Deep learning algorithm is an effective data mining method and has been used in many fields to solve practical problems.However,the deep learning algorithms often contain some hyper-parameters which may be continuous,...Deep learning algorithm is an effective data mining method and has been used in many fields to solve practical problems.However,the deep learning algorithms often contain some hyper-parameters which may be continuous,integer,or mixed,and are often given based on experience but largely affect the effectiveness of activity recognition.In order to adapt to different hyper-parameter optimization problems,our improved Cuckoo Search(CS)algorithm is proposed to optimize the mixed hyper-parameters in deep learning algorithm.The algorithm optimizes the hyper-parameters in the deep learning model robustly,and intelligently selects the combination of integer type and continuous hyper-parameters that make the model optimal.Then,the mixed hyper-parameter in Convolutional Neural Network(CNN),Long-Short-Term Memory(LSTM)and CNN-LSTM are optimized based on the methodology on the smart home activity recognition datasets.Results show that the methodology can improve the performance of the deep learning model and whether we are experienced or not,we can get a better deep learning model using our method.展开更多
Due to outstanding performance in cheminformatics,machine learning algorithms have been increasingly used to mine molecular properties and biomedical big data.The performance of machine learning models is known to cri...Due to outstanding performance in cheminformatics,machine learning algorithms have been increasingly used to mine molecular properties and biomedical big data.The performance of machine learning models is known to critically depend on the selection of the hyper-parameter configuration.However,many studies either explored the optimal hyper-parameters per the grid searching method or employed arbitrarily selected hyper-parameters,which can easily lead to achieving a suboptimal hyper-parameter configuration.In this study,Hyperopt library embedding with the Bayesian optimization is employed to find optimal hyper-parameters for different machine learning algorithms.Six drug discovery datasets,including solubility,probe-likeness,h ERG,Chagas disease,tuberculosis,and malaria,are used to compare different machine learning algorithms with ECFP6 fingerprints.This contribution aims to evaluate whether the Bernoulli Na?ve Bayes,logistic linear regression,Ada Boost decision tree,random forest,support vector machine,and deep neural networks algorithms with optimized hyper-parameters can offer any improvement in testing as compared with the referenced models assessed by an array of metrics including AUC,F1-score,Cohen’s kappa,Matthews correlation coefficient,recall,precision,and accuracy.Based on the rank normalized score approach,the Hyperopt models achieve better or comparable performance on 33 out 36 models for different drug discovery datasets,showing significant improvement achieved by employing the Hyperopt library.The open-source code of all the 6 machine learning frameworks employed in the Hyperopt python package is provided to make this approach accessible to more scientists,who are not familiar with writing code.展开更多
Based on generalized the variation method, by introducing Hamilton function and Lagrange multiplier, this paper proposed a linear quadratic optimal control strategy for an incomplete controllable system with fixed ter...Based on generalized the variation method, by introducing Hamilton function and Lagrange multiplier, this paper proposed a linear quadratic optimal control strategy for an incomplete controllable system with fixed terminal state and time. Applying the proposed optimal control to the simple two-input dual-stage actuator magnetic head positioning system with three degrees-of-freedom, the simulation results show that the system has no residual vibration at the terminal position and time, which can reduce the total access time during head positioning process. To verify the validation of the optimal control strategy of three degrees-of-freedom spring-mass models in actual magnetic head positioning of hard disk drives, a finite element model of an actual magnetic head positioning system is presented. Substituting the optimal control force from simple three degrees-of-freedom spring-mass models into the finite element model, the simulation results show that the magnetic head also has no residual vibration at the end of track-to-track travel. That is to say, the linear quadratic optimal control technique based on simple two-input dual- stage actuator system with three degrees-of-freedom proposed in this paper is of high reliability for the industrial application of an actual magnetic head positioning system.展开更多
In lean combustion mode,exhaust gas ratio(EGR)is a significant factor that affects fuel economy and combustion stability.A proper EGR level is beneficial for the fuel economy;however,the combustion stability(coefficie...In lean combustion mode,exhaust gas ratio(EGR)is a significant factor that affects fuel economy and combustion stability.A proper EGR level is beneficial for the fuel economy;however,the combustion stability(coefficient of variation(COV)in indicated mean effective pressure(IMEP))deteriorated monotonously with increasing EGR.The aim of this study is to achieve a trade-off between the fuel economy and combustion stability by optimizing the EGR set-point.A cost function(J)is designed to represent the trade-off and reduce the calibration burden for optimal EGR at different engine operating conditions.An extremum-seeking(ES)algorithm is adopted to search for the extreme value of J and obtain the optimal EGR at an operating point.Finally,a map of optimal EGR set-value is designed and experimentally validated on a real driving cycle.展开更多
This paper focuses on solving the distributed optimization problem with binary-valued intermittent measurements of local objective functions.In this paper,a binary-valued measurement represents whether the measured va...This paper focuses on solving the distributed optimization problem with binary-valued intermittent measurements of local objective functions.In this paper,a binary-valued measurement represents whether the measured value is smaller than a fixed threshold.Meanwhile,the“intermittent”scenario arises when there is a non-zero probability of not detecting each local function value during the measuring process.Using this kind of coarse measurement,the authors propose a discrete-time stochastic extremum seeking-based algorithm for distributed optimization over a directed graph.As is well-known,many existing distributed optimization algorithms require a doubly-stochastic weight matrix to ensure the average consensus of agents.However,in practical engineering,achieving doublestochasticity,especially for directed graphs,is not always feasible or desirable.To overcome this limitation,the authors design a row-stochastic matrix and a column-stochastic matrix as weight matrices in the proposed algorithm instead of relying on doubly-stochasticity.Under some mild conditions,the authors rigorously prove that agents can reach the average consensus and ultimately find the optimal solution.Finally,the authors provide a numerical example to illustrate the effectiveness of the algorithm.展开更多
Efficiency and emissions of spark-ignited engines are significantly affected by combustion phase which can usually be indicated by crank angle of 50% mass burnt （CA50）. Managing combustion phase at the optimal value...Efficiency and emissions of spark-ignited engines are significantly affected by combustion phase which can usually be indicated by crank angle of 50% mass burnt （CA50）. Managing combustion phase at the optimal value at which the maximal efficiency can be achieved is a challenging issue due to the cyclic variations of combustion process. This paper addresses this issue in two loops： CA50 set-point optimization （outer loop） and set-point tracking （inner loop） by controlling spark advance （SA）. Extremum seeking approach maximizing thermal efficiency is employed in the CA50 set-point optimization. A proportional- integral （PI） controller is adopted to make the moving average value of CA50 tracking the optimal CA50 set-point determined in the outer loop. Moreover, in order to obtain fast responses at steady and transient operations, feed-forward maps are designed for extremum seeking controller and PI controller, respectively. Finally, experimental validations are conducted on a six-cylinder gasoline at steady and transient operations to show the effectiveness of proposed control scheme.展开更多
Buildings are a major energy consumer and carbon emitter,therefore it is important to improve building energy efficiency to achieve our sustainable development goal.Deep reinforcement learning(DRL),as an advanced buil...Buildings are a major energy consumer and carbon emitter,therefore it is important to improve building energy efficiency to achieve our sustainable development goal.Deep reinforcement learning(DRL),as an advanced building control method,demonstrates great potential for energy efficiency optimization and improved occupant comfort.However,the performance of DRL is highly sensitive to hyper-parameters,and selecting inappropriate hyper-parameters may lead to unstable learning or even failure.This study aims to investigate the design and application of DRL in building energy system control,with a specific focus on improving the performance of DRL controllers through hyper-parameter optimization(HPO)algorithms.It also aims to provide quantitative evaluation and adaptive validation of these optimized controllers.Two widely used algorithms,deep deterministic policy gradient(DDPG)and soft actor-critic(SAC),are used in the study and their performance is evaluated in different building environments based on the BOPTEST virtual testbed.One of the focuses of the study is to compare various HPO techniques,including tree-structured Parzen estimator(TPE),covariance matrix adaptation evolution strategy(CMA-ES),and combinatorial optimization methods,to determine the efficacy of different hyper-parameter optimization methods for DRL.The study enhances HPO efficiency through parallel computation and conducts a comprehensive quantitative assessment of the optimized DRL controllers,considering factors such as reduced energy consumption and improved comfort.The results show that the HPO algorithms significantly improve the performance of the DDPG and SAC controllers.A reduction of 56.94%and 68.74%in thermal discomfort is achieved,respectively.Additionally,the study demonstrates the applicability of the HPO-based approach for enhancing DRL controller performance across diverse building environments,providing valuable insights for the design and optimization of building DRL controllers.展开更多
This paper proposes a comprehensive design scheme for the extremum seeking control(ESC)of the unmanned aerial vehicle(UAV)close formation flight.The proposed design scheme combines a Newton-Raphson method with an exte...This paper proposes a comprehensive design scheme for the extremum seeking control(ESC)of the unmanned aerial vehicle(UAV)close formation flight.The proposed design scheme combines a Newton-Raphson method with an extended Kalman filter(EKF)to dynamically estimate the optimal position of the following UAV relative to the leading UAV.To reflect the wake vortex effects reliably,the drag coefficient induced by the wake vortex is considered as a performance function.Then,the performance function is parameterized by the first-order and second-order terms of its Taylor series expansion.Given the excellent performance of nonlinear estimation,the EKF is used to estimate the gradient and the Hessian matrix of the parameterized performance function.The output feedback of the proposed scheme is determined by iterative calculation of the Newton-Raphson method.Compared with the traditional ESC and the classic ESC,the proposed design scheme avoids the slow continuous time integration of the gradient.This allows a faster convergence of relative position extremum.Furthermore,the proposed method can provide a smoother command during the seeking process as the second-order term of the performance function is taken into account.The convergence analysis of the proposed design scheme is accomplished by showing that the output feedback is a supermartingale sequence.To improve estimation performance of the EKF,a improved pigeon-inspired optimization(IPIO)is proposed to automatically tune the noise covariance matrix.Monte Carlo simulations for a three-UAV close formation show that the proposed design scheme is robust to the initial position of the following UAV.展开更多
Dear Editor,This letter explores optimal formation control for a network of unmanned surface vessels(USVs).By designing an individual objective function for each USV,the optimal formation problem is transformed into a...Dear Editor,This letter explores optimal formation control for a network of unmanned surface vessels(USVs).By designing an individual objective function for each USV,the optimal formation problem is transformed into a noncooperative game.Under this game theoretic framework,the optimal formation is achieved by seeking the Nash equilibrium of the regularized game.A modular structure consisting of a distributed Nash equilibrium seeker and a regulator is proposed.展开更多
A continuous-time stochastic model is constructed to analyze how to control rent seeking behaviors. Using the stochastic optimization methods based on the modem risky theory, a unique positive solution to the dynamic ...A continuous-time stochastic model is constructed to analyze how to control rent seeking behaviors. Using the stochastic optimization methods based on the modem risky theory, a unique positive solution to the dynamic model is derived. The effects of preference-related parameters on the optimal control level of rent seeking are discussed, and some policy measures are given. The results show that there exists a unique solution to the stochastic dynamic model under some macroeconomic assumptions, and that raising public expenditure may have reverse effects on rent seeking in an underdeveloped or developed economic environment.展开更多
In computer vision,convolutional neural networks have a wide range of uses.Images representmost of today’s data,so it’s important to know how to handle these large amounts of data efficiently.Convolutional neural ne...In computer vision,convolutional neural networks have a wide range of uses.Images representmost of today’s data,so it’s important to know how to handle these large amounts of data efficiently.Convolutional neural networks have been shown to solve image processing problems effectively.However,when designing the network structure for a particular problem,you need to adjust the hyperparameters for higher accuracy.This technique is time consuming and requires a lot of work and domain knowledge.Designing a convolutional neural network architecture is a classic NP-hard optimization challenge.On the other hand,different datasets require different combinations of models or hyperparameters,which can be time consuming and inconvenient.Various approaches have been proposed to overcome this problem,such as grid search limited to low-dimensional space and queuing by random selection.To address this issue,we propose an evolutionary algorithm-based approach that dynamically enhances the structure of Convolution Neural Networks(CNNs)using optimized hyperparameters.This study proposes a method using Non-dominated sorted genetic algorithms(NSGA)to improve the hyperparameters of the CNN model.In addition,different types and parameter ranges of existing genetic algorithms are used.Acomparative study was conducted with various state-of-the-art methodologies and algorithms.Experiments have shown that our proposed approach is superior to previous methods in terms of classification accuracy,and the results are published in modern computing literature.展开更多
High-dimensional and incomplete(HDI) matrices are primarily generated in all kinds of big-data-related practical applications. A latent factor analysis(LFA) model is capable of conducting efficient representation lear...High-dimensional and incomplete(HDI) matrices are primarily generated in all kinds of big-data-related practical applications. A latent factor analysis(LFA) model is capable of conducting efficient representation learning to an HDI matrix,whose hyper-parameter adaptation can be implemented through a particle swarm optimizer(PSO) to meet scalable requirements.However, conventional PSO is limited by its premature issues,which leads to the accuracy loss of a resultant LFA model. To address this thorny issue, this study merges the information of each particle's state migration into its evolution process following the principle of a generalized momentum method for improving its search ability, thereby building a state-migration particle swarm optimizer(SPSO), whose theoretical convergence is rigorously proved in this study. It is then incorporated into an LFA model for implementing efficient hyper-parameter adaptation without accuracy loss. Experiments on six HDI matrices indicate that an SPSO-incorporated LFA model outperforms state-of-the-art LFA models in terms of prediction accuracy for missing data of an HDI matrix with competitive computational efficiency.Hence, SPSO's use ensures efficient and reliable hyper-parameter adaptation in an LFA model, thus ensuring practicality and accurate representation learning for HDI matrices.展开更多
The real-time capability of integrated flight/propulsion optimal control （IFPOC） is studied. An appli- cation is proposed for IFPOC by combining the onboard hybrid aero-engine model with sequential quadratic pro- gr...The real-time capability of integrated flight/propulsion optimal control （IFPOC） is studied. An appli- cation is proposed for IFPOC by combining the onboard hybrid aero-engine model with sequential quadratic pro- gramming （SQP）. Firstly, a steady-state hybrid aero-engine model is designed in the whole flight envelope with a dramatic enhancement of real-time capability. Secondly, the aero-engine performance seeking control including the maximum thrust mode and the minimum fuel-consumption mode is performed by SQP. Finally, digital simu- lations for cruise and accelerating flight are carried out. Results show that the proposed method improves real- time capability considerably with satisfactory effectiveness of optimization.展开更多
This paper first proposes a distributed continuous-time Newton-Raphson algorithm for heterogeneous linear multi-agent systems over unbalanced digraphs.Then this approach extends to cases where the local cost functions...This paper first proposes a distributed continuous-time Newton-Raphson algorithm for heterogeneous linear multi-agent systems over unbalanced digraphs.Then this approach extends to cases where the local cost functions and Hessian matrices are unknown.While local exponential stability of the inverse Hessian matrix estimator has been established for single-agent systems,this paper proves local exponential stability in multi-agent systems,ensuring the stability of the proposed distributed Newton-Raphson extremum seeking algorithm.A numerical example demonstrates the effectiveness of the proposed algorithms.展开更多
Since the beginning of this year,the international environment has been complex and volatile,the international economic and trade order has suffered severe setbacks,and instability and uncertainty have increased signi...Since the beginning of this year,the international environment has been complex and volatile,the international economic and trade order has suffered severe setbacks,and instability and uncertainty have increased significantly.Faced with this complex situation,China's textile industry has adhered to the general principle of seeking progress while maintaining stability,steadily advancing the optimization and adjustment of its industrial structure,and deepening the transformation and upgrading of foreign trade.Supported by the country's more proactive and effective macroeconomic policies,the economy remained generally stable in the first half of the year,with exports maintaining growth despite significant pressure,and its resilience being consolidated and unleashed.Looking ahead to the second half of 2025,the textile industry will continue to face numerous challenges while consolidating its stable and positive development foundation.展开更多
基金Supported by the Anhui Province Sports Health Information Monitoring Technology Engineering Research Center Open Project (KF2023012)。
文摘Deep learning algorithm is an effective data mining method and has been used in many fields to solve practical problems.However,the deep learning algorithms often contain some hyper-parameters which may be continuous,integer,or mixed,and are often given based on experience but largely affect the effectiveness of activity recognition.In order to adapt to different hyper-parameter optimization problems,our improved Cuckoo Search(CS)algorithm is proposed to optimize the mixed hyper-parameters in deep learning algorithm.The algorithm optimizes the hyper-parameters in the deep learning model robustly,and intelligently selects the combination of integer type and continuous hyper-parameters that make the model optimal.Then,the mixed hyper-parameter in Convolutional Neural Network(CNN),Long-Short-Term Memory(LSTM)and CNN-LSTM are optimized based on the methodology on the smart home activity recognition datasets.Results show that the methodology can improve the performance of the deep learning model and whether we are experienced or not,we can get a better deep learning model using our method.
基金financial support provided by the National Key Research and Development Project(2019YFC0214403)Chongqing Joint Chinese Medicine Scientific Research Project(2021ZY023984)。
文摘Due to outstanding performance in cheminformatics,machine learning algorithms have been increasingly used to mine molecular properties and biomedical big data.The performance of machine learning models is known to critically depend on the selection of the hyper-parameter configuration.However,many studies either explored the optimal hyper-parameters per the grid searching method or employed arbitrarily selected hyper-parameters,which can easily lead to achieving a suboptimal hyper-parameter configuration.In this study,Hyperopt library embedding with the Bayesian optimization is employed to find optimal hyper-parameters for different machine learning algorithms.Six drug discovery datasets,including solubility,probe-likeness,h ERG,Chagas disease,tuberculosis,and malaria,are used to compare different machine learning algorithms with ECFP6 fingerprints.This contribution aims to evaluate whether the Bernoulli Na?ve Bayes,logistic linear regression,Ada Boost decision tree,random forest,support vector machine,and deep neural networks algorithms with optimized hyper-parameters can offer any improvement in testing as compared with the referenced models assessed by an array of metrics including AUC,F1-score,Cohen’s kappa,Matthews correlation coefficient,recall,precision,and accuracy.Based on the rank normalized score approach,the Hyperopt models achieve better or comparable performance on 33 out 36 models for different drug discovery datasets,showing significant improvement achieved by employing the Hyperopt library.The open-source code of all the 6 machine learning frameworks employed in the Hyperopt python package is provided to make this approach accessible to more scientists,who are not familiar with writing code.
基金Project supported by the National Natural Science Foundation of China （No. 10472038）；the Science Foundation of the Ministry of Education of China for Ph.D. Programme （No. 20050730016）；the National Science Foundation of China for 0utstanding Young Researchers （No. 10025208）.
文摘Based on generalized the variation method, by introducing Hamilton function and Lagrange multiplier, this paper proposed a linear quadratic optimal control strategy for an incomplete controllable system with fixed terminal state and time. Applying the proposed optimal control to the simple two-input dual-stage actuator magnetic head positioning system with three degrees-of-freedom, the simulation results show that the system has no residual vibration at the terminal position and time, which can reduce the total access time during head positioning process. To verify the validation of the optimal control strategy of three degrees-of-freedom spring-mass models in actual magnetic head positioning of hard disk drives, a finite element model of an actual magnetic head positioning system is presented. Substituting the optimal control force from simple three degrees-of-freedom spring-mass models into the finite element model, the simulation results show that the magnetic head also has no residual vibration at the end of track-to-track travel. That is to say, the linear quadratic optimal control technique based on simple two-input dual- stage actuator system with three degrees-of-freedom proposed in this paper is of high reliability for the industrial application of an actual magnetic head positioning system.
文摘In lean combustion mode,exhaust gas ratio(EGR)is a significant factor that affects fuel economy and combustion stability.A proper EGR level is beneficial for the fuel economy;however,the combustion stability(coefficient of variation(COV)in indicated mean effective pressure(IMEP))deteriorated monotonously with increasing EGR.The aim of this study is to achieve a trade-off between the fuel economy and combustion stability by optimizing the EGR set-point.A cost function(J)is designed to represent the trade-off and reduce the calibration burden for optimal EGR at different engine operating conditions.An extremum-seeking(ES)algorithm is adopted to search for the extreme value of J and obtain the optimal EGR at an operating point.Finally,a map of optimal EGR set-value is designed and experimentally validated on a real driving cycle.
基金supported by the National Natural Science Foundation of China under Grant No.62473272the Natural Science Foundation of Sichuan Province,China under Grant No.2024NSFSC0437。
文摘This paper focuses on solving the distributed optimization problem with binary-valued intermittent measurements of local objective functions.In this paper,a binary-valued measurement represents whether the measured value is smaller than a fixed threshold.Meanwhile,the“intermittent”scenario arises when there is a non-zero probability of not detecting each local function value during the measuring process.Using this kind of coarse measurement,the authors propose a discrete-time stochastic extremum seeking-based algorithm for distributed optimization over a directed graph.As is well-known,many existing distributed optimization algorithms require a doubly-stochastic weight matrix to ensure the average consensus of agents.However,in practical engineering,achieving doublestochasticity,especially for directed graphs,is not always feasible or desirable.To overcome this limitation,the authors design a row-stochastic matrix and a column-stochastic matrix as weight matrices in the proposed algorithm instead of relying on doubly-stochasticity.Under some mild conditions,the authors rigorously prove that agents can reach the average consensus and ultimately find the optimal solution.Finally,the authors provide a numerical example to illustrate the effectiveness of the algorithm.
文摘Efficiency and emissions of spark-ignited engines are significantly affected by combustion phase which can usually be indicated by crank angle of 50% mass burnt （CA50）. Managing combustion phase at the optimal value at which the maximal efficiency can be achieved is a challenging issue due to the cyclic variations of combustion process. This paper addresses this issue in two loops： CA50 set-point optimization （outer loop） and set-point tracking （inner loop） by controlling spark advance （SA）. Extremum seeking approach maximizing thermal efficiency is employed in the CA50 set-point optimization. A proportional- integral （PI） controller is adopted to make the moving average value of CA50 tracking the optimal CA50 set-point determined in the outer loop. Moreover, in order to obtain fast responses at steady and transient operations, feed-forward maps are designed for extremum seeking controller and PI controller, respectively. Finally, experimental validations are conducted on a six-cylinder gasoline at steady and transient operations to show the effectiveness of proposed control scheme.
基金supported by the National Natural Science Foundation of China(No.72371072)Jiangsu Association for Science&Technology Youth Science&Technology Talents Lifting Project(No.JSTJ-2023-JS001).
文摘Buildings are a major energy consumer and carbon emitter,therefore it is important to improve building energy efficiency to achieve our sustainable development goal.Deep reinforcement learning(DRL),as an advanced building control method,demonstrates great potential for energy efficiency optimization and improved occupant comfort.However,the performance of DRL is highly sensitive to hyper-parameters,and selecting inappropriate hyper-parameters may lead to unstable learning or even failure.This study aims to investigate the design and application of DRL in building energy system control,with a specific focus on improving the performance of DRL controllers through hyper-parameter optimization(HPO)algorithms.It also aims to provide quantitative evaluation and adaptive validation of these optimized controllers.Two widely used algorithms,deep deterministic policy gradient(DDPG)and soft actor-critic(SAC),are used in the study and their performance is evaluated in different building environments based on the BOPTEST virtual testbed.One of the focuses of the study is to compare various HPO techniques,including tree-structured Parzen estimator(TPE),covariance matrix adaptation evolution strategy(CMA-ES),and combinatorial optimization methods,to determine the efficacy of different hyper-parameter optimization methods for DRL.The study enhances HPO efficiency through parallel computation and conducts a comprehensive quantitative assessment of the optimized DRL controllers,considering factors such as reduced energy consumption and improved comfort.The results show that the HPO algorithms significantly improve the performance of the DDPG and SAC controllers.A reduction of 56.94%and 68.74%in thermal discomfort is achieved,respectively.Additionally,the study demonstrates the applicability of the HPO-based approach for enhancing DRL controller performance across diverse building environments,providing valuable insights for the design and optimization of building DRL controllers.
基金supported by the National Natural Science Foundation of China(Grant Nos.91948204,U20B2071,T2121003 and U1913602)Open Fund/Postdoctoral Fund of the Laboratory of Cognition and Decision Intelligence for Complex Systems,Institute of Automation,Chinese Academy of Sciences(Grant No.CASIA-KFKT-08)。
文摘This paper proposes a comprehensive design scheme for the extremum seeking control(ESC)of the unmanned aerial vehicle(UAV)close formation flight.The proposed design scheme combines a Newton-Raphson method with an extended Kalman filter(EKF)to dynamically estimate the optimal position of the following UAV relative to the leading UAV.To reflect the wake vortex effects reliably,the drag coefficient induced by the wake vortex is considered as a performance function.Then,the performance function is parameterized by the first-order and second-order terms of its Taylor series expansion.Given the excellent performance of nonlinear estimation,the EKF is used to estimate the gradient and the Hessian matrix of the parameterized performance function.The output feedback of the proposed scheme is determined by iterative calculation of the Newton-Raphson method.Compared with the traditional ESC and the classic ESC,the proposed design scheme avoids the slow continuous time integration of the gradient.This allows a faster convergence of relative position extremum.Furthermore,the proposed method can provide a smoother command during the seeking process as the second-order term of the performance function is taken into account.The convergence analysis of the proposed design scheme is accomplished by showing that the output feedback is a supermartingale sequence.To improve estimation performance of the EKF,a improved pigeon-inspired optimization(IPIO)is proposed to automatically tune the noise covariance matrix.Monte Carlo simulations for a three-UAV close formation show that the proposed design scheme is robust to the initial position of the following UAV.
基金supported by the National Key R&D Program of China(2022ZD0119604)the National Natural Science Foundation of China(NSFC),(62222308,62173181,62221004)+1 种基金the Natural Science Foundation of Jiangsu Province(BK20220139)the Young Elite Scientists Sponsorship Program by CAST(2021QNRC001)。
文摘Dear Editor,This letter explores optimal formation control for a network of unmanned surface vessels(USVs).By designing an individual objective function for each USV,the optimal formation problem is transformed into a noncooperative game.Under this game theoretic framework,the optimal formation is achieved by seeking the Nash equilibrium of the regularized game.A modular structure consisting of a distributed Nash equilibrium seeker and a regulator is proposed.
基金The Social Science Foundation of China(No06BJY011)
文摘A continuous-time stochastic model is constructed to analyze how to control rent seeking behaviors. Using the stochastic optimization methods based on the modem risky theory, a unique positive solution to the dynamic model is derived. The effects of preference-related parameters on the optimal control level of rent seeking are discussed, and some policy measures are given. The results show that there exists a unique solution to the stochastic dynamic model under some macroeconomic assumptions, and that raising public expenditure may have reverse effects on rent seeking in an underdeveloped or developed economic environment.
基金This research was supported by the Researchers Supporting Program(TUMAProject-2021-27)Almaarefa University,Riyadh,Saudi Arabia.
文摘In computer vision,convolutional neural networks have a wide range of uses.Images representmost of today’s data,so it’s important to know how to handle these large amounts of data efficiently.Convolutional neural networks have been shown to solve image processing problems effectively.However,when designing the network structure for a particular problem,you need to adjust the hyperparameters for higher accuracy.This technique is time consuming and requires a lot of work and domain knowledge.Designing a convolutional neural network architecture is a classic NP-hard optimization challenge.On the other hand,different datasets require different combinations of models or hyperparameters,which can be time consuming and inconvenient.Various approaches have been proposed to overcome this problem,such as grid search limited to low-dimensional space and queuing by random selection.To address this issue,we propose an evolutionary algorithm-based approach that dynamically enhances the structure of Convolution Neural Networks(CNNs)using optimized hyperparameters.This study proposes a method using Non-dominated sorted genetic algorithms(NSGA)to improve the hyperparameters of the CNN model.In addition,different types and parameter ranges of existing genetic algorithms are used.Acomparative study was conducted with various state-of-the-art methodologies and algorithms.Experiments have shown that our proposed approach is superior to previous methods in terms of classification accuracy,and the results are published in modern computing literature.
基金supported in part by the National Natural Science Foundation of China (62372385, 62272078, 62002337)the Chongqing Natural Science Foundation (CSTB2022NSCQ-MSX1486, CSTB2023NSCQ-LZX0069)the Deanship of Scientific Research at King Abdulaziz University, Jeddah, Saudi Arabia (RG-12-135-43)。
文摘High-dimensional and incomplete(HDI) matrices are primarily generated in all kinds of big-data-related practical applications. A latent factor analysis(LFA) model is capable of conducting efficient representation learning to an HDI matrix,whose hyper-parameter adaptation can be implemented through a particle swarm optimizer(PSO) to meet scalable requirements.However, conventional PSO is limited by its premature issues,which leads to the accuracy loss of a resultant LFA model. To address this thorny issue, this study merges the information of each particle's state migration into its evolution process following the principle of a generalized momentum method for improving its search ability, thereby building a state-migration particle swarm optimizer(SPSO), whose theoretical convergence is rigorously proved in this study. It is then incorporated into an LFA model for implementing efficient hyper-parameter adaptation without accuracy loss. Experiments on six HDI matrices indicate that an SPSO-incorporated LFA model outperforms state-of-the-art LFA models in terms of prediction accuracy for missing data of an HDI matrix with competitive computational efficiency.Hence, SPSO's use ensures efficient and reliable hyper-parameter adaptation in an LFA model, thus ensuring practicality and accurate representation learning for HDI matrices.
基金Supported by the Aeronautical Science Foundation of China(2010ZB52011)the Funding of Jiangsu Innovation Program for Graduate Education(CXLX11-0213)the Nanjing University of Aeronautics and Astronautics Research Funding(NS2010055)~~
文摘The real-time capability of integrated flight/propulsion optimal control （IFPOC） is studied. An appli- cation is proposed for IFPOC by combining the onboard hybrid aero-engine model with sequential quadratic pro- gramming （SQP）. Firstly, a steady-state hybrid aero-engine model is designed in the whole flight envelope with a dramatic enhancement of real-time capability. Secondly, the aero-engine performance seeking control including the maximum thrust mode and the minimum fuel-consumption mode is performed by SQP. Finally, digital simu- lations for cruise and accelerating flight are carried out. Results show that the proposed method improves real- time capability considerably with satisfactory effectiveness of optimization.
基金supported in part by the National Natural Science Foundation of China(NSFC)under Grant No.62373314in part by the Research Grants Council of the Hong Kong Special Administrative Region of China under Project City U/11207323in part by the NSFC-Excellent Young Scientists Fund(Hong Kong and Macao)under Grant No.62222318。
文摘This paper first proposes a distributed continuous-time Newton-Raphson algorithm for heterogeneous linear multi-agent systems over unbalanced digraphs.Then this approach extends to cases where the local cost functions and Hessian matrices are unknown.While local exponential stability of the inverse Hessian matrix estimator has been established for single-agent systems,this paper proves local exponential stability in multi-agent systems,ensuring the stability of the proposed distributed Newton-Raphson extremum seeking algorithm.A numerical example demonstrates the effectiveness of the proposed algorithms.
文摘Since the beginning of this year,the international environment has been complex and volatile,the international economic and trade order has suffered severe setbacks,and instability and uncertainty have increased significantly.Faced with this complex situation,China's textile industry has adhered to the general principle of seeking progress while maintaining stability,steadily advancing the optimization and adjustment of its industrial structure,and deepening the transformation and upgrading of foreign trade.Supported by the country's more proactive and effective macroeconomic policies,the economy remained generally stable in the first half of the year,with exports maintaining growth despite significant pressure,and its resilience being consolidated and unleashed.Looking ahead to the second half of 2025,the textile industry will continue to face numerous challenges while consolidating its stable and positive development foundation.
