Network control method, apparatus, device, and storage medium

Abstract

The application discloses a network control method and device, equipment and storage medium, and relates to the field of network control.The method comprises the following steps: acquiring time sequence characteristic information of original data, extracting a time sequence subset from the original data based on the time sequence characteristic information, constructing a to-be-mined sample set based on the time sequence subset, performing data mining on the to-be-mined sample set, obtaining a target control parameter, and controlling a target network control model based on the target control parameter.Because the application performs data mining on the constructed to-be-mined sample based on the time sequence characteristic information of the original data, accurate target control parameters are obtained, the time characteristics of dynamic data are fully considered, the induced time delay problem of network control is effectively avoided, and the network control efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network control, and particularly relates to a network control method, device, equipment and storage medium. BACKGROUND

[0002] A network control system is a system for establishing a loop by using a communication network, and generally comprises a controller, an actuator and a sensor, and is a typical closed-loop feedback control system. The network control system is prone to cause data transmission induced time delay due to the introduction of the network. The induced time delay exists in the network and will affect the stability of the system.

[0003] In the prior art, only static data in the network control system is considered in the network control process, and the induced time delay problem of the network control is not considered, so that the network control parameters are inaccurate, the network control efficiency is low, and good control of the network system cannot be achieved.

[0004] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0005] The main purpose of the present application is to provide a network control method, device, equipment and storage medium, which aims to solve the technical problems that in the prior art, only static data in the network control system is considered in the network control process, and the induced time delay problem of the network control is not considered, so that the network control parameters are inaccurate, and the network control efficiency is low.

[0006] To achieve the above purpose, the present application provides a network control method, which comprises the following steps:

[0007] obtaining time sequence feature information of original data;

[0008] extracting a time sequence subset from the original data based on the time sequence feature information, and constructing a to-be-mined sample set based on the time sequence subset;

[0009] performing data mining on the to-be-mined sample set to obtain a target control parameter;

[0010] controlling a target network control model based on the target control parameter.

[0011] Optionally, the data mining on the to-be-mined sample set to obtain a target control parameter comprises:

[0012] performing feature mode evaluation on the to-be-mined sample set;

[0013] converting the to-be-mined sample set into a mode evaluation sample set based on the feature mode evaluation result;

[0014] obtain a parameter prediction model based on the mode evaluation of the sample set;

[0015] input the sample set to be mined into the parameter prediction model for data mining to obtain a target control parameter.

[0016] Optionally, the feature mode evaluation of the sample set to be mined comprises:

[0017] obtain a time sequence in the sample set to be mined;

[0018] determine a target evaluation index according to the time sequence;

[0019] obtain a fractal dimension of each sample in the sample set to be mined based on the target evaluation index;

[0020] perform feature mode evaluation of the sample set to be mined based on the fractal dimension.

[0021] Optionally, the time sequence feature information of the original data comprises:

[0022] collect original data in the transmission network through each sensor of the transmission network;

[0023] obtain a collection period and a collection times of each sensor in the collection process;

[0024] obtain the time sequence feature information of the original data based on the collection period and the collection times.

[0025] Optionally, before the control of the target network control model based on the target control parameter, the method further comprises:

[0026] obtain time delay information between each transmission node in the transmission network;

[0027] construct a signal time sequence formula of the transmission network based on the time delay information;

[0028] construct a target network control model of the transmission network based on the signal time sequence formula.

[0029] Optionally, the construction of the target network control model of the transmission network based on the signal time sequence formula comprises:

[0030] obtain a dimension vector of the transmission network at each time based on the signal time sequence formula;

[0031] construct a target network control model of the transmission network based on the dimension vector of the transmission network at each time.

[0032] Optionally, the construction of the target network control model of the transmission network based on the dimension vector comprises:

[0033] obtaining a state transition matrix, an input matrix and a feedforward matrix of the transmission network;

[0034] constructing a target network control model of the transmission network based on the state transition matrix, the input matrix and the feedforward matrix and a dimension vector at each time in the transmission network.

[0035] In addition, to achieve the above object, the present application further provides a network control device, which comprises:

[0036] an information obtaining module, configured to obtain time sequence characteristic information of original data;

[0037] a sample extracting module, configured to extract a time sequence subset from the original data based on the time sequence characteristic information, and construct a to-be-mined sample set based on the time sequence subset;

[0038] a data mining module, configured to mine the to-be-mined sample set to obtain a target control parameter;

[0039] a network control module, configured to control a target network control model based on the target control parameter.

[0040] In addition, to achieve the above object, the present application further provides a network control device, which comprises a memory, a processor and a network control program stored in the memory and executable on the processor, and the network control program is configured to implement the steps of the network control method as described above.

[0041] In addition, to achieve the above object, the present application further provides a storage medium, which stores a network control program, and the network control program is executable on a processor to implement the steps of the network control method as described above.

[0042] The present application extracts a time sequence subset based on the time sequence characteristic information of original data, constructs a to-be-mined sample set based on the time sequence subset, mines the to-be-mined sample set to obtain a target control parameter, and controls a target network control model based on the target control parameter. Since the present application extracts a time sequence subset based on the time sequence characteristic information of original data, a to-be-mined sample set is constructed, the to-be-mined sample set is mined to obtain an accurate target control parameter, the time characteristics of dynamic data are fully considered, the induced time delay problem of network control is effectively avoided, the network control efficiency is improved, the network control performance is effectively improved, and the network control delay and packet loss rate are reduced. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 is a structural schematic diagram of a network control device of a hardware running environment related to an embodiment of the present application.

[0044] Figure 2 is a flow schematic diagram of a network control method of a first embodiment of the present application.

[0045] Figure 3 is a flow schematic diagram of a network control method of a second embodiment of the present application.

[0046] Figure 4 is a flow schematic diagram of a network control method of a third embodiment of the present application.

[0047] Figure 5 is a structural block diagram of a network control device of a first embodiment of the present application.

[0048] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0049] It should be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application.

[0050] Reference Figure 1 , Figure 1 is a structural schematic diagram of a network control device of a hardware running environment related to an embodiment of the present application.

[0051] As shown in Figure 1 , the network control device can include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection and communication between the components. The user interface 1003 can include a display, an input unit such as a keyboard, and can also include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a wireless fidelity (Wi-Fi) interface). The memory 1005 can be a high-speed random access memory (RAM), and can also be a stable non-volatile memory (NVM), such as a disk memory. The memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0052] Those skilled in the art can understand that Figure 1The structure shown in the figure does not constitute a limitation on the network control device, and can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0053] As shown in Figure 1 The memory 1005 as a storage medium can include an operating system, a network communication module, a user interface module, and a network control program.

[0054] In the network control device shown in Figure 1 The network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the network control device of the application can be arranged in the network control device, and the network control device calls the network control program stored in the memory 1005 through the processor 1001, and executes the network control method provided by the embodiment of the application.

[0055] The embodiment of the application provides a network control method, which refers to Figure 2 , Figure 2 The flowchart of the first embodiment of the network control method of the application.

[0056] In this embodiment, the network control method comprises the following steps:

[0057] Step S10: Obtain the time sequence characteristic information of the original data.

[0058] It should be understood that the execution subject of the method of the embodiment can be a network control device with data processing, network communication and program running functions, such as a computer or other devices or equipment capable of realizing the same or similar functions, which are described above as an example of the network control device.

[0059] It should be noted that the original data can be dynamic data with time characteristics collected by a sensor, or time characteristic data of control parameters, etc., and the above-mentioned sensor can be a time-driven sensor, etc. The above-mentioned time sequence characteristic information can be the time sequence of the original data and other time characteristics.

[0060] It should be understood that sample extraction is an important part of data mining. The time sequence data collected by each sensor is extracted as massive original data that needs to be analyzed and mined, and the spatial and temporal correlation in the massive data is obtained by time analysis of the original data to obtain the time sequence characteristic information of the original data.

[0061] Further, in order to accurately obtain the time sequence characteristic information, the above step S10 can comprise:

[0062] Collecting raw data in the transmission network through sensors in the transmission network;

[0063] Acquiring a collection period and a collection times of each sensor in the collection process;

[0064] Acquiring time sequence feature information of the raw data based on the collection period and the collection times.

[0065] It should be understood that the embodiment collects time-driven data from sensors. Sensor devices or sensor nodes can be used to collect data and store it in a database or file.

[0066] It can be understood that for the collected raw data, data preprocessing can be performed on the raw data, for example, preprocessing can include removing noise, outliers and missing values from the raw data, and extracting time sequence features from the preprocessed data, wherein the time sequence features can include time domain features and frequency domain features.

[0067] Step S20: Extracting a time sequence subset from the raw data based on the time sequence feature information, and constructing a to-be-mined sample set based on the time sequence subset.

[0068] It should be noted that the time sequence subset can be a corresponding data subset extracted from the raw data. The above to-be-mined sample can be a sample set that needs to be data-mined based on the data subset.

[0069] It can be understood that sample extraction is an important part of data mining. By extracting time sequence data collected by each sensor, the time sequence data is used as massive raw data that needs to be analyzed and mined, and according to the spatial and temporal correlation in the massive data, a corresponding time sequence subset is extracted to establish a sample set. The sensor measurement value time sequence and the control parameter time sequence are used as data mining samples to determine the network control system control parameters.

[0070] It should be understood that the embodiment can use ΔT and y1, y2, … y m respectively represent the sampling period and the control parameter, refer to the following formula 1, formula 2 and formula 3, formula 1 is an expression of a to-be-mined sample of k groups of dynamic data, formula 2 is an expression of a sensor measurement value time sequence, and formula 3 is an expression of a control parameter time sequence, wherein represents the sensor measurement value time sequence, Y k represents the control parameter time sequence, and t0 is an initial sampling time.

[0071]

[0072]

[0073] Y k = [y1(t0+kΔT), y2(t0+kΔT),..., y m (t0+kΔT)] T Formula 3

[0074] Step S30: data mining is performed on the to-be-mined sample set to obtain a target control parameter.

[0075] It should be noted that the data mining can be dynamic data mining. The target control parameter can be an optimal control parameter obtained after data mining.

[0076] It should be understood that the embodiment can perform data mining on the to-be-mined sample based on a time sequence, convert the to-be-mined sample into a pattern evaluation sample by using a pattern evaluation method, select a regression analysis method to determine a control parameter prediction model, input the pattern evaluation sample into the control parameter prediction model, and output an optimal target control parameter.

[0077] Step S40: controlling a target network control model based on the target control parameter.

[0078] It should be noted that the embodiment can represent a network generalized object by using a discrete data model, the target network control model can be a network control system mathematical model of a generalized controlled object, and the target network control model can represent the network generalized object by using a discrete mathematical model according to a signal time sequence formula of an induced time delay, as shown in the following formula 4, which is a mathematical model of the target network control model, where G = e AT is a signal transmission rate in a period; x k+1 represents a state of the system at time k+1, which is usually some physical quantity related to system dynamics, such as position, velocity, acceleration, etc.; x k represents a state of the system at time k; u(k) represents a control input of the system at time k; u(k-1) represents a control input of the system at time k-1; Γ0(k) is a control input coefficient matrix, which describes the action of the control input at the current time k. Usually, Γ0(k) is a column vector, and the number of rows is equal to the dimension of the state vector; Γ1(k) is a control input coefficient matrix, which describes the action of the control input at the previous time k-1. Usually, Γ1(k) is a column vector, and the number of rows is equal to the dimension of the state vector.

[0079] x k+1 = Gx k + Γ0(k)u(k) + Γ1(k)u(k-1) Formula 4

[0080] It can be understood that the embodiment is directed to the defect of not considering the induced delay in the prior art, by analyzing the signal with induced delay, and establishing a control parameter prediction model by using a dynamic data mining method, data mining is performed based on the time series of sensor measurement values and control parameters, the optimal control parameter is determined, and good control of the network system is realized.

[0081] The embodiment extracts a time sequence subset from the original data based on the time sequence characteristic information of the original data, constructs a to-be-mined sample set based on the time sequence subset, performs data mining on the to-be-mined sample set, obtains a target control parameter, and controls a target network control model based on the target control parameter; since the time sequence subset is extracted based on the time sequence characteristic information of the original data, the to-be-mined sample set needs to be mined, the accurate target control parameter is obtained by mining the constructed to-be-mined sample set, the time characteristics of dynamic data are fully considered, the induced delay problem of network control is effectively avoided, the network control efficiency is improved, the network control performance is effectively improved, and the network control delay and packet loss rate are reduced.

[0082] Reference Figure 3 , Figure 3 The flowchart of the second embodiment of the network control method is shown.

[0083] Based on the above-mentioned first embodiment, in the embodiment, the step S30 comprises:

[0084] Step S31: Feature mode evaluation is performed on the to-be-mined sample set.

[0085] It should be noted that the time sequence data is represented by a pattern evaluation value sequence through the feature mode evaluation process (feature evaluation generally selects a group of feature vectors to describe the specific mode of a signal. In the embodiment, it can refer to the feature change of the transmission variable between the controller, the sensor and the network).

[0086] Further, in order to effectively perform feature mode evaluation, the above-mentioned step S31 can comprise:

[0087] Step S311: The time sequence in the to-be-mined sample set is obtained.

[0088] Step S312: The target evaluation index is determined according to the time sequence;

[0089] Step S313: The fractal dimension of each sample in each to-be-mined sample set is obtained based on the target evaluation index.

[0090] Step S314: Feature mode evaluation is performed on the to-be-mined sample set based on the fractal dimension.

[0091] It should be noted that the target evaluation index can be an evaluation dimension in the feature mode evaluation process. The fractal dimension can be a measure used to describe the time complexity of each sample in the to-be-mined sample set.

[0092] It should be understood that after the to-be-mined sample is determined, the time series mining network control parameter is utilized to identify the control parameter measurement time series mode index set based on the following formula 5, wherein S ij represents the jth control parameter index of the measurement value x, i time series.

[0093] S i ={S il ,...,S ij ,...,S iq} Formula 5

[0094] Step S32: converting the to-be-mined sample set into a mode evaluation sample set based on the feature mode evaluation result.

[0095] It should be noted that the feature mode evaluation in this embodiment can be based on a fractal dimension evaluation algorithm.

[0096] It should be understood that this embodiment sets the time series of the to-be-determined control parameter The acquired time series is evaluated when the mode is S ij , and the expression thereof is referred to the following formula 6, wherein D is a fractal dimension evaluation algorithm.

[0097]

[0098] It can be understood that the time series in the kth group of dynamic data mining samples is evaluated in the mode, and the expression of the kth group of dynamic data mining samples is referred to the following formula 7, and the time series is represented by using the mode evaluation result. The sample set is converted into the following formula 8, wherein M and K respectively represent the time series.

[0099]

[0100] {(M k ,Y k )} k＝1,2,...,l Formula 8

[0101] Step S33: obtaining a parameter prediction model based on the mode evaluation sample set.

[0102] It should be noted that the parameter prediction model can be a pre-constructed network control parameter prediction model.

[0103] It should be understood that the embodiment converts the data sample into a pattern evaluation sample by using a pattern evaluation method, selects a regression analysis method to determine a control parameter prediction model, and uses the control parameter prediction model to determine the control parameter of the network control system. With respectively represent a measurement time sequence and a control parameter time sequence, a parameter prediction model is established, and an expression of the parameter prediction model is referred to as the following formula 9, wherein: X=[x1, x2,..., x n ] and Y=[y1, y2,..., y n ] are a measurement vector and a control parameter vector respectively, and N is the length of the measurement time sequence.

[0104] Y(τ) = f(X(τ), X(τ-Δt), X(τ-2Δt),..., X(τ-NΔt)) Formula 9

[0105] Step S34: inputting the to-be-mined sample set into the parameter prediction model for data mining to obtain a target control parameter.

[0106] It can be understood that the mining result output by the parameter prediction model is selected to test a test sample set, and the data mining result is output to determine the best control parameter of the network control system.

[0107] It should be understood that the embodiment takes the sensor measurement time sequence and the control parameter time sequence as the data mining sample, determines the to-be-mined sample, mines the network control parameter based on the time sequence, performs feature pattern evaluation, sets an association model according to the control requirement, and determines the best control parameter of the network control system, thereby effectively improving the state response performance and effectively compensating for the network time delay of the network control system. The scheme effectively improves the time delay of the network control system and improves the system response speed.

[0108] The embodiment performs feature pattern evaluation on the to-be-mined sample set, converts the to-be-mined sample set into a pattern evaluation sample set based on the feature pattern evaluation result, obtains a parameter prediction model based on the pattern evaluation sample set, inputs the to-be-mined sample set into the parameter prediction model for data mining, and obtains a target control parameter. Since the embodiment evaluates the to-be-mined sample in the time feature pattern, the time sequence data is represented by using a pattern evaluation value sequence, and then the to-be-mined sample is input into the parameter prediction model for data mining, thereby determining the best control parameter of the network control system and improving the accuracy and precision of data mining and ensuring the network control efficiency.

[0109] Reference Figure 4 , Figure 4 is a flowchart of a third embodiment of a network control method.

[0110] Based on the first embodiment, in the present embodiment, before the step S40, comprising:

[0111] Step S41: obtaining time delay information between each transmission node in the transmission network.

[0112] It should be noted that the transmission node can be an event driver and a time sensor, etc.

[0113] It should be understood that the network control system overall structure includes a network, an event-driven controller, an actuator, a controlled object and a time-driven sensor. The time-driven sensor and the event-driven controller are used in the designed network control system, the time-driven sensor node and the event-driven controller node are set as the controlled object, a unit quantity of data packets are transmitted through the controlled object node in the network, and the control amount consumed from the sending to the arrival of the data packets (from the start node to the last node, the node can be: time sensor, event driver, network, etc.) is calculated.

[0114] Step S42: constructing a signal timing formula of the transmission network based on the time delay information.

[0115] It can be understood that the total time delay of the control loop in the network is and 0≤τ k ≤T, T is the network delay time, and respectively represent the induced time delay from the event-driven controller to the actuator and from the time-driven sensor to the event-driven controller. The linear controlled object is represented by a state equation, and the expression is referred to the following formula 10, in which x(t)∈R n and u(t)∈R are the state and input of the controlled object to be controlled, A and B are suitable dimension matrices, and the following formula 11 is referred to, which is a signal timing formula with induced time delay in the network. When tk≤t<tk+τk, u(t)=uk-1, that is, the control input remains unchanged between the time period of time tk and time tk+τk, and is equal to the input uk-1 at the previous time; when tk+τk≤t<tk+T, u(t)=uk, that is, the control input remains unchanged between the time period of time tk+τk and the next time tk+T, and is equal to the input uk at the current time.

[0116] x(t)=Ax(t)+Bu(t) Formula 10

[0117]

[0118] Step S43: constructing a target network control model of the transmission network based on the signal timing formula.

[0119] It should be understood that, according to the signal timing formula of the induced time delay, the network generalized object is represented by a discrete mathematical model as follows in formula 12, where G = e AT is the signal transmission rate within a period; x k+1 represents the state of the system at time k+1, which is usually some physical quantity related to the dynamics of the system, such as position, velocity, acceleration, etc.; x k represents the state of the system at time k; u(k) represents the control input of the system at time k; u(k-1) represents the control input of the system at time k-1; Γ0(k) is a control input coefficient matrix that describes the effect of the control input at the current time k. Usually, Γ0(k) is a column vector with a number of rows equal to the dimension of the state vector; Γ1(k) is a control input coefficient matrix that describes the effect of the control input at the previous time k-1. Usually, Γ1(k) is a column vector with a number of rows equal to the dimension of the state vector.

[0120] x k+1 = Gx k + Γ0(k)u(k) + Γ1(k)u(k-1) formula 12

[0121] It can be understood that, are the signal discrete transmission rates in different time periods within a period. Since Γ0(k) and Γ1(k) are time-varying, the following formula 13 is referred to.

[0122]

[0123] Further, in order to construct a high-quality target network control model, the above step S43 can include:

[0124] Step S431: obtaining the dimension vector of the transmission network at each time based on the signal timing formula;

[0125] Step S432: obtaining the state transition matrix, input matrix and feedforward matrix of the transmission network;

[0126] Step S433: constructing the target network control model of the transmission network based on the state transition matrix, the input matrix and the feedforward matrix, and the dimension vector of the transmission network at each time.

[0127] It should be understood that the above formula 12 can be converted to the following formula 14 by the present embodiment, wherein: Δu(k) is a time delay hysteresis; is a network bandwidth limit; u(k-1) is a control input of the system at time k-1; Δu(k) represents a disturbance input at time k, i.e., an external interference to which the system is subjected, which can cause the system state to change; G is a state transition matrix, which describes the evolution rule of the system state without external input. Generally, G is a square matrix, and the dimension thereof is equal to the number of state variables; H is an input matrix, which describes the influence of the control input on the system state. Generally, H is a column vector or a matrix, and the number of rows thereof is equal to the dimension of the state vector; Γ0(k) is a disturbance input coefficient matrix, which describes the influence of the disturbance input on the system state. Generally, Γ0(k) is a column vector or a matrix, and the number of rows thereof is equal to the dimension of the state vector.

[0128] x k+1 = Gx k + Hu k-1 + Γ0(k) Δu(k) Formula 14

[0129] It can be understood that the matrix A is converted to the following formula 15, wherein: r i ≥ 0, i = 1, 2, 3, r1+r2+r3=n; J and J are diagonal matrices corresponding to each different eigenvalue and all repeated eigenvalues, respectively; Λ is a matrix containing the eigenvectors of A; A represents a matrix, and the dimension thereof is r1+r2+r3×r1+r2+r3; Λ represents a diagonal matrix, and the dimension thereof is r1+r2+r3×r1+r2+r3. The elements on the diagonal are composed of the eigenvalues of the three sub-matrices, and the remaining elements are 0.

[0130] A = Λ diag(0 r1×r2 ,J r1×r2 ,J r1×r2 ) Λ -1 Formula 15

[0131] It should be noted that diag(0 r1×r2 ,J r2×r2 ,Jρ r3×r3 ) represents a diagonal block matrix, and the dimension thereof is r1+r2+r3×r1+r2+r3. Among them, 0 r1×r1 is a r1×r1 zero matrix, J r2×r2 is a r2×r2 unit matrix, and J r3×r3 is a r3×r3 Jordan block matrix. This diagonal block matrix divides the matrix A into three block matrices, wherein the first block matrix is a zero matrix, the second block matrix is a unit matrix, and the third block matrix is a Jordan block matrix. Λ ^-1denotes the inverse of Λ, which is also of dimension r1+r2+r3 x r1+r2+r3.

[0132] It is understood that if there is only one zero eigenvalue and r nonzero eigenvalues in matrix A, A can be transformed into Equation 16 below, where J1 and J2 are diagonal blocks and Jordan blocks, respectively. Γ0(k) can be transformed into Equation 17 below, where Γ0(k) denotes a matrix of dimension r1+r2+r3 x r1+r2+r3. Λ denotes a diagonal matrix of dimension r1+r2+r3 x r1+r2+r3. The elements on the diagonal are composed of the eigenvalues of the three sub-matrices, and the other elements are 0.

[0133] A = Λ diag(0, J1, J2) Λ -1 Equation 16

[0134]

[0135] It is noted that diag in Equation 17 above denotes a diagonal block matrix of dimension r1+r2+r3 x r1+r2+r3. Where T-τk is a real number, eλ2(T-τk)-1λ2, …, eλn-ρ(T-τk)-1λn-ρ is a set of complex numbers, and J2 is a Jordan block matrix of dimension r2 x r2. This diagonal block matrix separates matrix Λ -1 B into three block matrices, where the first block matrix is a real number, the second block matrix is a set of complex numbers, and the third block matrix is a Jordan block matrix. Λ ^-1 denotes the inverse of Λ, which is also of dimension r1+r2+r3 x r1+r2+r3, and B denotes a matrix of dimension r1+r2+r3 x 1.

[0136] It is understood that let D = Λ diag(α1, α2, …, α n-r α * ), and E = Λ -1 B, by reasonably selecting α1, α2, …, α n-r , α * , F T (τ k )F(τ k )≤D holds, the following Equation 18 can be obtained, where D and E are constant matrices, and the mathematical model of the network control system representing the discretized generalized controlled object can be obtained, which is referred to as Equation 19 below, x k denotes the state vector of the system at time k, which is an n-dimensional column vector, and u k-1x(k) represents the input vector of the system at time k-1, which is an m-dimensional column vector, G represents the state transition matrix of the system, which is an n*n matrix, H represents the input matrix of the system, which is an n*m matrix; DFE represents the feedforward matrix of the system, which is an n*p matrix, and Δuk represents the input increment of the system at time k, which is a p-dimensional column vector.

[0137] Γ0(k)=DF(τ k )E Formula 18

[0138] x k+1 =Gx k +Hu k-1 +DFEΔu k Formula 19

[0139] The embodiment obtains the delay information between each transmission node in the transmission network, constructs a signal timing formula of the transmission network based on the delay information, and constructs a target network control model of the transmission network based on the signal timing formula. Since the embodiment constructs the target network control model in advance, the network generalized object is represented by using the discrete mathematical model, the network control is accurately performed, and the control delay and the packet loss rate are effectively reduced.

[0140] In addition, the embodiment of the present application also provides a storage medium, and the storage medium stores a network control program. When the network control program is executed by a processor, the steps of the network control method described above are implemented.

[0141] Since the storage medium adopts all the technical solutions of the above-mentioned embodiments, at least all the beneficial effects brought by the technical solutions of the above-mentioned embodiments are achieved, and here, they will not be repeated.

[0142] Referring to Figure 5 , Figure 5 FIG. 1 is a structural block diagram of a network control device according to a first embodiment of the present application.

[0143] As shown in Figure 5 , the network control device according to the embodiment of the present application comprises:

[0144] An information acquisition module 10 is configured to acquire time sequence characteristic information of original data.

[0145] A sample extraction module 20 is configured to extract a time sequence subset from the original data based on the time sequence characteristic information, and construct a to-be-mined sample set based on the time sequence subset.

[0146] A data mining module 30 is configured to perform data mining on the to-be-mined sample set, and obtain a target control parameter.

[0147] The network control module 40 is configured to control the target network control model based on the target control parameter.

[0148] Further, the data mining module 30 is further configured to evaluate the feature mode of the sample set to be mined, convert the sample set to be mined into a mode evaluation sample set based on the evaluation result of the feature mode, obtain a parameter prediction model based on the mode evaluation sample set, input the sample set to be mined into the parameter prediction model for data mining, and obtain the target control parameter.

[0149] Further, the network control device further comprises:

[0150] The feature mode evaluation module 50 is configured to obtain a time sequence in the sample set to be mined, determine a target evaluation index according to the time sequence, obtain a fractal dimension of each sample in the sample set to be mined based on the target evaluation index, and evaluate the feature mode of the sample set to be mined based on the fractal dimension.

[0151] The information acquisition module 10 is further configured to collect original data in the transmission network through each sensor of the transmission network, obtain a collection period and a collection times of each sensor in the collection process, and obtain time sequence feature information of the original data based on the collection period and the collection times.

[0152] The network control device further comprises:

[0153] The network control model construction module 60 is configured to obtain time delay information between each transmission node in the transmission network, construct a signal time sequence formula of the transmission network based on the time delay information, and construct a target network control model of the transmission network based on the signal time sequence formula.

[0154] The network control model construction module 60 is further configured to obtain a dimension vector of the transmission network at each time based on the signal time sequence formula, and construct the target network control model of the transmission network based on the dimension vector of the transmission network at each time.

[0155] The network control model construction module 60 is further configured to obtain a state transition matrix, an input matrix and a feedforward matrix of the transmission network, and construct the target network control model of the transmission network based on the state transition matrix, the input matrix, the feedforward matrix and the dimension vector of the transmission network at each time.

[0156] The embodiment obtains time sequence feature information of original data, extracts a time sequence subset from the original data based on the time sequence feature information, constructs a to-be-mined sample set based on the time sequence subset, performs data mining on the to-be-mined sample set, obtains a target control parameter, and controls a target network control model based on the target control parameter; since the time sequence subset is extracted based on the time sequence feature information of the original data, a to-be-mined sample set that needs to be subjected to data mining is constructed, accurate target control parameters are obtained by performing data mining on the constructed to-be-mined sample set, the time characteristics of dynamic data are fully considered, the induced time delay problem of network control is effectively avoided, the network control efficiency is improved, the network control performance is effectively improved, and the network control delay and packet loss rate are reduced.

[0157] It should be understood that the above is only illustrative, and does not constitute any limitation on the technical solutions of the present application. In specific applications, those skilled in the art can set up as needed, and the present application does not limit this.

[0158] It should be noted that the above-described workflow is only illustrative and does not limit the scope of protection of the present application. In actual applications, those skilled in the art can select part or all of them to achieve the purpose of the embodiment scheme according to actual needs, which is not limited here.

[0159] In addition, technical details not described in detail in the embodiment can be referred to the network control method provided by any embodiment of the present application, which will not be repeated here.

[0160] In addition, it should be noted that in this document, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or system. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of another identical element in the process, method, article or system that includes the element.

[0161] The above embodiment numbers of the present application are only for description, not representing the advantages and disadvantages of the embodiments.

[0162] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, can also be through hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part of the prior art contribution can be embodied in the form of software products, the computer software product is stored in a storage medium (such as read only memory (Read Only Memory, ROM) / RAM, disk, optical disk), including a number of instructions to make a terminal device (may be a mobile phone, computer, server, or network equipment, etc.) executes the method described in various embodiments of the present application.

[0163] The above is only the preferred embodiment of the present application, not therefore limit the patent scope of the present application, any equivalent structure or equivalent flow transformation made by using the content of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A network control method characterized by comprising: The network control method comprises: obtaining time sequence characteristic information of original data; extracting a time sequence subset from the original data based on the time sequence characteristic information, and constructing a to-be-mined sample set based on the time sequence subset; performing data mining on the to-be-mined sample set to obtain a target control parameter; controlling a target network control model based on the target control parameter; the data mining on the to-be-mined sample set to obtain the target control parameter comprises: performing feature mode evaluation on the to-be-mined sample set; converting the to-be-mined sample set into a mode evaluation sample set based on the feature mode evaluation result; obtaining a parameter prediction model based on the mode evaluation sample set; inputting the to-be-mined sample set into the parameter prediction model for data mining to obtain the target control parameter.

2. The network control method of claim 1, wherein, the feature mode evaluation on the to-be-mined sample set comprises: obtaining a time sequence in the to-be-mined sample set; determining a target evaluation index according to the time sequence; obtaining a fractal dimension of each sample in the to-be-mined sample set based on the target evaluation index; performing feature mode evaluation on the to-be-mined sample set based on the fractal dimension.

3. The network control method of claim 1, wherein, the time sequence characteristic information of the original data comprises: collecting original data in a transmission network through sensors of the transmission network; obtaining a collection period and a collection times of each sensor in the collection process; obtaining time sequence characteristic information of the original data based on the collection period and the collection times.

4. The network control method according to any one of claims 1 to 3, wherein, before the controlling the target network control model based on the target control parameter, further comprising: obtaining time delay information between transmission nodes in a transmission network; constructing a signal time sequence formula of the transmission network based on the time delay information; constructing a target network control model of the transmission network based on the signal time sequence formula.

5. The network control method of claim 4, wherein, the constructing the target network control model of the transmission network based on the signal time sequence formula comprises: obtaining a dimension vector of the transmission network at each time based on the signal time sequence formula; constructing the target network control model of the transmission network based on the dimension vector of the transmission network at each time.

6. The network control method of claim 5, wherein, the constructing the target network control model of the transmission network based on the dimension vector of the transmission network at each time comprises: obtaining a state transition matrix, an input matrix and a feedforward matrix of the transmission network; constructing the target network control model of the transmission network based on the state transition matrix, the input matrix, the feedforward matrix and the dimension vector of the transmission network at each time.

7. A network control device, characterized by the network control device comprises: an information acquisition module for obtaining time sequence characteristic information of original data; a sample extraction module for extracting a time sequence subset from the original data based on the time sequence characteristic information, and constructing a to-be-mined sample set based on the time sequence subset; The data mining module is configured to perform data mining on the sample set to be mined to obtain a target control parameter. The data mining module is further configured to perform feature pattern evaluation on the sample set to be mined, convert the sample set to be mined into a pattern evaluation sample set based on a feature pattern evaluation result, obtain a parameter prediction model based on the pattern evaluation sample set, and input the sample set to be mined into the parameter prediction model to perform data mining to obtain a target control parameter. The network control module is configured to control a target network control model based on the target control parameter.

8. A network control device, characterized by The network control device comprises a memory, a processor, and a network control program stored in the memory and executable on the processor. The network control program is configured to implement the network control method according to any one of claims 1 to 6.

9. A storage medium, characterized by The storage medium stores a network control program. The network control program is executable on the processor to implement the network control method according to any one of claims 1 to 6.

Images