An Evolutionary Game-based Approach to Vehicular Task Offloading in An MEC Environment

In the vehicle task offloading scenario, this study proposes a task offloading method based on a probabilistic performance-aware evolutionary game strategy to solve the problem of insufficient multi-task offloading efficiency in a cloud-edge hybrid environment. First, considering the time-varying fluctuation characteristics of edge server performance in a mixed environment composed of a central server and multiple edge servers, we use a method based on probabilistic performance-aware evolutionary game strategy to conduct probabilistic analysis on the historical performance data of edge cloud servers. Build an evolutionary game model. Next, we generate an evolutionary stable strategy (ESS) for service offloading to ensure each user performs task offloading with high satisfaction guaranteed. Finally, we compare and test different methods using the cloud edge resource location dataset and the cloud service performance test dataset by conducting simulation experiments on 24 consecutive time windows. Through simulation experiments, our framework shows better results than existing methods on multiple metrics.