NeFut Logo NeFut
Admin Login

[CS.AI] Beyond-Diagonal RIS Empowered Heterogeneous Edge Computing: A New Distributional RL Approach

Published at: 2026-07-17 22:00 Last updated: 2026-07-18 08:19
#AI #Machine Learning #optimization

Abstract

Active beyond-diagonal reconfigurable intelligent surfaces (BD-RISs) enable hybrid transmitting and reflecting mode for effective signal amplification and full-space coverage, presenting a promising solution for blockage-aware uplink offloading in heterogeneous mobile edge computing (MEC) systems. However, practical hybrid active BD-RISs realized by reciprocal devices generate cross-sector energy leakage, reshaping the system-level energy-latency tradeoff. This paper studies energy-aware offloading and resource allocation for reciprocal active BD-RIS-assisted heterogeneous MEC, where offloading decisions, CPU/GPU computation allocation, transmit powers, receive processing, and active BD-RIS are tightly coupled.

Method

The resulting problem is a high-dimensional mixed integer nonconvex problem, which is challenging for conventional per-instance optimization. To tackle this, we develop an end-to-end joint optimization framework based on a refined distributional soft actor-critic algorithm, termed DSAC-T. By modeling return distributions instead of only expected values, DSAC-T enhances policy stability under reward heterogeneity and feasibility-boundary sensitivity.

Results

Compared to other baseline algorithms, DSAC-T achieves the best energy-latency reward, with a feasibility ratio of 81.67%, and a fast online decision time of 0.0267 s per scenario.

Blogger's Review: The DSAC-T algorithm proposed in this paper excels in complex heterogeneous edge computing environments, especially in managing the trade-offs between energy and latency. By integrating distributional reinforcement learning, it significantly enhances policy stability and decision-making efficiency, pointing the way forward for future edge computing research.

Original Source: https://arxiv.org/abs/2607.13160

[h] Back to Home