On the Design of Federated Learning in Latency and Energy Constrained Computation Offloading Operations in Vehicular Edge Computing Systems

With the advent of smart vehicles, several new latency-critical and data-intensive applications are emerged in Vehicular Networks (VNs). Computation offloading has as a viable option allowing to resort nearby edge servers for remote processing within requested service latency requirement. Despite advantages, computation over resource-limited servers, together with vehicular mobility, is still challenging problem be solved. In particular, order avoid additional due out-of-coverage operations, Users (VUs) mobility introduces bound on amount data offloaded towards servers. Therefore, approaches have been used finding correct offloaded. Among others, Federated Learning (FL) highlighted one most promising solving techniques, given privacy concerns VNs limited communication resources. However, FL consumes resources during its operation therefore incurs an burden resource-constrained VUs. this work, we aim optimize VN performance terms energy consumption by considering both processes while selecting proper number iterations implemented. To end, first propose FL-inspired distributed learning framework VNs, then develop constrained optimization jointly minimize overall consumed. An evolutionary Genetic Algorithm proposed in-hand compared some benchmarks. The simulation results show effectiveness approach consumption.