CCO-DSS: A Cache-assisted Computation Offloading Based on DAG-structured Subtask Sorting

With the rapid development of Vehicular Edge Computing (VEC), efficiently managing task offloading and caching under task dependencies remains a critical challenge. Traditional cloud computing incurs high latency, whereas individual edge nodes struggle to handle large-scale dynamic workloads. Moreover, task dependencies complicate resource allocation and execution order, highlighting the need for intelligent coordination mechanisms. To address these challenges, this paper proposes a novel scheme called Cache-assisted Computation Offloading based on DAG-structured Subtask Sorting (CCO-DSS), which integrates task scheduling, cache management, and offloading decision-making. It combines dynamic priority-based scheduling, cooperative cache optimization, and a GA3C-based deep reinforcement learning strategy to enhance resource utilization while ensuring adherence to task execution order. Compared with baseline methods, simulation results demonstrate that CCO-DSS improves task throughput and reduces total system cost.