-the-art schemes.