Federated learning is a pivotal technique in vehicular edge computing (VEC), which allows distributed machine learning across vehicles while preserving data privacy. However, it is critical to select the vehicles to participate in the learning process effectively. In this context, the problem related to finding optimized vehicles to participate in the learning task is discussed. An optimized vehicle selection mechanism is needed in vehicular networks, as traffic flow occurs in a real environment, influenced by different conditions such as security, vehicle capacities, data freshness, weather conditions, etc. The primary key contributions of this research are as follows: (i) formulation of vehicle selection in vehicular edge computing as an optimization problem, and (ii) the development of a new approach using the Tabu search algorithm named TS-EVS, specifically designed for the problem, as it is shown to be NP-hard. We evaluated TS-EVS using a Python implementation, considering the outcome of the proposal in two scenarios taking into account the MEC applications and the number of deployed vehicles. Compared to the untrusted version of the proposal, the numerical results demonstrate how the suggested approach significantly improves accuracy and reduces learning time.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Vehicular Edge Computing: An Enhanced Vehicle Participant Selection System for Federated Learning




    Publication date :

    2025-05-12


    Size :

    898311 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English






    Blockchain-Enabled Federated Learning for Enhanced Collaborative Intrusion Detection in Vehicular Edge Computing

    Abou El Houda, Zakaria / Moudoud, Hajar / Brik, Bouziane et al. | IEEE | 2024


    Adaptive and Parallel Split Federated Learning in Vehicular Edge Computing

    Qiang, Xianke / Chang, Zheng / Hu, Yun et al. | ArXiv | 2024

    Free access