Zhenshou Song
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
Research Subjects
Zhenshou Song has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Chunquan Li
School of Information Engineering, Nanchang University, Nanchang, China
6 shared publications - Handing Wang
Xidian University, Xi'an, China
5 shared publications - Peter Xiaoping Liu
Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada
5 shared publications Jinghui Fan
School of Information Engineering, Nanchang University, Nanchang 330029, CHINA
4 shared publications- Yaochu Jin
Westlake University, Hangzhou, CHINA
3 shared publications - Bing Xue
Centre for Data Science and Artificial Intelligence & School of Engineering and Computer Science, Victoria University of Wellington, New Zealand
2 shared publications
Latest Publications
- 20 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 28 (6) , 1824-1837
Balancing Different Optimization Difficulty Between Objectives in Multiobjective Feature Selection
Multi-objective feature selection aims to find a set of feature subsets that achieves a trade-off between two objectives, i.e., reducing the number of selected features and improving the classification performance. However, these two objectives might not be always conflicting during the optimization process and have varying difficulties in optimization. Such characteristics pose a great challenge to existing multi-objective evolutionary approaches, which often treat two objectives equally. Specifically, a large number of feature subsets with few features may appear in the population and compete for survival opportunities with promising feature subsets located in not fully explored regions, leading to poor performance. To this end, we propose a two-archive evolutionary feature selection algorithm for multi-objective feature selection. In the proposed method, all individuals are equally allocated into two independent archives. A two-archive based solution generation strategy is proposed, where a dynamic dimensionality reduction operator is used to exploit small features subsets while a diversity-based mutation operator is utilized to find feature subsets with better classification performance. Moreover, a novel environmental selection scheme is proposed, which aims to improve the survival probability of promising feature subsets by providing different selection environments. Experimental results on 23 datasets demonstrate that the proposed algorithm is superior to the other five state-of-the-art algorithms.
- 20 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 28 (6) , 1824-1837
Balancing Different Optimization Difficulty Between Objectives in Multiobjective Feature Selection
Multi-objective feature selection aims to find a set of feature subsets that achieves a trade-off between two objectives, i.e., reducing the number of selected features and improving the classification performance. However, these two objectives might not be always conflicting during the optimization process and have varying difficulties in optimization. Such characteristics pose a great challenge to existing multi-objective evolutionary approaches, which often treat two objectives equally. Specifically, a large number of feature subsets with few features may appear in the population and compete for survival opportunities with promising feature subsets located in not fully explored regions, leading to poor performance. To this end, we propose a two-archive evolutionary feature selection algorithm for multi-objective feature selection. In the proposed method, all individuals are equally allocated into two independent archives. A two-archive based solution generation strategy is proposed, where a dynamic dimensionality reduction operator is used to exploit small features subsets while a diversity-based mutation operator is utilized to find feature subsets with better classification performance. Moreover, a novel environmental selection scheme is proposed, which aims to improve the survival probability of promising feature subsets by providing different selection environments. Experimental results on 23 datasets demonstrate that the proposed algorithm is superior to the other five state-of-the-art algorithms.
- 31 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 28 (5) , 1286-1300
Balancing Objective Optimization and Constraint Satisfaction in Expensive Constrained Evolutionary Multiobjective Optimization
In dealing with expensive constrained multi-objective optimization problems using surrogate-assisted evolutionary algorithms, it is a great challenge to reduce the negative impact caused by the approximate errors of surrogate models for constraints. To address this issue, we propose a Kriging-assisted evolutionary algorithm with two search modes to adaptively reduce the utilization frequency of surrogate models for constraints. To be more specific, an adaptively switching strategy analyzing the correlation between the objective optimization direction and constraint satisfaction direction is designed to determine whether to build the constraint surrogate models to assist the current evolutionary search. Accordingly, the proposed algorithm contains two search modes: 1) unconstrained surrogate-assisted search mode and 2) constrained surrogate-assisted search mode. In the first search mode, an existing surrogate-assisted evolutionary algorithm without considering constraint is introduced, which rapidly drives the population to move to the feasible region(s) while avoiding the negative effects of the constraint surrogate models. In the second search mode, a novel Kriging-assisted constrained multi-objective optimization algorithm is designed for locating constrained Pareto front in the feasible region. In addition, a data selection strategy is proposed to improve the efficiency and quality of surrogate models for constraint functions. The proposed method has been tested on numerous instances from three popular benchmark test suites. The experimental results demonstrate that the performance of the proposed algorithm outperforms other state-of-the-art methods.
- 31 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 28 (5) , 1286-1300
Balancing Objective Optimization and Constraint Satisfaction in Expensive Constrained Evolutionary Multiobjective Optimization
In dealing with expensive constrained multi-objective optimization problems using surrogate-assisted evolutionary algorithms, it is a great challenge to reduce the negative impact caused by the approximate errors of surrogate models for constraints. To address this issue, we propose a Kriging-assisted evolutionary algorithm with two search modes to adaptively reduce the utilization frequency of surrogate models for constraints. To be more specific, an adaptively switching strategy analyzing the correlation between the objective optimization direction and constraint satisfaction direction is designed to determine whether to build the constraint surrogate models to assist the current evolutionary search. Accordingly, the proposed algorithm contains two search modes: 1) unconstrained surrogate-assisted search mode and 2) constrained surrogate-assisted search mode. In the first search mode, an existing surrogate-assisted evolutionary algorithm without considering constraint is introduced, which rapidly drives the population to move to the feasible region(s) while avoiding the negative effects of the constraint surrogate models. In the second search mode, a novel Kriging-assisted constrained multi-objective optimization algorithm is designed for locating constrained Pareto front in the feasible region. In addition, a data selection strategy is proposed to improve the efficiency and quality of surrogate models for constraint functions. The proposed method has been tested on numerous instances from three popular benchmark test suites. The experimental results demonstrate that the performance of the proposed algorithm outperforms other state-of-the-art methods.
- 14 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 53 (10) , 6268-6280
A Surrogate-Assisted Evolutionary Framework With Regions of Interests-Based Data Selection for Expensive Constrained Optimization
Optimization problems whose evaluations of the objective and constraints involve costly numerical simulations or physical experiments are referred to as expensive constrained optimization (ECO) problems. Such problems can be solved by evolutionary algorithms (EAs) in conjunction with computationally cheap surrogates that separately approximate the expensive objective and constraint functions. During the process of the ECO, the interested regions of surrogate models for the objective and constraints usually have a small overlap only. Specifically, the surrogate model for the objective function should focus on the prediction accuracy in the promising region, while the models for constraint functions should concentrate on the accuracy at the boundary of the feasible region. However, most existing methods neglect such differences and train those different models using the same training data, barely resulting in satisfactory performance. Therefore, we propose a general framework for solving expensive optimization problems with inequality constraints. In the proposed framework, the objective and constraints are separately trained with two different sets of training data to enhance the prediction accuracy and reliability in the interested regions. A novel infill sampling criterion is tailored to decide whether potentially better or more uncertain solutions should be sampled. Moreover, a new strategy, termed search intensity adjustment, is designed for adjusting the number of search generations on new surrogate models. We attempt to embed three competitive constrained EAs into our framework to verify its generality. The experimental results obtained on numerous benchmark functions from CEC2006, CEC2010, and CEC2017 have demonstrated the superiority of our approach over three state-of-the-art surrogate-assisted EAs.
- 14 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 53 (10) , 6268-6280
A Surrogate-Assisted Evolutionary Framework With Regions of Interests-Based Data Selection for Expensive Constrained Optimization
Optimization problems whose evaluations of the objective and constraints involve costly numerical simulations or physical experiments are referred to as expensive constrained optimization (ECO) problems. Such problems can be solved by evolutionary algorithms (EAs) in conjunction with computationally cheap surrogates that separately approximate the expensive objective and constraint functions. During the process of the ECO, the interested regions of surrogate models for the objective and constraints usually have a small overlap only. Specifically, the surrogate model for the objective function should focus on the prediction accuracy in the promising region, while the models for constraint functions should concentrate on the accuracy at the boundary of the feasible region. However, most existing methods neglect such differences and train those different models using the same training data, barely resulting in satisfactory performance. Therefore, we propose a general framework for solving expensive optimization problems with inequality constraints. In the proposed framework, the objective and constraints are separately trained with two different sets of training data to enhance the prediction accuracy and reliability in the interested regions. A novel infill sampling criterion is tailored to decide whether potentially better or more uncertain solutions should be sampled. Moreover, a new strategy, termed search intensity adjustment, is designed for adjusting the number of search generations on new surrogate models. We attempt to embed three competitive constrained EAs into our framework to verify its generality. The experimental results obtained on numerous benchmark functions from CEC2006, CEC2010, and CEC2017 have demonstrated the superiority of our approach over three state-of-the-art surrogate-assisted EAs.
- Top Cited Papers
- 16 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 25 (6) , 1013-1027
A Kriging-Assisted Two-Archive Evolutionary Algorithm for Expensive Many-Objective Optimization
Only a small number of function evaluations can be afforded in many real-world multiobjective optimization problems (MOPs) where the function evaluations are economically/computationally expensive. Such problems pose great challenges to most existing multiobjective evolutionary algorithms (EAs), which require a large number of function evaluations for optimization. Surrogate-assisted EAs (SAEAs) have been employed to solve expensive MOPs. Specifically, a certain number of expensive function evaluations are used to build computationally cheap surrogate models for assisting the optimization process without conducting expensive function evaluations. The infill sampling criteria in most existing SAEAs take all requirements on convergence, diversity, and model uncertainty into account, which is, however, not the most efficient in exploiting the limited computational budget. Thus, this article proposes a Kriging-assisted two-archive EA for expensive many-objective optimization. The proposed algorithm uses one influential point-insensitive model to approximate each objective function. Moreover, an adaptive infill criterion that identifies the most important requirement on convergence, diversity, or uncertainty is proposed to determine an appropriate sampling strategy for reevaluations using the expensive objective functions. The experimental results on a set of expensive multi/many-objective test problems have demonstrated its superiority over five state-of-the-art SAEAs.
- 16 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Evolutionary Computation
- Vol. 25 (6) , 1013-1027
Top Cited PapersA Kriging-Assisted Two-Archive Evolutionary Algorithm for Expensive Many-Objective Optimization
Only a small number of function evaluations can be afforded in many real-world multiobjective optimization problems (MOPs) where the function evaluations are economically/computationally expensive. Such problems pose great challenges to most existing multiobjective evolutionary algorithms (EAs), which require a large number of function evaluations for optimization. Surrogate-assisted EAs (SAEAs) have been employed to solve expensive MOPs. Specifically, a certain number of expensive function evaluations are used to build computationally cheap surrogate models for assisting the optimization process without conducting expensive function evaluations. The infill sampling criteria in most existing SAEAs take all requirements on convergence, diversity, and model uncertainty into account, which is, however, not the most efficient in exploiting the limited computational budget. Thus, this article proposes a Kriging-assisted two-archive EA for expensive many-objective optimization. The proposed algorithm uses one influential point-insensitive model to approximate each objective function. Moreover, an adaptive infill criterion that identifies the most important requirement on convergence, diversity, or uncertainty is proposed to determine an appropriate sampling strategy for reevaluations using the expensive objective functions. The experimental results on a set of expensive multi/many-objective test problems have demonstrated its superiority over five state-of-the-art SAEAs.
- 24 March 2021
- Published by Springer Nature
- p. 531-542
Pareto-Based Bi-indicator Infill Sampling Criterion for Expensive Multiobjective Optimization
- 24 March 2021
- Published by Springer Nature
- p. 531-542
Pareto-Based Bi-indicator Infill Sampling Criterion for Expensive Multiobjective Optimization
- Open Access
- 12 June 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 77629-77641
An Improved Brain Storming Optimization Algorithm for Estimating Parameters of Photovoltaic Models
Estimating parameters for various photovoltaic (PV) models is of great importance in simulating, evaluating, and controlling PV systems. To achieve the effective and accurate parameters of PV models, this paper presents an improved Brain Storming Optimization (IBSO). In IBSO, a new individuals’ generation scheme is developed to balance the global and local search capability in the entire iterative process. Furthermore, an improved individual clustering scheme is developed to decrease the computational cost of K-means in the original BSO. In addition, a self-adaptive individuals’ update scheme is introduced to further balance the global and local exploration capability. The proposed IBSO is employed to identify the parameters of different PV models, i.e., single diode, double diode, and PV module. The experiment results show that IBSO can generate an extremely promising performance compared with other state-of-the-art algorithms, especially in terms of accuracy and reliability.
- 12 June 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 77629-77641
Open AccessAn Improved Brain Storming Optimization Algorithm for Estimating Parameters of Photovoltaic Models
Estimating parameters for various photovoltaic (PV) models is of great importance in simulating, evaluating, and controlling PV systems. To achieve the effective and accurate parameters of PV models, this paper presents an improved Brain Storming Optimization (IBSO). In IBSO, a new individuals’ generation scheme is developed to balance the global and local search capability in the entire iterative process. Furthermore, an improved individual clustering scheme is developed to decrease the computational cost of K-means in the original BSO. In addition, a self-adaptive individuals’ update scheme is introduced to further balance the global and local exploration capability. The proposed IBSO is employed to identify the parameters of different PV models, i.e., single diode, double diode, and PV module. The experiment results show that IBSO can generate an extremely promising performance compared with other state-of-the-art algorithms, especially in terms of accuracy and reliability.
- Open Access
- 21 February 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 28211-28229
An Enhanced Brain Storm Sine Cosine Algorithm for Global Optimization Problems
The conventional sine cosine algorithm (SCA) does not appropriately balance exploration and exploitation, causing premature convergence, especially for complex optimization problems, such as the complex shifted or shifted rotated problems. To address this issue, this paper proposes an enhanced brain storm SCA (EBSSCA), where an enhanced brain storm (EBS) strategy is employed to improve the population diversity, and by combining it with two different update equations, two new individual update strategies (IUS: IUS-I and IUS-II) are developed to make effective balance between exploration and exploitation during the entire iterative search process. Double sets of benchmark suites involving 46 popular functions and two real-world problems are employed to compare EBS-SCA with other metaheuristic algorithms. Experimental results validate that the proposed EBS-SCA achieves the best overall performance including the global search ability, convergence speed, and scalability.
- 21 February 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 28211-28229
Open AccessAn Enhanced Brain Storm Sine Cosine Algorithm for Global Optimization Problems
The conventional sine cosine algorithm (SCA) does not appropriately balance exploration and exploitation, causing premature convergence, especially for complex optimization problems, such as the complex shifted or shifted rotated problems. To address this issue, this paper proposes an enhanced brain storm SCA (EBSSCA), where an enhanced brain storm (EBS) strategy is employed to improve the population diversity, and by combining it with two different update equations, two new individual update strategies (IUS: IUS-I and IUS-II) are developed to make effective balance between exploration and exploitation during the entire iterative search process. Double sets of benchmark suites involving 46 popular functions and two real-world problems are employed to compare EBS-SCA with other metaheuristic algorithms. Experimental results validate that the proposed EBS-SCA achieves the best overall performance including the global search ability, convergence speed, and scalability.
- Open Access
- 4 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 78193-78213
A Vector Grouping Learning Brain Storm Optimization Algorithm for Global Optimization Problems
- 4 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 78193-78213
Open AccessA Vector Grouping Learning Brain Storm Optimization Algorithm for Global Optimization Problems
- Open Access
- 29 November 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 76700-76727
A Double Evolutionary Learning Moth-Flame Optimization for Real-Parameter Global Optimization Problems
The original moth-flame optimization (MFO) algorithm neither generates high-performance flames nor utilizes the flames to offer enough effective search guidance for moths in solution spaces, causing the degeneration of the global search capability and convergence speed in confronting complicated problems. To overwhelm those imperfections, this paper proposes a double-evolutionary learning MFO algorithm (DELMFO), where two different evolutionary learning strategies, namely, the differential evolution flame generation (DEFG) and dynamic flame guidance (DFG) strategy, are presented to generate high-performance flames and dynamically guide the search of moths, respectively. By constructing the cascading collaboration between DEFG and DFG, the DELMFO offers a positive feedback channel that makes the personal best historical solutions (PBHSs), flames, and moths promote each other. This improves the global search capability and accelerates convergence speed. The DELMFO is compared with six MFO algorithms and nine popular stochastic optimization algorithms on the CEC2013 test suite. Furthermore, the DELMFO also is further compared with 10 stochastic optimization algorithms on the CEC2017 test suite. Experimental results show that the DELMFO obtains the competitive performance on the global search capability, convergence speed, and scalability among all the algorithms.
- 29 November 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 76700-76727
Open AccessA Double Evolutionary Learning Moth-Flame Optimization for Real-Parameter Global Optimization Problems
The original moth-flame optimization (MFO) algorithm neither generates high-performance flames nor utilizes the flames to offer enough effective search guidance for moths in solution spaces, causing the degeneration of the global search capability and convergence speed in confronting complicated problems. To overwhelm those imperfections, this paper proposes a double-evolutionary learning MFO algorithm (DELMFO), where two different evolutionary learning strategies, namely, the differential evolution flame generation (DEFG) and dynamic flame guidance (DFG) strategy, are presented to generate high-performance flames and dynamically guide the search of moths, respectively. By constructing the cascading collaboration between DEFG and DFG, the DELMFO offers a positive feedback channel that makes the personal best historical solutions (PBHSs), flames, and moths promote each other. This improves the global search capability and accelerates convergence speed. The DELMFO is compared with six MFO algorithms and nine popular stochastic optimization algorithms on the CEC2013 test suite. Furthermore, the DELMFO also is further compared with 10 stochastic optimization algorithms on the CEC2017 test suite. Experimental results show that the DELMFO obtains the competitive performance on the global search capability, convergence speed, and scalability among all the algorithms.
- Open Access
- 30 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 19304-19323
A Brain Storm Optimization With Multi-Information Interactions for Global Optimization Problems
The original brain storm optimization (BSO) fails to consider some potential information interactions in its individual update pattern, causing the premature convergence for complex problems. To address this problem, we propose a BSO algorithm with multi-information interactions (MIIBSO). First, a multi-information interaction (MII) strategy is developed, thoroughly considering various information interactions among individuals. Specially, this strategy contains three new MII patterns. The first two patterns aim to reinforce information interaction capability between individuals. The third pattern provides interactions between the corresponding dimensions of different individuals. The collaboration of the above three patterns is established by an individual stagnation feedback mechanism, contributing to preserve the diversity of the population and enhance the global search capability for MIIBSO. Second, a random grouping (RG) strategy is introduced to replace both the K-means algorithm and cluster center disruption of the original BSO algorithm, further enhancing the information interaction capability and reducing the computational cost of MIIBSO. Finally, a dynamic difference step-size (DDS), which can offer individual feedback information and improve search range, is designed to achieve an effective balance between global and local search capability for MIIBSO. By combining the MII strategy, RG, and DDS, MIIBSO achieves the effective improvement in the global search ability, convergence speed, and computational cost. MIIBSO is compared with 11 BSO algorithms and five other algorithms on the CEC2013 test suit. The results confirm that MIIBSO obtains the best global search capability and convergence speed amongst the 17 algorithms.
- 30 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 19304-19323
Open AccessA Brain Storm Optimization With Multi-Information Interactions for Global Optimization Problems
The original brain storm optimization (BSO) fails to consider some potential information interactions in its individual update pattern, causing the premature convergence for complex problems. To address this problem, we propose a BSO algorithm with multi-information interactions (MIIBSO). First, a multi-information interaction (MII) strategy is developed, thoroughly considering various information interactions among individuals. Specially, this strategy contains three new MII patterns. The first two patterns aim to reinforce information interaction capability between individuals. The third pattern provides interactions between the corresponding dimensions of different individuals. The collaboration of the above three patterns is established by an individual stagnation feedback mechanism, contributing to preserve the diversity of the population and enhance the global search capability for MIIBSO. Second, a random grouping (RG) strategy is introduced to replace both the K-means algorithm and cluster center disruption of the original BSO algorithm, further enhancing the information interaction capability and reducing the computational cost of MIIBSO. Finally, a dynamic difference step-size (DDS), which can offer individual feedback information and improve search range, is designed to achieve an effective balance between global and local search capability for MIIBSO. By combining the MII strategy, RG, and DDS, MIIBSO achieves the effective improvement in the global search ability, convergence speed, and computational cost. MIIBSO is compared with 11 BSO algorithms and five other algorithms on the CEC2013 test suit. The results confirm that MIIBSO obtains the best global search capability and convergence speed amongst the 17 algorithms.