Distributionally Robust Optimization for Computation Offloading in Aerial Access Networks

TL;DR

The paper tackles computation offloading in aerial access networks (AANs) with MEC under uncertain task sizes. It builds a distributionally robust optimization framework using historical-data-based uncertainty sets and develops the MEC-based distributionally robust latency optimization algorithm (MDRLOA), transforming the problem into a tractable linear program via relaxation and dualization. Key contributions include an $L_1$-based uncertainty-set construction, a reformulation of the minimax problem into solvable subproblems, and an algorithm that integrates offloading decisions, UAV/HAP computation, and energy constraints. Empirical results show that MDRLOA achieves approximately 18.8% latency reduction compared with deterministic offloading and about 6.9% energy savings versus a robust baseline, highlighting its robustness and practical impact for remote-area MEC in 6G networks.

Abstract

With the rapid increment of multiple users for data offloading and computation, it is challenging to guarantee the quality of service (QoS) in remote areas. To deal with the challenge, it is promising to combine aerial access networks (AANs) with multi-access edge computing (MEC) equipments to provide computation services with high QoS. However, as for uncertain data sizes of tasks, it is intractable to optimize the offloading decisions and the aerial resources. Hence, in this paper, we consider the AAN to provide MEC services for uncertain tasks. Specifically, we construct the uncertainty sets based on historical data to characterize the possible probability distribution of the uncertain tasks. Then, based on the constructed uncertainty sets, we formulate a distributionally robust optimization problem to minimize the system delay. Next,we relax the problem and reformulate it into a linear programming problem. Accordingly, we design a MEC-based distributionally robust latency optimization algorithm. Finally, simulation results reveal that the proposed algorithm achieves a superior balance between reducing system latency and minimizing energy consumption, as compared to other benchmark mechanisms in the existing literature.

Figures (3)

• Figure 1: Computation offloading in an AAN.
• Figure 2: Performance of different optimization mechanisms.
• Figure 3: Performance of MDRLOA with different parameters.