Two-Timescale Task Scheduling, Resource Allocation and Service Deployment for Collaborative MEC Networks: A Hierarchical Method

Collaborative mobile edge computing (MEC) networks achieve higher efficiency through cooperative computing among interconnected edge servers (ESs), which necessitates periodic configurations of computing modules on ESs, i.e., service deployment. A novel hierarchical collaborative MEC network framework is proposed in this paper, where pre-configured MEC services are deployed according to both long-timescale deployment capacity and short-timescale changes in mobile channels and user requests. The key objective is to minimize the long-term weighted sum of service deployment cost and the system delay. To solve the two-timescale optimization problem, a hierarchical dual-layer algorithm is proposed: 1) The short-timescale decisions are solved iteratively through block coordinate descent in the inner-layer. 2) By observing the inner-layer solutions, the long-timescale problem is reformulated as a binary problem and solved using branch-and-bound framework augmented with cutting plane technique. Based on this nested optimization framework, we achieve a convergence-guaranteed local optimum solution for the two-timescale NP-hard problem. Numerical results show that the proposed algorithm outperforms benchmarks in both system cost and fairness, achieving superior optimization trade-offs between deployment cost and system delay.