Linwei Jiang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Linwei Jiang has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Baihong Xie
School of Biomedical Engineering, Sun Yat-Sen University, Shenzhen, China
4 shared publications - Huanhao Zeng3 shared publications
Zhaoyue Wang
3 shared publications- Zhenlin Qin
Institute of Physics, Henan Academy of Sciences, Henan, China
3 shared publications Huanhao Zeng
China Telecom Research Institute, Guangzhou, China
1 shared publicationLika Li
China Telecom Research Institute, Guangzhou, China
1 shared publication
Latest Publications
- Open Access
- 24 August 2026
- Published by Elsevier in SSRN Electronic Journal
Cell Prioritization for Resource-Constrained Green RAN Operations: A Commercial-Network Study
The growing energy and carbon footprint of large-scale radio access network (RAN) has made green operation an increasingly important infrastructure-level problem. Existing research has advanced downstream energy-saving control, yet comparatively less attention has been paid to upstream decisions on where limited operational resources and energy-saving capabilities should be deployed. A key challenge is that energy, traffic, spatial coverage, and technical--operational features are heterogeneous in form and difficult to compare within a unified cell-level decision space. This paper proposes an interpretable, decision-conditioned operational-value profiling and prioritization framework for upstream green RAN decision support. It transforms heterogeneous historical evidence into comparable cell-level value representations and adapts priority formation to decision contexts and resource constraints. The framework is evaluated using two years of real commercial-network data covering 19,796 cells and 403,602 directed coverage relationships. Seven rolling 90-day out-of-time cycles provide large-scale offline validation across changing network conditions. Relative to the operator rule, the proposed method increases the eligible-cohort deployment rate by 10.2 percentage points and cell-averaged energy saving by 29.8\%, while reducing four operational-risk rates by 1.99--3.37 percentage points. Live deployments in ten preselected high-complexity multi-coverage regions further preserve the expected improvement directions in benefit, risk control, and overall operational value. These results indicate that upstream operational-value prioritization can improve resource allocation and role coordination in complex RAN environments and provide actionable guidance for downstream green control.
- 24 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessCell Prioritization for Resource-Constrained Green RAN Operations: A Commercial-Network Study
The growing energy and carbon footprint of large-scale radio access network (RAN) has made green operation an increasingly important infrastructure-level problem. Existing research has advanced downstream energy-saving control, yet comparatively less attention has been paid to upstream decisions on where limited operational resources and energy-saving capabilities should be deployed. A key challenge is that energy, traffic, spatial coverage, and technical--operational features are heterogeneous in form and difficult to compare within a unified cell-level decision space. This paper proposes an interpretable, decision-conditioned operational-value profiling and prioritization framework for upstream green RAN decision support. It transforms heterogeneous historical evidence into comparable cell-level value representations and adapts priority formation to decision contexts and resource constraints. The framework is evaluated using two years of real commercial-network data covering 19,796 cells and 403,602 directed coverage relationships. Seven rolling 90-day out-of-time cycles provide large-scale offline validation across changing network conditions. Relative to the operator rule, the proposed method increases the eligible-cohort deployment rate by 10.2 percentage points and cell-averaged energy saving by 29.8\%, while reducing four operational-risk rates by 1.99--3.37 percentage points. Live deployments in ten preselected high-complexity multi-coverage regions further preserve the expected improvement directions in benefit, risk control, and overall operational value. These results indicate that upstream operational-value prioritization can improve resource allocation and role coordination in complex RAN environments and provide actionable guidance for downstream green control.
- Open Access
- 29 July 2026
- Published by Elsevier in SSRN Electronic Journal
Conflict-Aware Evidence Arbitration for Risk-Constrained Energy Saving in Radio Access Network
Operational decisions in large engineering systems must balance expected benefit, safety risk, feasibility, and accountability under incomplete and conflicting information. This problem is particularly acute in radio access network (RAN) energy-saving operations, where dense 5G deployments, coupled coverage relations, heterogeneous devices, and changing service states make traffic-only decisions unreliable. This paper proposes a Conflict-Aware Evidence Arbitration Network (CAEAN) for risk-constrained RAN energy-saving decision support. The proposed framework organizes tool-grounded domain agents that analyze traffic, coverage, user behavior, hardware execution, historical operation, and feedback from prior operations for the same object--action pair. Their outputs are transformed into role-aware evidence tokens encoding utility, risk, confidence, warning cues, agent identity, and action identity. A set-transformer-inspired encoder then models cross-agent interactions, while an adaptive risk--utility gate fuses opportunity- and risk-oriented representations. Multi-task factor-estimation heads, combined with expert feasibility and safety constraints, produce actionable recommendations with risk guards and explanation traces. We evaluate the framework using real-world RAN energy-saving data, operational agent-platform logs, and observed or matched post-deployment outcomes under an offline replay protocol. Compared with statistical, expert-rule, and learning-based baselines, the proposed method achieves a better benefit--risk trade-off. Relative to the strongest operational-feedback reinforcement-learning method, it improves energy-saving gain by 1.3 percentage points and conflict-resolution accuracy by 4.6 percentage points, while reducing fault-level misjudgment to 0.38%. The results show that learning to arbitrate heterogeneous agent evidence can preserve recoverable energy-saving opportunities while maintaining operational safety.
- 29 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessConflict-Aware Evidence Arbitration for Risk-Constrained Energy Saving in Radio Access Network
Operational decisions in large engineering systems must balance expected benefit, safety risk, feasibility, and accountability under incomplete and conflicting information. This problem is particularly acute in radio access network (RAN) energy-saving operations, where dense 5G deployments, coupled coverage relations, heterogeneous devices, and changing service states make traffic-only decisions unreliable. This paper proposes a Conflict-Aware Evidence Arbitration Network (CAEAN) for risk-constrained RAN energy-saving decision support. The proposed framework organizes tool-grounded domain agents that analyze traffic, coverage, user behavior, hardware execution, historical operation, and feedback from prior operations for the same object--action pair. Their outputs are transformed into role-aware evidence tokens encoding utility, risk, confidence, warning cues, agent identity, and action identity. A set-transformer-inspired encoder then models cross-agent interactions, while an adaptive risk--utility gate fuses opportunity- and risk-oriented representations. Multi-task factor-estimation heads, combined with expert feasibility and safety constraints, produce actionable recommendations with risk guards and explanation traces. We evaluate the framework using real-world RAN energy-saving data, operational agent-platform logs, and observed or matched post-deployment outcomes under an offline replay protocol. Compared with statistical, expert-rule, and learning-based baselines, the proposed method achieves a better benefit--risk trade-off. Relative to the strongest operational-feedback reinforcement-learning method, it improves energy-saving gain by 1.3 percentage points and conflict-resolution accuracy by 4.6 percentage points, while reducing fault-level misjudgment to 0.38%. The results show that learning to arbitrate heterogeneous agent evidence can preserve recoverable energy-saving opportunities while maintaining operational safety.
- Open Access
- 28 July 2026
- Published by Elsevier in SSRN Electronic Journal
Conflict-Aware Evidence Arbitration for Risk-Constrained Energy Saving in Radio Access Network
Operational decisions in large engineering systems must balance expected benefit, safety risk, feasibility, and accountability under incomplete and conflicting information. This problem is particularly acute in radio access network (RAN) energy-saving operations, where dense 5G deployments, coupled coverage relations, heterogeneous devices, and changing service states make traffic-only decisions unreliable. This paper proposes a Conflict-Aware Evidence Arbitration Network (CAEAN) for risk-constrained RAN energy-saving decision support. The proposed framework organizes tool-grounded domain agents that analyze traffic, coverage, user behavior, hardware execution, historical operation, and feedback from prior operations for the same object--action pair. Their outputs are transformed into role-aware evidence tokens encoding utility, risk, confidence, warning cues, agent identity, and action identity. A set-transformer-inspired encoder then models cross-agent interactions, while an adaptive risk--utility gate fuses opportunity- and risk-oriented representations. Multi-task factor-estimation heads, combined with expert feasibility and safety constraints, produce actionable recommendations with risk guards and explanation traces. We evaluate the framework using real-world RAN energy-saving data, operational agent-platform logs, and observed or matched post-deployment outcomes under an offline replay protocol. Compared with statistical, expert-rule, and learning-based baselines, the proposed method achieves a better benefit--risk trade-off. Relative to the strongest operational-feedback reinforcement-learning method, it improves energy-saving gain by 1.3 percentage points and conflict-resolution accuracy by 4.6 percentage points, while reducing fault-level misjudgment to 0.38%. The results show that learning to arbitrate heterogeneous agent evidence can preserve recoverable energy-saving opportunities while maintaining operational safety.
- 28 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessConflict-Aware Evidence Arbitration for Risk-Constrained Energy Saving in Radio Access Network
Operational decisions in large engineering systems must balance expected benefit, safety risk, feasibility, and accountability under incomplete and conflicting information. This problem is particularly acute in radio access network (RAN) energy-saving operations, where dense 5G deployments, coupled coverage relations, heterogeneous devices, and changing service states make traffic-only decisions unreliable. This paper proposes a Conflict-Aware Evidence Arbitration Network (CAEAN) for risk-constrained RAN energy-saving decision support. The proposed framework organizes tool-grounded domain agents that analyze traffic, coverage, user behavior, hardware execution, historical operation, and feedback from prior operations for the same object--action pair. Their outputs are transformed into role-aware evidence tokens encoding utility, risk, confidence, warning cues, agent identity, and action identity. A set-transformer-inspired encoder then models cross-agent interactions, while an adaptive risk--utility gate fuses opportunity- and risk-oriented representations. Multi-task factor-estimation heads, combined with expert feasibility and safety constraints, produce actionable recommendations with risk guards and explanation traces. We evaluate the framework using real-world RAN energy-saving data, operational agent-platform logs, and observed or matched post-deployment outcomes under an offline replay protocol. Compared with statistical, expert-rule, and learning-based baselines, the proposed method achieves a better benefit--risk trade-off. Relative to the strongest operational-feedback reinforcement-learning method, it improves energy-saving gain by 1.3 percentage points and conflict-resolution accuracy by 4.6 percentage points, while reducing fault-level misjudgment to 0.38%. The results show that learning to arbitrate heterogeneous agent evidence can preserve recoverable energy-saving opportunities while maintaining operational safety.
- Open Access
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 32134-32145
Research on Base Station Energy-Saving Optimization Method Integrating Spatial Features and Temporal Modeling
As mobile communication networks continue to scale, energy-saving optimization for base stations is increasingly constrained by the need for fine-grained, timely decision-making under complex and dynamic operating conditions. In practical deployments, the effectiveness and stability of existing energy-saving strategies are primarily limited by two factors: (i) the lack of high-precision and frequently updated base station scene labels that accurately reflect heterogeneous service environments, and (ii) the prohibitive computational cost of modeling complete annual traffic patterns for large-scale networks, which forces most solutions to rely on short-term or sampled data, thereby weakening their ability to capture long-term traffic regularities. The combination of these issues restricts achievable energy-saving depth and undermines operational stability. To address these challenges, this paper proposes a scenario-aware energy-saving optimization framework that integrates semantic scene perception with long-term spatio-temporal traffic modeling. First, fine-grained and dynamically updated scene labels are generated through the fusion of natural language processing on configuration texts and cloud-based geographic map image recognition, enabling accurate characterization of base station service environments without manual annotation. Second, a GeoAutoformer-Transformer model is developed by embedding geographic-aware graph attention into a decompositionbased Transformer architecture, allowing the model to efficiently learn long-term temporal patterns and spatial dependencies among neighboring base stations while remaining computationally tractable for large-scale deployments. In addition, a masking mechanism is introduced to explicitly distinguish between normal zero-traffic intervals caused by energy-saving shutdowns and abnormal zero values resulting from data loss or transmission failures, thereby improving the robustness and reliability of traffic prediction and subsequent energy-saving decisions. Through jointly extending effective energy-saving durations and recommending more suitable energy-saving strategies for eligible base stations, the proposed framework achieves a 17–19% improvement in overall energy-saving efficiency. At the same time, more accurate scene perception and comprehensive traffic modeling contribute to more stable network excellence rates and reduced negative feedback, indicating improved operational reliability. Extensive experiments conducted on a large-scale real-world dataset comprising approximately 50,000 base stations from a nationwide commercial network demonstrate the effectiveness and general applicability of the proposed framework across diverse geographic regions and service scenarios.
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 32134-32145
Open AccessResearch on Base Station Energy-Saving Optimization Method Integrating Spatial Features and Temporal Modeling
As mobile communication networks continue to scale, energy-saving optimization for base stations is increasingly constrained by the need for fine-grained, timely decision-making under complex and dynamic operating conditions. In practical deployments, the effectiveness and stability of existing energy-saving strategies are primarily limited by two factors: (i) the lack of high-precision and frequently updated base station scene labels that accurately reflect heterogeneous service environments, and (ii) the prohibitive computational cost of modeling complete annual traffic patterns for large-scale networks, which forces most solutions to rely on short-term or sampled data, thereby weakening their ability to capture long-term traffic regularities. The combination of these issues restricts achievable energy-saving depth and undermines operational stability. To address these challenges, this paper proposes a scenario-aware energy-saving optimization framework that integrates semantic scene perception with long-term spatio-temporal traffic modeling. First, fine-grained and dynamically updated scene labels are generated through the fusion of natural language processing on configuration texts and cloud-based geographic map image recognition, enabling accurate characterization of base station service environments without manual annotation. Second, a GeoAutoformer-Transformer model is developed by embedding geographic-aware graph attention into a decompositionbased Transformer architecture, allowing the model to efficiently learn long-term temporal patterns and spatial dependencies among neighboring base stations while remaining computationally tractable for large-scale deployments. In addition, a masking mechanism is introduced to explicitly distinguish between normal zero-traffic intervals caused by energy-saving shutdowns and abnormal zero values resulting from data loss or transmission failures, thereby improving the robustness and reliability of traffic prediction and subsequent energy-saving decisions. Through jointly extending effective energy-saving durations and recommending more suitable energy-saving strategies for eligible base stations, the proposed framework achieves a 17–19% improvement in overall energy-saving efficiency. At the same time, more accurate scene perception and comprehensive traffic modeling contribute to more stable network excellence rates and reduced negative feedback, indicating improved operational reliability. Extensive experiments conducted on a large-scale real-world dataset comprising approximately 50,000 base stations from a nationwide commercial network demonstrate the effectiveness and general applicability of the proposed framework across diverse geographic regions and service scenarios.