.png){kind=icon}
Return
Parking Vehicle-Assisted Task Offloading in Edge Computing: An Environment-Adaptive Multi-Strategy Dynamic Multi-Objective Evolutionary Algorithm
.png){kind=icon}
PRE
AI
DOI：10.1109/tevc.2026.3740890.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Vehicle edge computing (VEC) supports low-latency, computation-intensive vehicular applications. To address the challenges of limited computational resources and suboptimal task offloading in the Internet of Vehicles (IoV) environment, this paper proposes a service architecture that integrates edge servers (ESs) with the resources of parked vehicles (PVs). By grouping PVs into parked vehicle clusters (PVCs) alongside roadside units (RSUs) as service nodes, the coverage of edge computing is extended. Tasks are modeled as directed acyclic graphs (DAGs) to capture dependencies, and the offloading problem is formulated as a dynamic multi-objective optimization (DMOP) that minimizes both average computational latency (ACL) and average energy consumption (AEC). To solve this problem, we propose a transformer-based, environmentadaptive, multi-strategy dynamic multi-objective evolutionary algorithm (DMOEA/D-TBEAMS). The algorithm introduces a multi-dimensional environmental change perception mechanism and a multi-strategy fusion response mechanism, which adaptively adjusts the response according to the degree of change. In the static optimization phase, the Pareto adaptive scalarizing method (PaS) is used to balance convergence and diversity. Simulation results show that compared to mainstream algorithms, the proposed algorithm achieves faster convergence speed, better solution set distribution, and improved stability in dynamic task offloading scenarios, demonstrating an effective trade-off between latency and energy consumption.
Keywords:
Computation Offloading
Dynamic Multi-Objective Optimization
Evolutionary Algorithm
Vehicle Edge Computing
Vehicular Collaboration
Journal
IF:
12
Papers:
1.9K
Citations:
2.4W
Organization
View more organizations
Cited Papers
No cited papers available