Secured planning and scheduling in blockchain for the Internet of Vehicles using adaptive deep actor-critic reinforcement learning with deformable attention procedures

An efficient, secure planning and scheduling framework for IoV is designed with blockchain. In the developed model, Adaptive Deep Actor-Critic Reinforcement Learning with Deformable Attention (ADACRL-DA) is employed to execute the secured scheduling and path planning in blockchain for the IoV. In the developed ADACRL-DA, different parameters are tuned using Enhanced Exploration-based Magnificent Frigatebird Optimization (EE-MFO), which helps to enhance the secure planning and scheduling process. In the proposed framework, traffic situation, task processing, and communication are tuned using MFO, which helps to reduce computation latency as well as travel time. Later, scheduling issues in various vehicles are converted into Markov Decision Processes (MDP), and also proactive load balancing is accomplished through ADACRL-DA among road infrastructure. Later, multiple experiments are executed in the developed framework over the classical schemes to verify the overall efficiency while scheduling and path‑planning processes are executed.