In the fog, haze and other severe atmospheric conditions, the image will be severely degraded, the practical value of the image will be affected greatly. In this paper, the removal of haze for the degraded image is studied. Firstly, the basic principle of fog imaging is analyzed. Then, through the analysis of the dark channel prior principle, a novel dark channel prior haze removal algorithm with the peak signal-to-noise ratio is proposed to improve the dehazing image details clear degree effectively. Finally, the simulation results are compared with histogram equalization, Multi-scale Retinex and other classical dark channel prior algorithms to verify the effectiveness of the proposed algorithm.
A novel coupled fault isolation scheme is proposed to deal with multiple, simultaneous and coupled nonlinear faults for traction motor with structural damage. The structural damage reduced-order model is firstly built by introducing the relative balance factor associated with big data driven classifier algorithm. The proposed discrete-time fault model has been established with the help of Delta operator and support vector machine (SVM). Meanwhile, fault isolation condition and threshold determination techniques are used to isolate the coupled faults. Finally, through a coupled fault isolation experiments using real data of fault-injection benchmark for traction drive control system, the effectiveness of the proposed algorithm is justified.
It is difficult to effectively forecast overall sales volume of short life cycle experiential products using traditional methods with data point forecasts during the preparation period, because of fuzzification of variables and a lack of historical sales data. To address this problem, a method for forecasting interval reliability classification of short life cycle experiential product during preparation period is proposed based on rough set and evidence theory. This method use rough set theory to get reliability function from each variables and use evidence theory to set up comprehensive reliability. The comprehensive reliability value of the training set is used to construct the classification reliability interval, so as to judge the classification reliability interval from the comprehensive reliability value of the test set , and obtain the classification reliability interval of the test set finally. 621 samples selected from domestic films released during 2016-2018 are used to verify the effectiveness of the proposed method, and the intersectional verification results shows the forecasting method has great accuracy.
A sufficient conditions for the convergence of D-type open-loop iterative learning controller is proposed for multiple input multiple output continuous repeat system in the presence of bounded external state disturbance and actuator failure. Under the conditions of the initial state is equal and the initial state exist bounded offset, based on λ-norm, this paper proposed the sufficient conditions for the convergence of the iterative learning controller respectively. It is proved that in the sense of λ norm, under the two conditions, the convergence conditions of the iterative controller are the same, only the upper bound of tracking error between the output of fault system and the expected trajectory is different. Based on Schur's complement principle, linear matrix inequality is given to determine the optimal control gain at a given convergence rate. Numerical simulations verify the effectiveness and feasibility of the proposed control strategy.
In order to improve the accuracy of image annotation, and solve the problem of how to integrate text features and visual features of images, this paper puts forward an image annotation algorithm based on dual theme space. Firstly, the visual features and text annotation of image are represented as two views of the same object. Based on the multiple statistical analysis theory of partial least squares (PLS), the semantic dual relationship between the two feature spaces is considered to extract the dual shared semantic information. Then, a non-probabilistic annotation model is constructed on the symmetric space composed of dual theme. Finally, the predicted annotation vectors are calculated by the projection of visual features on dual theme space, the annotation texts of a new image are selected by setting the threshold. The algorithm performance is tested on the public data set of Core5K, experiments show that the proposed algorithm based on dual mode semantic space can effectively improve the performance of image annotation and the number of annotation accuracy.
Potato shape is one of the important indicators of potato grading, this paper discusses a potato shape classification method based on machine vision technology, combined with the principal component analysis-support vector machine (PCA-SVM) algorithm. This method extracts the eigenvectors of eleven-dimensional in a single potato region,which can represent shapes, and principal component analysis (PCA) is used to reduce the dimension of the feature vector and extract the principal component features of the shape. Then, the principal component feature is brought into the support vector machine(SVM)for modeling, and the grid search method (GS) is used to optimize the parameters of SVM. In the detection, the images of potato samples were taken into the PCA model and the optimized SVM model by using the ten-fold cross validation (CV) algorithm to classify the potatoes. Experimental results show that the algorithm proposed in this paper has the higher sorting speed as well as the higher accuracy (97.3 %). The method is feasible for potato shape sorting and can be used for automatic potato grading.
Key words: Potato, machine vision, shape sorting, principal component
The scheduling problem model of distributed heterogeneous parallel machine is presented in the background of the actual production problems in industrial production. Then, a hybrid fruit fly optimization algorithm is designed to minimize the maximum completion time of the considered problem. Firstly, the competition mechanism is added in the initialization phase of the algorithm, which effectively improves the quality of the initial solution. Secondly, the adaptive search radius is introduced in the smell search stage to effectively search the solution space. Finally, the three-phase local search is integrated into the update phase of the algorithm, so global search and local search can achieve a better balance. Simulation experiments and algorithm comparisons verify the effectiveness and robustness of the proposed hybrid fruit fly optimization algorithm.
Due to end effect exists in Hilbert transform and low pass filtering of Hilbert vibration decomposition,an improved Hilbert vibration decomposition method is proposed. Mirror-symmetric extension method is adopted in signal preprocessing in Hilbert transform, low pass filtering and synchronous detection, and the end effect caused by Hilbert transform and low pass filtering was eliminated. Subsequently, instantaneous frequency obtained by Hilbert transform and low pass filtering process was regarded as reference frequency. Finally, several components with different amplitudes were obtained by using synchronous detection and iterative operation. The simulation signal and experiment data generated from roller bearing demonstrated that the proposed method has excellent performance in vibration signal decomposition and end effect suppression, and it can effectively diagnose the roller bearing defect appeared on inner and outer race. So it has certain practical engineering application value.
A novel super twisted decoupling nonsingular fast terminal sliding mode control (STDNFTSMC) method is proposed for the existence of singular phenomena and sliding mode jitter in a class of four order under actuated system decoupling terminal sliding mode control. The under actuated system is divided into two subsystems, and nonsingular fast terminal sliding mode surfaces are designed, respectively. The saturation function of sliding surface of one subsystem is used to construct an intermediate variable, and the variable is introduced into the sliding surface of another subsystem to construct the sliding surface of the whole system. Equivalent control method and super-twisted reaching law are used to solve the control law of the system. The chattering phenomenon is effectively eliminated and the robustness of the system is improved. The Lyapunov method is used to prove the asymptotic stability of the sliding surface of each system. The simulation results show the effectiveness of the method.
This paper investigates the problem of H2/H∞ control for nonlinear time-delay stochastic systems with state, control, and disturbance–dependent noise. Firstly, a sufficient condition for the existence of the nonlinear stochastic H2/H∞ control with time-delay is presented in terms of coupled Hamilton–Jacobi equations (HJEs). Secondly, by using the T-S fuzzy model, the stochastic H2/H∞ controller can be designed via solving a set of linear matrix inequalities (LMIs) instead of coupled HJEs. Finally, a numerical example is employed to show the effectiveness of the results obtained.
A discrete wolf pack algorithm is proposed to simulate the specific characteristics of the permutation flow shop scheduling problem and simulate the hunting behavior of wolves in nature. Using the coding method based on the workpiece sequence, the opposition learning initializes the population to improve the convergence speed of the algorithm. The redesign of the wandering behavior, summoning behavior and siege behavior in the original wolf group algorithm makes the algorithm not easy to fall into local optimum. At the same time, the Taguchi experimental design method is used to analyze the sensitivity of the algorithm parameter settings, and the optimal parameter combination is determined. Finally, the discrete wolf pack algorithm is used to simulate and test the standard test sets of Car, Reeves and Taillard. Compared with other intelligent optimization algorithms, the feasibility of the proposed algorithm is verified. It provides a more effective method for solving the permutation flow shop scheduling problem.
Wind power and photovoltaic output are uncertain. In view of the mismatched usable power and system natural load, an integrated dispatching system is put forward, which integrates the generation-side resources such as wind power, PV, thermal power and the demand-side resources such as electric vehicles and adjustable load. On the one hand, based on the fuzzy chance constrained programming model, the non-deterministic problems are equivalently rewritten for problem modeling and solving. On the other hand, considering the cost of electricity generation and consumption in the optimization objectives, the generation-side and demand-side interaction mechanism was applied in the constraints, so as to improve the supply-and-demand relationship between the renewable energy's output and load at a lower cost. Influences of the participation degrees of electric vehicles and demand response on the dispatching results were tested by digital simulation. The results show that the proposed method can not only match the load and output, but also effectively raise the economic efficiency.
With the increasingly prominent information security issues at home and abroad, the need for localization of information system equipment and the independence of information operation and maintenance is increasingly urgent. At the same time, with the continuous deepening of the application of information systems, a significant increase in the number of users and services has caused a bottleneck in system performance, and performance optimization is urgently needed. There are many successful cases in China that are subject to third-party integration companies, with closed source software and low controllability. Therefore, an optimization scheme of X86-based distributed storage architecture platform is proposed. This solution uses distributed storage, multi-node redundancy in computing nodes, storage nodes, and switches. It also uses a dual-active high-availability database system with primary and standby databases, which increases system reliability, reduces operation and maintenance costs, and achieves information Deep autonomous operation and maintenance of communication. The experiments of the feasibility analysis of the optimization method prove that the optimization scheme of the new platform is always better than the existing related optimization schemes.
According to the principle of time scale separation, a terminal sliding mode control method is proposed based on adaptive weighted reaching law to solve the attitude control problem of flapping-wing aircraft with high nonlinearity, internal model parameter disturbance and external disturbance. The inner and outer loop sliding mode controllers are designed by the length of adjustment time. An adaptive weighted reaching law is designed, which combines the advantages of power and exponential reaching law to eliminate chattering and improve the tracking speed of the system. The adaptive control method is used to estimate the model parameter disturbance on-line, which compensates the output of the sliding mode controller and reduces the steady-state error of the system. The simulation results show that the adaptive terminal sliding mode control not only eliminates the sliding mode chattering problem, but also effectively overcomes the impact of external and internal disturbances.
Terminal security throughput has been the focus of attention. In order to optimize airport security procedures, reduce the number of security channels and reduce operating costs, an optimization model based on Petri net and queuing theory is proposed. The homogeneous Markov chain is built by using Petri nets, and the screening process is analyzed. The token, library, change of Petri nets model respectively mapped to passengers and baggage, their state and each step of the security. The weight of limiting factors on security procedures is analyzed and the security check process bottlenecks is found out. So a new model is put forward, and verify its reliability. Based on the reliability of the model, the security check layout is analyzed and its layout is adjusted. The results show that the standard deviation of token number and transition utilization in Petri net decreases, and the waiting time tends to equalization, which improve the efficiency. According to queuing theory, the number of open channels is optimized, and corresponding improvement measures are given.
Aiming at the four levels closed-loop supply chain system composed of suppliers, manufacturers, sellers and third-party recycling, the system dynamics method is used to solve the quality control problem of four levels closed-loop supply chain. Firstly, the cause-and-effect loop diagram of four levels closed-loop supply chain quality control is plotted by Vensim simulation software, and the relationship between the influencing factors and the model boundary are found out. Secondly, the equations of each variable are designed, and the system flow diagram and the quality control model of the four levels closed-loop supply chain are further established. Meanwhile, rationality and authenticity of the model are tested. Finally, a numerical example is carried out to verify that the improvement of raw material quality fluctuation level and service quality level can improve the total profit of members at all levels of the four levels closed-loop supply chain, and it also can improve the overall performance of the closed-loop supply chain.
In order to overcome the defects existing in the optimization strategy of warehouse layout problem, this paper proposes an improved clonal selection algorithm based on the introduction of vaccine strategy. This algorithm adopts roulette algorithm in the process of vaccine selection and vaccination. Cloning scale, combining antibody affinity and antibody concentration, is calculated in the process of cloning and proliferation. The algorithm also retains excellent antibodies during cloning inhibition and introduces random antibodies. These optimization strategies which improve the convergence speed and optimization efficiency of the algorithm and increase the population diversity are used to realize the storage layout which satisfies the demand of the access job and the given condition constraint. Simulation analyses and comparisons show that this algorithm has the characteristics of faster convergence speed and shorter access path, thus greatly improving storage efficiency.
In the cloud environment, the virtual machine deployment directly affects the overall performance of the data center. The concept of virtual machines affinity is proposed according to the relationship attributes among virtual machines. And a deployment strategy based on multi population genetic algorithm combined with the penalty function method is applied considering the load balancing of the physical machines and the affinity of the virtual machines. In order to avoid the local optimum, Gauss learning is carried out on the optimal individuals. Simulation results show that the deployment strategy with high load balancing and good affinity could be achieved through the multi population genetic algorithm, which has strong robustness, fast convergence speed and can solve the virtual machine deployment problem effectively in the cloud environment.
A new multi-model extended Kalman filter algorithm is proposed for a kind of nonlinear systems, in which, the state equation includes a linear part and a nonlinear part, the measurement equation is a linear function. Firstly, the multi-model Kalman filter method is improved and named as IMMEKF. Secondly, the original system is divided into a linear part and a non-linear part. Thirdly, in the process of time update, the Kalman filter algorithm is used to predict the state of the linear part, and the improved multi-model extended Kalman filter algorithm is utilized to predict the state of the nonlinear part. Then, in the measurement update process, a sequential updating method is given to correct the predicted value of the linear part and the non-linear part gradually. The final simulation results illustrate the nonlinear filtering property of the two filtering methods.
Based on the communication topology from algebraic graph theory, the consensus protocol and coordination control of multiple unmanned aerial vehicles have always been a hot research topic in the field of automatic control. Since the attitude dynamics model of unmanned aerial vehicles is an Euler quaternion, the design of attitude synchronization controller of multiple unmanned aerial vehicles is very difficult to perform. Under time-varying communication topology, in this paper, the attitude synchronization control problem of multiple unmanned aerial vehicles is investigated. Furthermore, the implementation conditions of attitude synchronization are analyzed and the synchronization control algorithm is also designed. In the case of undirected graph and time-varying communication topology, it is proved that the attitude control algorithm can not only realize the attitude synchronization of multiple unmanned aerial vehicles, but also can guarantee the final angular velocity converge to zero. Under different circumstances of fixed switching cycles, time-varying, fast time-varying, slower time-varying, and isolated point in communication topology, based on the above results, the attitude synchronization control problems of multiple unmanned aerial vehicles are compared by computer simulation, and the effectiveness and the robustness of the proposed attitude synchronization control algorithm are also verified.
In view of the non-linearity, time-varying and large delay characteristics of SCR flue gas denitrification system, a multi-model switching DMC-PID cascade predictive control method for variable conditions operation is proposed under the background of coal-fired units participating in deep peak shaving. Firstly, the local linear model of SCR denitrification system under different working conditions is established by using system identification method. Secondly, the parameters of DMC-PID controller are designed according to the principle of quadratic optimal control. Finally, the switching conditions are designed according to the unit load instructions, and the DMC-PID cascade predictive control of denitrification system under variable conditions is realized by using multi-mode switching control method. The results show that compared with DMC-PID cascade predictive control without switching control and cascade PID multi-mode switching control with switching control, the multi-mode switching DMC-PID cascade predictive control designed in this paper has the advantages of small output fluctuation, small overshoot, short adjustment time, rapid response, strong anti-interference ability and good robustness under variable operating conditions, and is suitable for industrial process control with non-linearity, time-varying, large delay and variable operating conditions.
The reasonable planning of energy storage capacity is carried out from the point of view of stabilizing the random fluctuation of new energy output, and the joint probability model of new energy storage system is used to describe the output of micro power supply. A method of probabilistic power flow calculation based on improved sample ranking method is proposed to simplify the correlation control method of multi input random variables of traditional stochastic power flow. In order to reduce the system active power loss, reduce the probability of node voltage exceeding the limit, and consider the operation cost of energy storage device, a multi-objective optimization model is constructed, and the particle swarm optimization algorithm combined with chaotic optimization and linear decreasing inertia weight is used to solve the problem. Finally, an example analysis is carried out in Matlab, and the results show that the stochastic power flow method combined with the joint probabilistic model can effectively solve the multi-objective optimization model, and the system uncertainty is restrained under the constraint condition.
This paper considers two-agent scheduling with release times where agent A and agent B have to share m parallel machines for processing their jobs. The objective of agent A is to minimize the total completion time, while the objective of agent B is to ensure its total completion time under a fixed value. Firstly, the considered problem is proved to be NP-hard under the single machine condition. Secondly, a pseudo-polynomial-time algorithm is proposed by using the dynamic programming (DP) method, and then a fully polynomial-time approximation algorithm is further provided.
With the development of the economy, distributed permutation production becomes more and more widespread in all walks of life, more attention is put on distributed models. An adaptive estimation of distribution algorithm(AEDA) is proposed for solving no-wait distributed permutation flow shop with Sequence Dependent Setup Times and Arrival times(NDPFSP with SDSTs and RDs) and minimizing the largest completion time. First of all, this paper proposes the earliest completion factory with arrival time(ECFAT), which is more suitable for this problem, canmake a judgment in the generation process of solution and improve the quality of the current solution faster. Next, according to different problem scales, the proposed local search can adjust the depth of local search to ensure its ability for solving this problem.
For industrial users, the prediction of air-conditioning load can be great helpful to reduce the energy consumption. Besides the nonlinear character, the user's load data is also very sensitive to the external disturbance and with daily periodic characteristics. This makes the traditional ARMA and SVR methods hard to achieve good prediction. Therefore, a method of combining ARMA model and SVR model with daily periodic characteristics is presented. Firstly, combined with the daily periodic characteristic of the raw data, the ARMA model is used for linear prediction. For the nonlinear characteristics of original data which retained in the prediction residuals of ARMA model, the SVR model is used to predict the nonlinear part of the residuals and modify the prediction result, obtaining the final predicted value. Experimental results using actual data show that the proposed method can significantly improve the prediction performance.
For the single-valued neutrosophic multi-attribute decision making (MADM) problems that the attribute weights are completely unknown, a novel MADM method is developed on the basis of single-valued neutrosophic entropy. First, a new axiomatic definition of single-valued neutrosophic entropy is introduced. Then, by using the cosine function, a single-valued neutrosophic information entropy formula is constructed to measure the uncertainty of SVNV, and it is proved that the constructed formula satisfies the four axiomatic requirements of single-valued neutrosophic entropy. In addition, a programming model is proposed to determine optimal attribute weights with the principle of minimum uncertain information, and a single-valued neutrosophic MADM method is investigated that the attribute weights are completely unknown. In the end, a numerical example of selection for data product service provider is provided, and the rationality and effectiveness of the developed method are certified by comparing with the existing method.
For the multi-attribute decision-making (MADM) problems that the attribute weights are completely unknown under the interval-valued normal fuzzy environment, two interval-valued normal fuzzy information aggregation algorithms are designed, a novel interval-valued normal fuzzy MADM model is investigated. First, the interval-valued normal fuzzy operational laws are defined. Then, from the arithmetical and geometric point of view, the interval-valued normal fuzzy weighted averaging operator (IVNFWA) and interval-valued normal fuzzy weighted geometric operator (IVNFWG) are proposed. The relationship between these two operators is analyzed. Finally, combining these two proposed operators with distances, a new model is developed to deal with the interval-valued normal fuzzy MADM problems in which the attribute weights are completely unknown, and then apply the developed model to research on the selection of database system. The results obtained from performance analysis show that the proposed approach is correct, feasible and efficient.
In order to improve the performance of position sensorless control system of surface mounted permanent magnet synchronous motor (SPMSM), the extended state observer is applied to sensorless control of surface-mounted permanent magnet synchronous motor to estimate the rotor position angle and speed. It is proposed to estimate the rotor position by constructing an extended state observer with stator current as the main variable in two-phase stationary coordinate system and observing the armature back-EMF. An extended state observer with estimated rotor position angle as the main variable is constructed for speed estimation. This method has high accuracy of rotor position angle and speed estimation when the resistance and inductance parameters of the motor change. It is suitable for middle and high speed sensorless control of surface-mounted permanent magnet synchronous motor. The simulation and experimental results show the effectiveness of the proposed method.