Computational Resource Management of Edge Clouds for Vehicle-to-Network Services With Resource Limit
- 20 March 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automation Science and Engineering
- Vol. 23 (15455955) , 7379-7395
Abstract
This paper studies the online management of the computational resources between multiple edge clouds to minimize the operational cost for a vehicle-to-network (V2N) service provider, subject to stochastic trajectories of vehicles, the quality-of-service (QoS) and the resource limit. Due to the random mobility of vehicles, it poses challenges in real-time migration management when the computational capacity of each edge cloud and the stringent delay requirement of V2N services are both constrained, resulting in strong temporal-spatial coupling of the migration decisions. The problem could even be intractable to solve as the number of vehicles grows. To tackle these issues, we first propose a multi-layered cloud framework to gather and coordinate migration information between vehicles. Then, a multiagent rollout with feasibility construction approach is developed, where the migration decision can be optimized sequentially for each vehicle based on the coordinated migration information from the central cloud. To handle the potentially infeasible solutions caused by the computational resource constraint, we propose a heuristic method to determine the migration priority at each overloading edge cloud, and an integer program is constructed that can be easily solved to produce feasible solutions. The numerical results show the efficiency of the proposed approach in both economical performance and computational complexity compared to the benchmarks.Keywords
Funding Information
- National Key Research and Development Program of China (2023YFB2407900)
- National Natural Science Foundation of China (623B2083; 62192752; 62192755)
This publication has 57 references indexed in Scilit:
- Optimizing Energy Efficiency in Vehicular Edge-Cloud Networks Through Deep Reinforcement Learning-Based Computation OffloadingIEEE Access, 2024
- Network slicing with load-balancing for task offloading in vehicular edge computingVehicular Communications, 2022
- A Probabilistic Architecture of Long-Term Vehicle Trajectory Prediction for Autonomous DrivingEngineering, 2022
- Reducing the network overhead of user mobility–induced virtual machine migration in mobile edge computingSoftware: Practice and Experience, 2018
- A Survey on Virtual Machine Migration: Challenges, Techniques, and Open IssuesIEEE Communications Surveys & Tutorials, 2018
- Mobility prediction model-based service migration procedure for follow me cloud to support QoS and QoE2016
- Heterogeneous Vehicular Networking: A Survey on Architecture, Challenges, and SolutionsIEEE Communications Surveys & Tutorials, 2015
- Impact of Cloudlets on Interactive Mobile Cloud Applications2012
- Statistical inference for time varying origin–destination matricesTransportation Research Part B: Methodological, 2008
- Rollout Algorithms for Combinatorial OptimizationJournal of Heuristics, 1997