PASS-Enhanced MEC: Joint Optimization of Task Offloading and Uplink PASS Beamforming
2025
Read PASS-Enhanced MEC: Joint Optimization of Task Offloading and Uplink PASS Beamforming (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-22948.txt.
Introduction
Zhaoming Hu, with Ruikang Zhong, Xidong Mu, Dengao Li, and Yuanwei Liu, studies how a new antenna design could speed up edge computing over wireless links. Pinching antennas sit on waveguides and can shift position to form short direct links. That helps cut signal loss and blockage at high frequencies.
The team jointly chooses where tasks should run and how the uplink beams should form, aiming to cut total delay. They frame the choice as a learning problem and solve it with a load-aware policy method. Their results suggest the design could make offloaded phone and device tasks finish faster.
Catalog record
| YEAR | 2025 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-22948
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
PASS-Enhanced MEC: Joint Optimization of Task Offloading and Uplink PASS Beamforming Abstract A pinching-antenna system (PASS)-enhanced mobile edge computing (MEC) architecture is investigated to improve the task offloading efficiency and latency performance in dynamic wireless environments. By leveraging dielectric waveguides and flexibly adjustable pinching antennas, PASS establishes short-distance line-of-sight (LoS) links while effectively mitigating the significant path loss and potential signal blockage, making it a promising solution for high-frequency MEC systems. We formulate a network latency minimization problem to joint optimize uplink PASS beamforming and task offloading. The resulting problem is modeled as a Markov decision process (MDP) and solved via the deep reinforcement learning (DRL) method. To address the instability introduced by the max \max operator in the objective function, we propose a load balancing-aware proximal policy optimization (LBPPO) algorithm. LBPPO incorporates both node-level and waveguide-level load balancing information into the policy design, maintaining computational and transmission delay equilibrium, respectively. Simulation results demonstrate that the proposed PASS-enhanced MEC with adaptive uplink PASS beamforming exhibit stronger convergence capability than fixed-PA baselines and conventional MIMO-assisted MEC, especially in scenarios with a large number of UEs or high transmit power. Zhaoming Hu 1 2 , Ruikang Zhong 3 , Xidong Mu 4 , Dengao Li 1 2 , Yuanwei Liu, 5 I Introduction With the rapid evolution toward beyond fifth-generation (B5G) and sixth-generation (6G) wireless networks, emerging applications such as immersive metaverse, intelligent transportation, and industrial automation impose stringent requirements on both communication and computation performance. Massive numbers of heterogeneous user equipments (UEs) continuously generate intensive workloads that demand high data rates, ultra-low latency, and reliable service guarantees. Traditional cloud computing paradigm, although powerful in computational capability, is limited by long-distance backhaul transmission, which introduces excessive delays and is inadequate for delay-sensitive and computation-intensive applications. To address this issue, mobile edge computing (MEC) has been proposed as a promising paradigm that brings computing resources closer to UEs by deploying servers at the network edge, such as access points (APs) and base stations (BSs) [ 1 ] . By offloading computation tasks to nearby MEC servers, mobile terminals can significantly reduce task execution latency and alleviate their energy consumption constraints, thereby improving the overall quality of service (QoS). The performance of MEC is critically dependent on wireless link quality, which is susceptible to user mobility, blockages, and fading, making reliable connectivity a key challenge. Advanced antenna technologies are fundamental to enhancing link reliability and spectral efficiency. While conventional approaches like massive multiple-input multiple-output (MIMO) [ 2 ] use spatial diversity to improve capacity, and emerging technologies like reconfigurable intelligent surfaces (RISs) dynamically shape the propagation environment, they face limitations in dynamic MEC scenarios. Issues such as hardware complexity, limited reconfiguration agility, and poor adaptability to mobile users motivate the development of more flexible and efficient wireless solutions. As a groundbreaking architecture recently prototyped by NTT DOCOMO [ 3 ] , the pinching-antenna system (PASS) offers a hardware-reconfigurable solution. It overcomes the key limitations of conventional multi-antenna systems, namely free-space path loss and line-of-sight (LoS) blockage. Unlike traditional approaches, PASS employs low-loss dielectric waveguides as its primary signal transport medium [ 4 ] . Small dielectric particles, referred to as pinching antennas (PAs), can be dynamically attached or detached at arbitrary positions along the waveguide. This “pinching” mechanism enables flexible signal radiation into free space, effectively forming near-wired LoS links when PAs are placed in close proximity to users. Such a design not only drastically reduces propagation loss but also allows real-time adjustment of the number and layout of antennas without complex radio frequency (RF) circuitry. Recent studies have begun to explore the communication potential of PASS. [ 5 ] evaluated the performance of PASS with different multiple access technologies, while [ 6 ] investigated a secure wireless communication framework. These works highlight the advantages of PASS in supporting scalable, low-cost, and high-quality connectivity, making it a highly suitable candidate for enhancing MEC systems in dynamic wireless environments. Motivated by the above observations, this paper investigates a PASS-enhanced MEC system, where a network latency minimization problem is formulated to joint optimize uplink PASS beamforming and task offloading. As a long-term stochastic optimization, the problem can be modeled as a Markov decision process (MDP) and effectively addressed using deep reinforcement learning (DRL), which is well suited for sequential decision-making in dynamic environments. Nevertheless, the presence of \max operator in the objective function often leads to instability and slow convergence. To overcome this issue, we propose a load balancing-aware proximal policy optimization (LBPPO) algorithm. By incorporating both node-level and waveguide-level load balancing information into the policy design, LBPPO enhances training stability and convergence, and improving latency performance in PASS-assisted MEC systems. II System Model and Problem Formulation Fig. 1: System model of PASS-enabled MEC system. As illustrated in Fig. 1 , we consider a PASS-enhanced wireless MEC system. The system consists of a BS equipped with N waveguides and K UEs with L antennas. In addition, there are M PAs on each waveguide. Considering a three-dimensional (3D) Cartesian coordinate system, we assume that the users are located on the x-y plane with a squared area of d_{x}\times d_{y} , and the system operates over an extended temporal horizon comprising T discrete time slots. The \mathbf{p}_{k,\tau}=(x_{k,\tau},y_{k,\tau},0) denotes the position of the k -th UE at time slot t , where \forall k\in\mathcal{K}=\left\{1,2,...,K\right\} , and \forall\tau\in\mathcal{T}=\left\{1,2,...,T\right\} . Specifically, x_{k,\tau} follows a uniform distribution in \left[0,d_{x}\right] and y_{k,\tau} follows a uniform distribution in \left[-\frac{d_{y}}{2},\frac{d_{y}}{2}\right] . Without the loss of generality, it is assumed that the waveguides with the length of d_{x} is deployed at the height of H and in parallel to the x -axis. The location of the m -th PA on the n -th waveguide at time slot \tau is denoted as \mathbf{p}_{m,n,\tau}=(x_{m,n,\tau},y_{m,n,\tau},H) , and all m PAs on the same waveguide n share the same y -coordinate, i.e., y_{1,n,\tau}=y_{2,n,\tau}=...=y_{M,n,\tau} for all n\in\mathcal{N} and \tau\in\mathcal{T} . In the proposed system, UEs are equipped with wireless communication circuits and computing processors that have limited computational capabilities, and they handle computation tasks involving substantial amounts of input data (measured in bits). These tasks can be processed locally by the UEs or offloaded to the BS, which is equipped with an MEC server. The PASS, which can transmit an incident signal, is deployed to assist the UEs’ task offloading. Given the negligible bandwidth consumption of downlink result delivery in MEC systems, our subsequent analysis will focus on the uplink PASS scenario. II-A Signal Model In consideration of large-scale channel characteristics, the analysis is confined to LoS channels to explore fundamental performance limits. Within this framework, the small-scale fading model is simplified to focus on beamforming dynamics. The channel gain is normalized to exclude large-scale path loss, while phase variations induced by relative UE movement are explicitly modeled through propagation distance differences across the antenna array. Specifically, consider a uniform linear array (ULA) of L antennas at the UE, aligned along the y -axis with its centroid at \mathbf{p}_{k,\tau} . The location of the l -th antenna is then given by \displaystyle\mathbf{p}_{k,l,\tau}=\left(x_{k,\tau},\ y_{k,\tau}+\delta_{l},\ 0\right),\ \delta_{l}=\left(l-\frac{L+1}{2}\right)\Delta_{l}, (1) where \Delta_{l} denotes the UE’s antenna spacing. Under these conditions, the channel matrix \mathbf{H}_{k,n,\tau}\in\mathbb{C}^{M\times L} between the PA on the n -th waveguide and the