Task Offloading and Resource Allocation Using Novel Deep Learning‐Based Optimization in Vehicle Edge Networks

ABSTRACT Within the internet of things (IoT), the vehicle edge network (VEN) has emerged as a novel research hotspot. However, offloading computational tasks from vehicles to edge servers can introduce additional processing and communication delays, which may significantly reduce the quality of service (QoS) in VENs. Hence, this paper proposes task offloading and resource allocation (TORA) in VENs using a novel intelligent deep learning‐based optimization methodology. The delay associated with the offloading task to edge servers as well as the mathematical model is described. Here, the fitness function is performed by the novel enhanced gated recurrent unit (EGRU), in which the parameters of GRU are tuned by the optimization algorithm called stork optimization algorithm (SOA). The optimal solution is finally attained by the iteration process. Contrast studies are carried out to confirm the effectiveness of this approach. The experimental findings demonstrate the superiority of the proposed EGRU‐SOA over other systems in terms of delay reduction and QoS improvement. The proposed EGRU‐SOA for the TORA in VEN model is 32.14%, 48.15%, 48.65%, 64.88%, 20.69%, 69.57%, and 73.68% better than other considered existing methods in terms of cost, resource utilization, task rejection rate, delay, energy consumption, response time, and task offloading percentage, respectively.