| Abstract |
DySOFT (Dynamic Auction with Stochastic Optimization for Flexible Time Datacenters) is a truthful reverse auction-based framework for managing vehicular cloud resources in smart cities. It addresses the dynamic challenges of resource availability and task scheduling in parking lot-based vehicular clouds, also known as Flexible Time Datacenters. DySOFT uses machine learning to predict vehicle capabilities and task completion times, enabling efficient resource allocation and mitigating the risks of early vehicle departures through preemptive migration strategies. Simulation results highlight the framework's adaptability to unpredictable conditions, optimizing resource utilization, minimizing task delays, and ensuring reliable system performance. © 2024 IEEE. |
