Distributed Hierarchical Machine Learning for Joint Resource Allocation and Slice Selection in In-Network Edge Systems

2025-11-17

AI Summary

To address the challenges of joint wireless/computing resource management and slice selection in dynamic edge environments under high load—particularly in metaverse applications—this paper proposes a sliced-edge architecture integrating computational intranets with multi-access edge computing (MEC). We design a distributed hierarchical DeepSets-S model featuring relaxation-aware normalization and task-specific decoders to ensure permutation equivariance over variable-length device sets. Our approach combines decomposition of mixed-integer nonlinear programming (MINLP), offline optimal-data-driven training, and COIN/MEC co-scheduling. Experiments demonstrate that subtask accuracy reaches ≥95%, offloading accuracy improves to 88.24%, inference latency decreases by 86.1%, system cost approaches the global optimum (deviation ≤6.1%), and resource utilization significantly outperforms baseline methods.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

The Metaverse promises immersive, real-time experiences; however, meeting its stringent latency and resource demands remains a major challenge. Conventional optimization techniques struggle to respond effectively under dynamic edge conditions and high user loads. In this study, we explore a slice-enabled in-network edge architecture that combines computing-in-the-network (COIN) with multi-access edge computing (MEC). In addition, we formulate the joint problem of wireless and computing resource management with optimal slice selection as a mixed-integer nonlinear program (MINLP). Because solving this model online is computationally intensive, we decompose it into three sub-problems (SP1) intra-slice allocation, (SP2) inter-slice allocation, and (SP3) offloading decision and train a distributed hierarchical DeepSets-based model (DeepSets-S) on optimal solutions obtained offline. In the proposed model, we design a slack-aware normalization mechanism for a shared encoder and task-specific decoders, ensuring permutation equivariance over variable-size wireless device (WD) sets. The learned system produces near-optimal allocations with low inference time and maintains permutation equivariance over variable-size device sets. Our experimental results show that DeepSets-S attains high tolerance-based accuracies on SP1/SP2 (Acc1 = 95.26% and 95.67%) and improves multiclass offloading accuracy on SP3 (Acc = 0.7486; binary local/offload Acc = 0.8824). Compared to exact solvers, the proposed approach reduces the execution time by 86.1%, while closely tracking the optimal system cost (within 6.1% in representative regimes). Compared with baseline models, DeepSets-S consistently achieves higher cost ratios and better utilization across COIN/MEC resources.

Problem

• Optimizing joint wireless and computing resource allocation with slice selection
• Reducing computational complexity of MINLP problems in dynamic edge environments
• Achieving low-latency resource management for Metaverse applications at network edge

Innovation

• Distributed hierarchical DeepSets model for resource allocation
• Decomposed MINLP into three sub-problems for optimization
• Slack-aware normalization with shared encoder and decoders