Auction-Based Scheduling for Efficient Execution of Stochastic Tasks in Diverse Vehicular Clouds with Flexible Time Datacenter Integration
Syed R. Rizvi, Susan Zehra, Stephan Olariu, Samy El-Tawab · 2024
In the realm of vehicular cloud computing, the Flexible Time Datacenter (FTDC) addresses the dynamic and unpredictable nature of vehicle availability and computational resource allocation. This paper presents the FTDC framework, which leverages a truthful reverse auction mechanism for efficient and strategic resource allocation between vehicles and tasks. The framework ensures that vehicles with computational resources are effectively matched with tasks based on their specific requirements. A key component of this framework is its robust task mitigation strategy, which addresses challenges arising from unexpected vehicle departures or task interruptions. This strategy involves dynamic task reallocation, where tasks are promptly reassigned to alternative vehicles or re-auctioned to ensure deadlines are met. The system incorporates penalties for premature departures and incentives for reliable behavior to maintain system integrity and efficiency. The FTDC framework is evaluated through a series of simulations that demonstrate its effectiveness in managing dynamic task allocation, resource utilization, and system reliability. The results highlight the framework's ability to adapt to unpredictable conditions while maintaining high performance and minimizing task delays.