Yao Jiang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yao Jiang has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Yuanzhao Deng
Faculty of Data Science, City University of Macau, Macau SAR, China
6 shared publications - Yan Ai
City University of Macau
3 shared publications - Jixin Ma
The School of Computing and Mathematical Sciences University of Greenwich London UK
2 shared publications - Yuesheng Zhu
Peking University
2 shared publications - Fang Fang Yan
College of Finance and Economics, Guangdong University of Science and Technology, Dongguan, Guangdong, China
1 shared publication - Guoqiang Luo
Faculty of Artificial Intelligence and Big Data, Guangzhou Vocational and Technical University of Science and Technology, Guangzhou 510555, China
1 shared publication
Latest Publications
- Open Access
XLTFormer: Regime-Adaptive Time–Frequency Expert Routing for Multi-Step Load and Energy Forecasting
Multi-step load forecasting remains challenging because real-world demand series exhibit non-stationarity, multi-scale periodicity, and regime-dependent temporal structure. Performance degradation often arises not from insufficient model size alone, but from mismatch between the active operating regime and the inductive bias of a fixed forecasting backbone. This paper proposes XLTFormer, a regime-adaptive framework that treats multi-step prediction as descriptor-augmented routing over heterogeneous structural priors. Memory, attention, decomposition, cross-variable interaction, patch-based representation, and frequency-domain modeling are organized as route-level experts, while the methodological contribution centers on four mechanisms: regime-descriptor-guided sample-level routing, post-mixture peak-context correction, lightweight expert-preserving training, and validation-gated deployment selection. Across three real-load benchmarks, XLTFormer attains the lowest mean error on all five reported metrics under the ten-seed protocol and remains top-ranked overall under the controlled-capacity comparison. In the retained rerun, the final deployment repeatedly selects the cross-variable route on all three datasets, with WAPE/sMAPE values of 3.7274/3.7320 on Electricity, 3.9576/3.8643 on PJM, and 7.2619/7.5058 on GEFCom2014 Load. Relative to the strongest competing baseline, the WAPE improvement reaches 2.11% on Electricity, 1.11% on PJM, and 1.02% on GEFCom2014 Load, while the corresponding sMAPE reduction reaches 2.40%, 1.29%, and 1.08%, respectively. Seed-level paired tests, dependence-aware block bootstrap, and Friedman ranking analysis further show that this retained deployment advantage is stable rather than incidental. The temporal-routing audit nevertheless indicates that the sample-level gate still redistributes route responsibility across local regimes, especially when peak, ramp, or weather-linked stress increases. These results show that the practical forecasting gain comes from the full reliability-oriented framework, in which adaptive routing shapes route interaction and validation-gated selection determines the final retained output, rather than from attributing the deployed accuracy to raw adaptive fusion alone. Overall, the framework provides an effective direction for accurate, stable, and practically relevant multi-step load forecasting.
Open AccessXLTFormer: Regime-Adaptive Time–Frequency Expert Routing for Multi-Step Load and Energy Forecasting
Multi-step load forecasting remains challenging because real-world demand series exhibit non-stationarity, multi-scale periodicity, and regime-dependent temporal structure. Performance degradation often arises not from insufficient model size alone, but from mismatch between the active operating regime and the inductive bias of a fixed forecasting backbone. This paper proposes XLTFormer, a regime-adaptive framework that treats multi-step prediction as descriptor-augmented routing over heterogeneous structural priors. Memory, attention, decomposition, cross-variable interaction, patch-based representation, and frequency-domain modeling are organized as route-level experts, while the methodological contribution centers on four mechanisms: regime-descriptor-guided sample-level routing, post-mixture peak-context correction, lightweight expert-preserving training, and validation-gated deployment selection. Across three real-load benchmarks, XLTFormer attains the lowest mean error on all five reported metrics under the ten-seed protocol and remains top-ranked overall under the controlled-capacity comparison. In the retained rerun, the final deployment repeatedly selects the cross-variable route on all three datasets, with WAPE/sMAPE values of 3.7274/3.7320 on Electricity, 3.9576/3.8643 on PJM, and 7.2619/7.5058 on GEFCom2014 Load. Relative to the strongest competing baseline, the WAPE improvement reaches 2.11% on Electricity, 1.11% on PJM, and 1.02% on GEFCom2014 Load, while the corresponding sMAPE reduction reaches 2.40%, 1.29%, and 1.08%, respectively. Seed-level paired tests, dependence-aware block bootstrap, and Friedman ranking analysis further show that this retained deployment advantage is stable rather than incidental. The temporal-routing audit nevertheless indicates that the sample-level gate still redistributes route responsibility across local regimes, especially when peak, ramp, or weather-linked stress increases. These results show that the practical forecasting gain comes from the full reliability-oriented framework, in which adaptive routing shapes route interaction and validation-gated selection determines the final retained output, rather than from attributing the deployed accuracy to raw adaptive fusion alone. Overall, the framework provides an effective direction for accurate, stable, and practically relevant multi-step load forecasting.
- Open Access
- 29 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 137-145
A Cross-Modal Multi-Task Transformer Model for Small-Sample Data Pricing
- 29 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 137-145
Open AccessA Cross-Modal Multi-Task Transformer Model for Small-Sample Data Pricing
- Open Access
The State of Health Prediction of Li-Ion Batteries Based on ISMA-HKELM
Accurately predicting the State of Health (SOH) of lithium-ion batteries (LIBs) is essential to ensure their long-term stable and safe operation. This paper proposes a novel model, the ISMA-HKELM, which is an Improved Slime Mould Algorithm (ISMA)-optimized Hybrid Kernel Extreme Learning Machine (HKELM), designed for high-precision SOH estimation. We first selected the equal voltage rise time and equal voltage drop time as indirect health indicators, and their validity was rigorously confirmed through Pearson and Spearman correlation tests. Subsequently, the ISMA was utilized to effectively tune the key parameters of the HKELM model. Experimental results demonstrate that the ISMA-HKELM model exhibits superior prediction performance across multiple public datasets, achieving an average value of more than 0.99. Furthermore, the model shows significantly lower Mean Absolute Error (MAE), Mean Bias Error (MBE), and Root Mean Square Error (RMSE) compared to other control models. These results fully prove the advancement and validity of the ISMA-HKELM model for LIB SOH estimation.Open AccessThe State of Health Prediction of Li-Ion Batteries Based on ISMA-HKELM
Accurately predicting the State of Health (SOH) of lithium-ion batteries (LIBs) is essential to ensure their long-term stable and safe operation. This paper proposes a novel model, the ISMA-HKELM, which is an Improved Slime Mould Algorithm (ISMA)-optimized Hybrid Kernel Extreme Learning Machine (HKELM), designed for high-precision SOH estimation. We first selected the equal voltage rise time and equal voltage drop time as indirect health indicators, and their validity was rigorously confirmed through Pearson and Spearman correlation tests. Subsequently, the ISMA was utilized to effectively tune the key parameters of the HKELM model. Experimental results demonstrate that the ISMA-HKELM model exhibits superior prediction performance across multiple public datasets, achieving an average value of more than 0.99. Furthermore, the model shows significantly lower Mean Absolute Error (MAE), Mean Bias Error (MBE), and Root Mean Square Error (RMSE) compared to other control models. These results fully prove the advancement and validity of the ISMA-HKELM model for LIB SOH estimation. - Open Access
MIGS: A Modular Edge Gateway with Instance-Based Isolation for Heterogeneous Industrial IoT Interoperability
The exponential proliferation of the Internet of Things (IoT) has catalyzed a paradigm shift in industrial automation and smart city infrastructure. However, this rapid expansion has engendered significant heterogeneity in communication protocols, creating critical barriers to seamless data integration and interoperability. Conventional gateway solutions frequently exhibit limited flexibility in supporting diverse protocol stacks simultaneously and often lack granular user controllability. To mitigate these deficiencies, this paper proposes a novel, modular IoT gateway architecture, designated as MIGS (Modular IoT Gateway System). The proposed architecture comprises four distinct components: a Management Component, a Southbound Component, a Northbound Component, and a Cache Component. Specifically, the Southbound Component employs instance-based isolation and independent task threading to manage heterogeneous field devices utilizing protocols such as Modbus, MQTT, and OPC UA. The Northbound Component facilitates reliable bidirectional data transmission with cloud platforms. A dedicated Cache Component is integrated to decouple data acquisition from transmission, ensuring data integrity during network latency. Furthermore, a web-based Control Service Module affords comprehensive runtime management. We explicate the data transmission methodology and formulate a theoretical latency model to quantify the impact of the Python Global Interpreter Lock (GIL) and serialization overhead. Functional validation and theoretical analysis confirm the system’s efficacy in concurrent multi-protocol communication, robust data forwarding, and operational flexibility. The MIGS framework significantly enhances interoperability within heterogeneous IoT environments, offering a scalable solution for next-generation industrial applications.
Open AccessMIGS: A Modular Edge Gateway with Instance-Based Isolation for Heterogeneous Industrial IoT Interoperability
The exponential proliferation of the Internet of Things (IoT) has catalyzed a paradigm shift in industrial automation and smart city infrastructure. However, this rapid expansion has engendered significant heterogeneity in communication protocols, creating critical barriers to seamless data integration and interoperability. Conventional gateway solutions frequently exhibit limited flexibility in supporting diverse protocol stacks simultaneously and often lack granular user controllability. To mitigate these deficiencies, this paper proposes a novel, modular IoT gateway architecture, designated as MIGS (Modular IoT Gateway System). The proposed architecture comprises four distinct components: a Management Component, a Southbound Component, a Northbound Component, and a Cache Component. Specifically, the Southbound Component employs instance-based isolation and independent task threading to manage heterogeneous field devices utilizing protocols such as Modbus, MQTT, and OPC UA. The Northbound Component facilitates reliable bidirectional data transmission with cloud platforms. A dedicated Cache Component is integrated to decouple data acquisition from transmission, ensuring data integrity during network latency. Furthermore, a web-based Control Service Module affords comprehensive runtime management. We explicate the data transmission methodology and formulate a theoretical latency model to quantify the impact of the Python Global Interpreter Lock (GIL) and serialization overhead. Functional validation and theoretical analysis confirm the system’s efficacy in concurrent multi-protocol communication, robust data forwarding, and operational flexibility. The MIGS framework significantly enhances interoperability within heterogeneous IoT environments, offering a scalable solution for next-generation industrial applications.
- 27 November 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (11) , 115021
Euler optimizer: a novel metaheuristic algorithm for 3D UAV path planning and engineering design problems
Three-dimensional (3D) path planning for unmanned aerial vehicles (UAVs) and real-world engineering design problems has been a hot research issue. Traditional methods show limitations in dealing with these complex nonlinear models. To overcome these challenges, this paper proposes a new metaheuristic method, Euler optimizer (ELO), inspired by Euler’s method for the numerical solution of differential equations.ELO simulates Euler’s method to predict the position of the next point by the derivative and step size of the current point to perform the position updating, and sets the exploration weighting factor and the exploration factors that are used to tune the exploration and exploitation capabilities of the algorithm. We compared ELO with 11 representative state-of-the-art algorithms using the CEC2017 and CEC2022 benchmark suites, and performed the Wilcoxon rank sum test and Friedman’s test. The results show that ELO outperforms the other comparative algorithms by 90%, 96%, 99%, and 84% on CEC2017 (30/50/100 dimensions) and CEC2022 (20 dimensions), respectively. Finally, ELO is applied to UAV 3D path planning and five real-world engineering design problems. The experimental results show that ELO achieves the best compared to all 11 comparison algorithms, demonstrating ELO’s effectiveness and extensiveness in engineering optimization problems.
- 27 November 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (11) , 115021
Euler optimizer: a novel metaheuristic algorithm for 3D UAV path planning and engineering design problems
Three-dimensional (3D) path planning for unmanned aerial vehicles (UAVs) and real-world engineering design problems has been a hot research issue. Traditional methods show limitations in dealing with these complex nonlinear models. To overcome these challenges, this paper proposes a new metaheuristic method, Euler optimizer (ELO), inspired by Euler’s method for the numerical solution of differential equations.ELO simulates Euler’s method to predict the position of the next point by the derivative and step size of the current point to perform the position updating, and sets the exploration weighting factor and the exploration factors that are used to tune the exploration and exploitation capabilities of the algorithm. We compared ELO with 11 representative state-of-the-art algorithms using the CEC2017 and CEC2022 benchmark suites, and performed the Wilcoxon rank sum test and Friedman’s test. The results show that ELO outperforms the other comparative algorithms by 90%, 96%, 99%, and 84% on CEC2017 (30/50/100 dimensions) and CEC2022 (20 dimensions), respectively. Finally, ELO is applied to UAV 3D path planning and five real-world engineering design problems. The experimental results show that ELO achieves the best compared to all 11 comparison algorithms, demonstrating ELO’s effectiveness and extensiveness in engineering optimization problems.
- 15 September 2025
- Published by Springer Nature in Cluster Computing
- Vol. 28 (12) , 1-54
Adams-Bashforth-Moulton optimizer: a novel metaheuristic algorithm for solving engineering optimization problems
- 15 September 2025
- Published by Springer Nature in Cluster Computing
- Vol. 28 (12) , 1-54
Adams-Bashforth-Moulton optimizer: a novel metaheuristic algorithm for solving engineering optimization problems