Chixin Xiao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Chixin Xiao has a total of 49 co-authors. Here are the top co-authors by number of shared publications:
- Maoxin He
Xiangtan University, Xiangtan, China
6 shared publications - Minjie Zhang
University of Wollongong, Wollongong, NSW, Australia
6 shared publications - Dechen Jiang
Xiangtan University, Xiangtan, China
5 shared publications Danny Sutanto
3 shared publicationsKashem M. Muttaqi
3 shared publications- Zhao Yang Dong
City University of Hong Kong
3 shared publications
Latest Publications
- Open Access
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1102-1110
Anisotropic Random-Feature Surrogate-Assisted Evolutionary Algorithm for High-Dimensional Risk-Based Energy Management Optimization
In modern power systems, the day-ahead Energy Resource Management (ERM) problem, particularly under high penetration of Distributed Energy Resources (DERs), often becomes a high-dimensional challenge due to the uncertainties inherent in renewable generation, load consumption, and market prices. The complex, nested relationships between input control vectors and system outputs effectively transform these problems into black-box models, where only inputs and outputs are observable. This poses significant challenges for optimization algorithms, especially in terms of computational efficiency, under conditions of costly function evaluations and a limited number of evaluations. To address these challenges, this study proposes an anisotropic random-feature Extreme Learning Machine (ELM) surrogate integrated with evolutionary algorithms, designed for high-dimensional optimization problems under uncertainty. By incorporating variable-level scaling and anisotropic transformations, the proposed surrogate dynamically adapts to the heterogeneous correlations across different dimensions, thereby enhancing optimization efficiency while maintaining the rapid training and generalization capabilities of ELM. Experiments on the GECCO competition platform and widely used large-scale benchmark suites show that the proposed method achieves superior solution quality compared with state-of-the-art evolutionary algorithms.
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1102-1110
Open AccessAnisotropic Random-Feature Surrogate-Assisted Evolutionary Algorithm for High-Dimensional Risk-Based Energy Management Optimization
In modern power systems, the day-ahead Energy Resource Management (ERM) problem, particularly under high penetration of Distributed Energy Resources (DERs), often becomes a high-dimensional challenge due to the uncertainties inherent in renewable generation, load consumption, and market prices. The complex, nested relationships between input control vectors and system outputs effectively transform these problems into black-box models, where only inputs and outputs are observable. This poses significant challenges for optimization algorithms, especially in terms of computational efficiency, under conditions of costly function evaluations and a limited number of evaluations. To address these challenges, this study proposes an anisotropic random-feature Extreme Learning Machine (ELM) surrogate integrated with evolutionary algorithms, designed for high-dimensional optimization problems under uncertainty. By incorporating variable-level scaling and anisotropic transformations, the proposed surrogate dynamically adapts to the heterogeneous correlations across different dimensions, thereby enhancing optimization efficiency while maintaining the rapid training and generalization capabilities of ELM. Experiments on the GECCO competition platform and widely used large-scale benchmark suites show that the proposed method achieves superior solution quality compared with state-of-the-art evolutionary algorithms.
- Open Access
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1129-1136
Dual-Layer Synchronous Ring Cellular UMDA Based On Anisotropic RBF Surrogate for High-Dimensional Evolutionary Optimization in Energy Management
Day-ahead energy resource management faces uncertainty in renewables, demand, prices, and EV operations, yet existing methods often struggle with high dimensionality and expensive evaluations. This paper develops a Dual-layer Synchronous optimization framework that integrates a Ring Cellular Encode-Decode UMDA with an anisotropic radial basis function surrogate (DLRC-ARBF). The lower layer encodes decision variables and samples from univariate marginal distributions, which reduces the effective selection in each dimension while maintaining population diversity. In parallel, the upper layer learns dimension-specific length scales with an ARBF model to capture axis-dependent variability and to generate directionally scaled candidate solutions around elites. Both layers synchronously propose offspring each generation, and their joint subpopulations are evaluated on the true objective. The results show that DLRC-ARBF surpasses all other participants on the IEEE CEC/GECCO 2025 competition benchmark under the same evaluation budget, achieving the best overall performance.
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1129-1136
Open AccessDual-Layer Synchronous Ring Cellular UMDA Based On Anisotropic RBF Surrogate for High-Dimensional Evolutionary Optimization in Energy Management
Day-ahead energy resource management faces uncertainty in renewables, demand, prices, and EV operations, yet existing methods often struggle with high dimensionality and expensive evaluations. This paper develops a Dual-layer Synchronous optimization framework that integrates a Ring Cellular Encode-Decode UMDA with an anisotropic radial basis function surrogate (DLRC-ARBF). The lower layer encodes decision variables and samples from univariate marginal distributions, which reduces the effective selection in each dimension while maintaining population diversity. In parallel, the upper layer learns dimension-specific length scales with an ARBF model to capture axis-dependent variability and to generate directionally scaled candidate solutions around elites. Both layers synchronously propose offspring each generation, and their joint subpopulations are evaluated on the true objective. The results show that DLRC-ARBF surpasses all other participants on the IEEE CEC/GECCO 2025 competition benchmark under the same evaluation budget, achieving the best overall performance.
- 1 October 2025
- Published by Allerton Press in Applied Solar Energy
- Vol. 61 (5) , 563-575
Solar Energy Interval Prediction Based on Online Learning and Kernel Density Estimation
- 1 October 2025
- Published by Allerton Press in Applied Solar Energy
- Vol. 61 (5) , 563-575
Solar Energy Interval Prediction Based on Online Learning and Kernel Density Estimation
- Open Access
- 14 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 959-962
Hybrid-Surrogate-Assisted Particle Swarm Optimization with Adaptive Surrogate Selection for Solving High-Dimensional Expensive Optimization Problems
- 14 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 959-962
Open AccessHybrid-Surrogate-Assisted Particle Swarm Optimization with Adaptive Surrogate Selection for Solving High-Dimensional Expensive Optimization Problems
- Open Access
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1399-1407
Differential Evolution Based on Light-Weight-Surrogate for Solving High-Dimensional Energy Management Problem
In modern power engineering, given the high penetration of distributed energy resources, a risk-based and day-ahead management problem can often be formulated as a high-dimensional problem. In addition, the mathematical mapping from the input control vector to the quantiles of the system output becomes too complex due to the nested relationship between the different systems. Hence, this type of problem is often approached as a black box model, i.e., the decision maker can only observe the input and output ends. However, as a practical engineering problem, it often requires the corresponding optimization in limited time or evaluations, which poses a challenge to conventional differential evolution algorithms. This paper proposes a novel differential evolution algorithm with a compatible surrogate operator that provides dynamic rational exploration during the iterations to achieve reliable performance in limited time and space complexity. The surrogate also facilitates faster exploration of the optimal solution region compared with the state-of-the-art under the same configurations and conditions. A GECCO competition testbed is utilized to evaluate the performance, where the simulation results demonstrate that the proposed algorithm is more effective than most DE, PSO (and their typical variants).
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1399-1407
Open AccessDifferential Evolution Based on Light-Weight-Surrogate for Solving High-Dimensional Energy Management Problem
In modern power engineering, given the high penetration of distributed energy resources, a risk-based and day-ahead management problem can often be formulated as a high-dimensional problem. In addition, the mathematical mapping from the input control vector to the quantiles of the system output becomes too complex due to the nested relationship between the different systems. Hence, this type of problem is often approached as a black box model, i.e., the decision maker can only observe the input and output ends. However, as a practical engineering problem, it often requires the corresponding optimization in limited time or evaluations, which poses a challenge to conventional differential evolution algorithms. This paper proposes a novel differential evolution algorithm with a compatible surrogate operator that provides dynamic rational exploration during the iterations to achieve reliable performance in limited time and space complexity. The surrogate also facilitates faster exploration of the optimal solution region compared with the state-of-the-art under the same configurations and conditions. A GECCO competition testbed is utilized to evaluate the performance, where the simulation results demonstrate that the proposed algorithm is more effective than most DE, PSO (and their typical variants).
- 19 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A Hybrid of Multidisciplinary Design and Multi-Objective Optimization for Solving Combined Heat and Power Environmental Economic Dispatch
This paper proposes an approach taking advantage of multidisciplinary design optimization (MDO) and multi-objective optimization (MOO) to solve the combined heat and power economic/environmental dispatch (CHPEED) problem. The challenge in solving the CHPEED problem, coupling the electrical and thermal disciplines, needs to optimize the operational cost and the emission degree with more complicated constraints and feasible decision space than the conventional multi-objective optimization models in a pure power discipline. The proposed approach explores the CHPEED problem, considering the separated optimal requirements in the power and thermal disciplines and taking advantage of the interactions between such two fields. The simulations analysis shows that the proposed approach gets better performance by adopting MDO strategies to transfer the CHPEED problem into a series of standard optimization problems, and which can be solved through a set of optimizers in parallel.
- 19 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A Hybrid of Multidisciplinary Design and Multi-Objective Optimization for Solving Combined Heat and Power Environmental Economic Dispatch
This paper proposes an approach taking advantage of multidisciplinary design optimization (MDO) and multi-objective optimization (MOO) to solve the combined heat and power economic/environmental dispatch (CHPEED) problem. The challenge in solving the CHPEED problem, coupling the electrical and thermal disciplines, needs to optimize the operational cost and the emission degree with more complicated constraints and feasible decision space than the conventional multi-objective optimization models in a pure power discipline. The proposed approach explores the CHPEED problem, considering the separated optimal requirements in the power and thermal disciplines and taking advantage of the interactions between such two fields. The simulations analysis shows that the proposed approach gets better performance by adopting MDO strategies to transfer the CHPEED problem into a series of standard optimization problems, and which can be solved through a set of optimizers in parallel.
- 12 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 57 (2) , 1860-1871
Online Sequential Extreme Learning Machine Algorithm for Better Predispatch Electricity Price Forecasting Grids
The predispatch price forecast plays a key element in the electricity market. However, such a forecast usually depends on the traditional offline batch-learning technologies, which cannot respond in time to the unexpected changes in the local power system environment. Further, the predispatch local price forecast is often affected by the dynamic price changes from the neighboring regions. This article proposes a novel online learning forecast approach to overcome the above issues to provide a better predispatch price forecast by using the online sequential extreme learning machine (OS-ELM) algorithm. The article proposes a novel data structure in the form of a 2-D orthogonal list and two corresponding OS-ELM modules. One module provides the rolling day-ahead price prediction and prediction intervals using the day-by-day online training update, while the other provides the rolling 30-min prediction using the 2-h-by-2-h online training update. The proposed approach can continuously perceive any unexpected events and any price fluctuations from the neighboring regions in the nonlinear patterns. The proposed approach is validated using simulation studies based on the data from the Australian electricity market, and the simulation results show that the proposed approach can help in improving the forecast accuracy, especially when unexpected changes occur both locally and in the neighboring area.
- 12 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 57 (2) , 1860-1871
Online Sequential Extreme Learning Machine Algorithm for Better Predispatch Electricity Price Forecasting Grids
The predispatch price forecast plays a key element in the electricity market. However, such a forecast usually depends on the traditional offline batch-learning technologies, which cannot respond in time to the unexpected changes in the local power system environment. Further, the predispatch local price forecast is often affected by the dynamic price changes from the neighboring regions. This article proposes a novel online learning forecast approach to overcome the above issues to provide a better predispatch price forecast by using the online sequential extreme learning machine (OS-ELM) algorithm. The article proposes a novel data structure in the form of a 2-D orthogonal list and two corresponding OS-ELM modules. One module provides the rolling day-ahead price prediction and prediction intervals using the day-by-day online training update, while the other provides the rolling 30-min prediction using the 2-h-by-2-h online training update. The proposed approach can continuously perceive any unexpected events and any price fluctuations from the neighboring regions in the nonlinear patterns. The proposed approach is validated using simulation studies based on the data from the Australian electricity market, and the simulation results show that the proposed approach can help in improving the forecast accuracy, especially when unexpected changes occur both locally and in the neighboring area.