WO2020199114A1 - 控制方法、装置、电子设备以及存储介质 - Google Patents
控制方法、装置、电子设备以及存储介质 Download PDFInfo
- Publication number
- WO2020199114A1 WO2020199114A1 PCT/CN2019/080965 CN2019080965W WO2020199114A1 WO 2020199114 A1 WO2020199114 A1 WO 2020199114A1 CN 2019080965 W CN2019080965 W CN 2019080965W WO 2020199114 A1 WO2020199114 A1 WO 2020199114A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- coupling strength
- control
- network
- control node
- nodes
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/02—Details
Definitions
- the present invention relates to the technical field of complex network control, in particular to a control method, device, electronic equipment and storage medium.
- the complex network traction control must first consider three key issues. The first is to select which nodes to perform traction control; the second is to select how many nodes to perform traction control; the third is to select which traction controller.
- the present invention provides a control method, device, electronic equipment and storage medium to solve the technical problem that the criterion adopted by the existing control method cannot accurately reflect the coupling strength of the control node, resulting in poor network control effect.
- the present invention provides a control method, including: selecting control nodes from the network nodes to obtain a control node set; constructing a network state equation according to the control node set; wherein the network state equation includes a coupling term and a feedback term.
- the coupling term Including the node coupling coefficient; determine the coupling strength matrix according to the coupling terms and feedback terms, and obtain the maximum and minimum eigenvalues of the coupling strength matrix; determine the first constant and the second constant according to the maximum eigenvalue, the minimum eigenvalue and the coupling coefficient; Determine the coupling strength coefficient according to the maximum characteristic value, the minimum characteristic value, the first constant and the second constant; determine whether the number of nodes in the control node set is less than the preset number; if the judgment result is yes, then when the coupling strength coefficient is greater than When the preset standard coupling strength coefficient is used, the coupling strength coefficient is used to update the standard coupling strength coefficient and the control node set is updated until the number of nodes in the control node set is equal to the preset number; if the judgment result is no , Then output the control node set corresponding to the standard coupling strength coefficient to use the nodes in the control node set to control the network.
- a network state equation introducing control nodes is constructed, and the maximum and minimum eigenvalues, the first and second constants are obtained according to the network state equation, and the maximum eigenvalue, the minimum eigenvalue, and the A constant and a second constant determine the coupling strength coefficient, and the obtained network coupling strength coefficient can more accurately reflect the coupling strength of the system, and a control node set with a larger coupling strength coefficient is selected to control the network to achieve a better control network.
- ⁇ represents the coupling strength coefficient
- ⁇ N represents the maximum eigenvalue
- ⁇ 1 represents the minimum eigen value
- ⁇ 1 represents the first constant
- ⁇ 2 represents the second constant
- ⁇ 1 ⁇ 2 ⁇ 0 ⁇ 1 ⁇ 2 ⁇ 0.
- the ratio of the first constant to the minimum eigenvalue is subtracted from the ratio of the second constant to the maximum eigenvalue to obtain the coupling strength coefficient, so that the coupling strength coefficient more accurately reflects the coupling strength, And use this as a criterion for the controllability of the network system, and select a set of control nodes with a larger coupling strength coefficient to control the network to achieve better control effects.
- the updating the set of control nodes specifically includes: selecting a node from the nodes of the network to add to the set of control nodes, so as to update the set of control nodes.
- the convergence speed can be avoided and the optimization can enter the local optimum.
- the specific expression of the coupling strength matrix is:
- B represents the coupling strength matrix
- L represents the Laplacian matrix of the network
- D represents the feedback matrix of the network.
- control device is introduced below, and its implementation principle and technical effect are similar to the above-mentioned method principle and technical effect, and will not be repeated here.
- the present invention provides a control device, which is characterized by comprising: an obtaining module for selecting a control node from network nodes to obtain a control node set; a building module for constructing a network state equation based on the control node set; Among them, the network state equation includes coupling terms and feedback terms, and the coupling terms include node coupling coefficients; the determination module is used to determine the coupling strength matrix according to the coupling terms and feedback terms, and obtain the maximum and minimum eigenvalues of the coupling strength matrix; determine The module is also used to determine the first constant and the second constant according to the maximum characteristic value, the minimum characteristic value and the coupling coefficient; the determining module is also used to determine the coupling strength coefficient according to the maximum characteristic value, the minimum characteristic value, the first constant and the second constant; The judgment module is used to judge whether the number of nodes in the control node set is less than the preset number; the update module is used to if the judgment result is yes, when the coupling strength coefficient is greater than the prese
- the determining module is specifically configured to: determine the coupling strength range according to the following formula:
- ⁇ represents the coupling strength coefficient
- ⁇ N represents the minimum eigenvalue
- ⁇ 1 represents the maximum eigen value
- ⁇ 1 represents the first constant
- ⁇ 2 represents the second constant
- ⁇ 1 ⁇ 2 ⁇ 0 ⁇ 1 ⁇ 2 ⁇ 0.
- the update module is specifically configured to select a node from the nodes of the network to add to the control node set, so as to update the control node set.
- the specific expression of the coupling strength matrix is:
- B represents the coupling strength matrix
- L represents the Laplacian matrix of the network
- D represents the feedback matrix of the network.
- the present invention provides an electronic device, including: at least one processor and a memory;
- the memory stores computer execution instructions
- At least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the control method involved in the first aspect and the optional solution.
- the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions.
- the processor executes the computer-executable instructions, the first aspect and the alternatives involved Control Method.
- the present invention provides a control method, device, electronic equipment, and storage medium.
- control nodes are selected from the network to form a control node set, and a network state equation that introduces the control nodes is constructed to obtain the network state equation Maximum and minimum eigenvalues, first and second constants.
- the obtained network coupling strength coefficient can more accurately reflect the coupling strength of the system.
- the coupling strength coefficient is used as a criterion to judge the control performance of the network, and the control node set with a larger coupling strength coefficient is selected to control the network to achieve a better control network.
- Fig. 1 is a schematic flowchart of a control method according to an exemplary embodiment of the present invention
- Figure 2(a) is a distribution diagram of the control node set obtained by using the existing control method
- Fig. 2(b) is a distribution diagram of the control node set obtained by using the control method provided by the embodiment shown in Fig. 1;
- Figure 2(c) is the distribution diagram of the control node set obtained by other control methods
- Figure 3(a) describes the function of network 1 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies
- Figure 3(b) describes the function of network 2 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies
- Figure 3(c) describes the function of network 3 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies
- Figure 3(d) describes the function of network 4 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies
- Fig. 4(a) shows the result of the criterion ⁇ proposed by the network 1 in the traction control point set and different maximum optimization strategies in the present invention
- Figure 4(b) shows the results of the criterion ⁇ proposed by the present invention in the network 2 in the traction control point set and different maximum optimization strategies;
- Figure 4(c) shows the results of the criterion ⁇ proposed by the present invention in the network 3 in the traction control point set and different maximum optimization strategies;
- Figure 4(d) shows the results of the criterion ⁇ proposed by the present invention in the network 4 in the traction control point set and different maximum optimization strategies;
- Figure 5 shows the ratio of the intersection of networks 1 to 4 in the union of traction control nodes
- Fig. 6 is a schematic structural diagram of a control device according to an exemplary embodiment of the present invention.
- Fig. 7 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present invention.
- the present invention provides a control method, device, electronic equipment and storage medium to solve the technical problem that the criterion adopted by the existing control method cannot accurately reflect the coupling strength of the control node, resulting in poor network control effect.
- Fig. 1 is a schematic flowchart of a control method according to an exemplary embodiment of the present invention. As shown in Figure 1, the control method provided in this embodiment includes the following steps:
- a network system it is composed of multiple nodes, and multiple nodes are randomly selected from the nodes of the network system as control nodes to form a control node set.
- the control node in the control node set is used to control the network to balance the network.
- ik represents the control node
- x ik represents the state of the control node ik
- f( ⁇ ) represents the nonlinear variation equation of the node itself
- c represents the coupling coefficient between nodes
- j represents the node
- 1 ⁇ j ⁇ N N is the number of network nodes
- l ij represents the network Lap
- H (h ij ) z ⁇ z
- h ij represents the coupling relationship between the states of adjacent nodes
- x j represents the state of the j-th node
- d represents the feedback coefficient
- Df is the Jacobian matrix of the function f
- ⁇ is the real variable
- the maximum Lyapunov exponent of the equation is a function of the real variable ⁇ .
- S103 Determine the coupling strength matrix according to the coupling item and the feedback item, and obtain the maximum eigenvalue and the minimum eigenvalue of the coupling strength matrix.
- B represents the coupling strength matrix
- L represents the Laplacian matrix of the network
- D represents the feedback matrix of the network.
- the feedback matrix of the network is composed of the feedback coefficient d, and D is N ⁇ N dimensions. For non-control nodes, the feedback coefficient Is 0.
- the obtained eigenvalues are: ⁇ 1 , ⁇ 2 ,..., ⁇ N , and 0> ⁇ 1 ⁇ 2 ⁇ ... ⁇ N.
- S104 Determine the first constant and the second constant according to the maximum characteristic value, the minimum characteristic value and the coupling coefficient.
- first constant and the second constant are obtained according to the following set of inequalities:
- ⁇ 1 represents the first constant
- ⁇ 2 represents the second constant
- ⁇ 1 and ⁇ 2 are real numbers.
- S105 Determine the coupling strength coefficient according to the maximum characteristic value, the minimum characteristic value, the first constant and the second constant.
- the coupling strength range is determined according to the following formula:
- S106 Determine whether the number of nodes in the control node set is less than the preset number; if the judgment result is all yes, then go to S107, if the judgment result is no, then go to S108;
- the upper limit of the number of control nodes is set according to the control complexity requirements.
- the standard coupling strength coefficient can be randomly set. If the number of nodes in the control node set is less than the preset number, the control node set can be optimized continuously, and the control node set can be optimized in the following manner:
- control node set is updated randomly, and the next cycle is entered.
- S108 Output the control node set corresponding to the standard coupling strength coefficient, so as to use the nodes in the control node set to control the network.
- the number of nodes in the control node set is greater than the preset number, stop optimizing the control node set, and output the control node set corresponding to the standard coupling strength coefficient.
- the coupling strength coefficient of the control node set is the largest currently found Coupling strength coefficient, using the node control network in the control node set, can achieve the best control effect.
- construct the network state equation that introduces the control node and obtain the maximum and minimum eigenvalues, the first and second constants according to the network state equation, and according to the maximum eigenvalue, the minimum eigenvalue, the first constant and the second constant
- the constant determines the coupling strength coefficient, and the obtained network coupling strength coefficient can more accurately reflect the coupling strength of the system.
- the control node set with a larger coupling strength coefficient is selected to control the network to achieve a better control network.
- S201 calculates the traditional ratio criterion R of the analog network and the proposed criterion ⁇ and v according to the network topology structure parameters such as the adjacency matrix A of the network.
- v is the largest eigenvalue of the matrix [Df+ ⁇ H T ],
- Figure 2(a) is a distribution diagram of the control node set obtained by using the existing control method.
- nodes 1, 4, and 7 represent the traction control set in the optimal case of R.
- Fig. 2(b) is a distribution diagram of a set of control nodes obtained by using the control method provided by the embodiment shown in Fig. 1.
- nodes 4, 7 and 8 represent the traction control set under the optimal condition of ⁇ .
- Figure 2(c) is a distribution diagram of the control node set obtained by using other existing control methods; as shown in Figure 2(c), nodes 2, 8 and 11 represent the optimal traction control set.
- Network 1 is ItalyPowerGrid
- network 2 is PDZBase
- network 3 is US Air
- network 4 is Neural.
- Figure 3(a) describes the function of network 1 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies.
- Figure 3(b) describes the function of network 2 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies.
- Figure 3(c) describes the function of network 3 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies.
- Figure 3(d) describes the function of network 4 in the traditional ratio criterion R, traction control point set and different maximum optimization strategies.
- R opt , ⁇ opt and v opt respectively represent the optimization of R, ⁇ and v, the traction node is set in The result of the traditional ratio criterion R.
- the ordinate represents the proportion of the pulling point of the point to all the nodes.
- R opt and ⁇ opt line chart trends are almost the same, which shows that ⁇ can describe the traction control ability the same as R. But in Figure 3(a) and Figure 3(b), there is a relatively small inconsistency between ⁇ and R.
- the optimal control nodes of ⁇ and v cannot reach the maximum value of R (where v is According to the calculation result of the optimal Lyapunov exponent, that is, the optimal control speed), that is to say, in the actual network, the actual applications of R, ⁇ and v are different.
- Figure 4(a) shows the results of the proposed criterion ⁇ of the network 1 in the traction control point set and different maximum optimization strategies.
- Figure 4(b) shows the results of the criterion ⁇ proposed in the network 2 of the invention in the traction control point set and different
- Figure 4(c) shows the results of the proposed criterion ⁇ in the traction control point set of the network 3 and different maximum optimization strategies.
- Figure 4(d) shows the network 4 the criterion ⁇ proposed by the invention in the traction The set of control points and the results of different maximum optimization strategies.
- Figure 4 (a), Figure 4 (b), Figure 4 (c) and Figure 4 (d) verify the different applications of ⁇ relative to R.
- Figure 5 shows the ratio of the intersection of networks 1 to 4 in the union of traction control nodes.
- the ordinate is the proportion of the traction control node union set by the intersection of two different standards.
- the blank column represents the intersection of the traditional ratio R and the newly proposed criterion ⁇ occupies the traction control node union
- the bar with oblique lines represents the ratio of the intersection of the traditional ratio R and the optimal control speed to the traction control node union
- the bar with black filling represents the intersection of the optimal control speed and the newly proposed criterion ⁇ .
- Control the ratio of node union.
- the intersection of the traditional ratio R and the newly proposed criterion ⁇ has the largest ratio compared to the others. In network 3 and network 4, it even reaches 0.7 or more, but there is a big difference between network 1 and network 2. This shows that the scenarios used by R and ⁇ are different, and this difference also causes the difference between Figure 4 and Figure 5.
- the new criterion ⁇ of the present invention does have a certain improvement over the traditional criterion R.
- the traditional criterion R cannot accurately describe the coupling strength range.
- the coupling strength determined by the new criterion ⁇ The range is larger than the traditional criterion R.
- control device is introduced below, and its implementation principle and technical effect are similar to the above-mentioned method principle and technical effect, and will not be repeated here.
- Fig. 6 is a schematic structural diagram of a control device according to an exemplary embodiment of the present invention.
- the control device 200 provided by the present invention includes: an obtaining module 201 for selecting control nodes from network nodes to obtain a control node set; a building module 202 for constructing a network state equation according to the control node set; Among them, the network state equation includes a coupling term and a feedback term, and the coupling term includes a node coupling coefficient; the determining module 203 is used to determine the coupling strength matrix according to the coupling term and the feedback term, and obtain the maximum eigenvalue and the minimum eigenvalue of the coupling strength matrix; The determining module 203 is also used to determine the first constant and the second constant according to the maximum characteristic value, the minimum characteristic value and the coupling coefficient; the determining module 203 is also used to determine the coupling according to the maximum characteristic value, the minimum characteristic value, the first constant and the second constant Strength coefficient; judging module 204, for judging whether the number of
- the determining module 203 is specifically configured to determine the coupling strength range according to the following formula:
- R represents the coupling strength coefficient
- ⁇ N represents the maximum eigenvalue
- ⁇ 1 represents the minimum eigen value
- ⁇ 1 represents the first constant
- ⁇ 2 represents the second constant
- ⁇ 1 ⁇ 2 ⁇ 0 ⁇ 1 ⁇ 2 ⁇ 0.
- the update module 206 is specifically configured to select a node from the nodes of the network to add to the control node set, so as to update the control node set.
- the specific expression of the coupling strength matrix is:
- A represents the coupling strength matrix
- L represents the Laplacian matrix of the network
- D represents the feedback matrix of the network.
- the electronic equipment provided in this application can be used to implement the above-mentioned aircraft prompting method, and its content and effects can be referred to the method part, which will not be repeated in this application.
- Fig. 7 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present invention.
- the electronic device 300 of this embodiment includes: a processor 301 and a memory 302, where
- the memory 302 is used to store computer execution instructions
- the processor 301 is configured to execute computer-executable instructions stored in the memory to implement each step executed by the receiving device in the foregoing embodiment. For details, refer to the related description in the foregoing method embodiment.
- the memory 302 may be independent or integrated with the processor 201.
- the electronic device 300 further includes a bus 303 for connecting the memory 302 and the processor 301.
- the embodiment of the present invention also provides a computer-readable storage medium, and the computer-readable storage medium stores computer-executable instructions.
- the processor executes the computer-executable instructions, the control method described above is implemented.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
一种控制方法、装置、电子设备以及存储介质,该方法包括:获得控制节点集合(S101);根据控制节点集合构建网络状态方程(S102);根据网络状态确定最大特征值、最小特征值,以确定耦合强度范围(S103、S104、S105);判断所述控制节点集合中节点数量是否小于预设数量(S106);若判断结果为是,则当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量(S107);若判断结果为否,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络(S108)。本方法所获得网络耦合强度系数能够更加准确地反应系统的耦合强度候选工作范围。
Description
本发明涉及复杂网络控制技术领域,尤其涉及一种控制方法、装置、电子设备以及存储介质。
在复杂网络中存在牵引控制现象,即有选择性的对网络中少部分节点施加控制而使得整个网络达到所期望的行为。
复杂网络牵引控制首先要考虑三个关键问题。第一个是选择哪些节点进行牵引控制;第二个是选择多少个节点进行牵引控制;第三个是选择什么牵引控制器。针对第一个问题和第二问题,现有方法通常在选择完控制节点后,计算判据R=λ
1/λ
N值,若判据R>α
2/α
1,判断判据是否大于选择其他控制节点下的判据R,若是,则选择数值更大的判据对应的控制节点作为最优控制节点,利用最优控制节点对网络进行牵引控制。
然而,判据R=λ
1/λ
N并不能精确反应控制节点的耦合强度,导致现有的控制方法对网络控制效果差的技术问题。
发明内容
本发明提供一种控制方法、装置、电子设备以及存储介质,以解决现有的控制方法所采用判据无法精确反应控制节点的耦合强度而导致对网络控制效果差的技术问题。
第一方面,本发明提供一种控制方法,包括:从网络节点中选取控制节点以获得控制节点集合;根据控制节点集合构建网络状态方程;其中,网络状态方程包括耦合项和反馈项,耦合项包括节点耦合系数;根据耦合项和反馈项确定耦合强度矩阵,并获得耦合强度矩阵的最大特征值和最小特征值;根据最大特征值、最小特征值以及耦合系数确定第一常量和第二常量;根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数;判断所述控制节点集合中节点数量是否小于预设数量;若判断结果为是,则当所述耦合强度系数大 于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量;若判断结果为否,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
在本发明提供的一种控制方法中,构建引入控制节点的网络状态方程,并根据网络状态方程获得最大和最小特征值、第一和第二常量,并根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数,所获得网络耦合强度系数能够更加准确地反应系统的耦合强度,选取耦合强度系数更大的控制节点集合对网络进行控制,实现更好的控制网络。
可选地,根据如下公式确定耦合强度范围:
其中,ω表示耦合强度系数,λ
N表示最大特征值,λ
1表示最小特征值,α
1表示第一常量,α
2表示第二常量,且α
1<α
2<0。
在本发明提供的一种控制方法中,将第一常量与最小特征值的比值同第二常量与最大特征值的比值相减,以获得耦合强度系数,使得耦合强度系数更加精确反应耦合强度,并以此作为网络系统可控性的判据,选择耦合强度系数更大的控制节点集合对网络进行控制,以达到更好的控制效果。
可选地,所述更新所述控制节点集合,具体包括:从网络的节点中选择一个节点新增至所述控制节点集合,以更新所述控制节点集合。
在本发明提供的一种控制方法中,通过在现有的控制节点集合基础之上增加一个节点,获得更新后的控制节点集合,可以避免收敛速度过快,使优化进入局部最优。
可选地,耦合强度矩阵的具体表达式为:
B=-L-D
其中,B表示耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络的反馈矩阵。
下面对控制装置进行介绍,其实现原理和技术效果与上述方法原理和技术效果类似,此处不再赘述。
第二方面,本发明提供一种控制装置,其特征在于,包括:获得模块,用于从网络节点中选取控制节点以获得控制节点集合;构建模块,用于根据控制 节点集合构建网络状态方程;其中,网络状态方程包括耦合项和反馈项,耦合项包括节点耦合系数;确定模块,用于根据耦合项和反馈项确定耦合强度矩阵,并获得耦合强度矩阵的最大特征值和最小特征值;确定模块还用于根据最大特征值、最小特征值以及耦合系数确定第一常量和第二常量;确定模块还用于根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数;判断模块,用于判断所述控制节点集合中节点数量是否小于预设数量;更新模块,用于若判断结果为是,则当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量;输出模块,用于若判断结果为否,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
可选地,确定模块具体用于:根据如下公式确定耦合强度范围:
其中,ω表示耦合强度系数,λ
N表示最小特征值,λ
1表示最大特征值,α
1表示第一常量,α
2表示第二常量,且α
1<α
2<0。
可选地,更新模块具体用于:从网络的节点中选择一个节点新增至所述控制节点集合,以更新所述控制节点集合。
可选地,耦合强度矩阵的具体表达式为:
B=-L-D
其中,B表示耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络的反馈矩阵。
第三方面,本发明提供一种电子设备,包括:至少一个处理器和存储器;
其中,存储器存储计算机执行指令;
至少一个处理器执行存储器存储的计算机执行指令,使得至少一个处理器执行第一方面以及可选方案涉及的控制方法。
第四方面,本发明提供一种计算机可读存储介质,其特征在于,计算机可读存储介质中存储有计算机执行指令,当处理器执行计算机执行指令时,实现第一方面以及可选方案涉及的控制方法。
本发明提供了一种控制方法、装置、电子设备以及存储介质,在控制方法中,从网络中选择控制节点以构成控制节点集合,并构建引入控制节点的网络 状态方程,以根据网络状态方程获得最大和最小特征值、第一和第二常量。并以此确定耦合强度系数,所获得网络耦合强度系数能够更加准确地反应系统的耦合强度。将耦合强度系数作为评判网络控制性能的判据,选取耦合强度系数更大的控制节点集合对网络进行控制,实现更好的控制网络。
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍,显而易见地,下面描述中的附图是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。
图1为本发明根据一示例性实施例示出的控制方法的流程示意图;
图2(a)为采用现有的控制方法得到的控制节点集合分布图;
图2(b)为采用图1所示实施例提供的控制方法得到的控制节点集合分布图;
图2(c)为采用其他控制方法得到的控制节点集合分布图;
图3(a)描述了网络1在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数;
图3(b)描述了网络2在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数;
图3(c)描述了网络3在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数;
图3(d)描述了网络4在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数;
图4(a)表示网络1本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果;
图4(b)表示网络2本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果;
图4(c)表示网络3本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果;
图4(d)表示网络4本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果;
图5表示在网络1至网络4中交集所占牵引控制节点并集的比例;
图6为本发明根据一示例性实施例示出的控制装置的结构示意图;
图7为本发明根据一示例性实施例示出的电子设备的结构示意图。
为使本发明实施例的目的、技术方案和优点更加清楚,下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
本发明提供一种控制方法、装置、电子设备以及存储介质,以解决现有的控制方法所采用判据无法精确反应控制节点的耦合强度而导致对网络控制效果差的技术问题。
图1为本发明根据一示例性实施例示出的控制方法的流程示意图。如图1所示,本实施例提供的控制方法包括如下步骤:
S101、从网络节点中选取控制节点以获得控制节点集合。
更具体地,对于网络系统,其由多个节点构成,从网络系统的节点中随机选取多个节点作为控制节点,并构成控制节点集合。控制节点集合中控制节点用于控制网络,以使网络达到平衡。
S102、根据控制节点集合构建网络状态方程。
更具体地,针对控制节点集合中控制节点,构建如下网络状态方程:
其中,ik表示控制节点,x
ik表示控制节点ik状态,
表示控制节点ik状态的微分,f(·)表示节点自身的非线性变化方程,c代表节点间耦合系数,j表示节点,1≤j≤N,N为网络节点数量,l
ij代表网络拉普拉斯矩阵的系数,H=(h
ij)
z×z,h
ij表示临近节点状态之间的耦合关系,x
j表示第j个节点的状态,d表示反馈系数,
表示静止状态,并且
服从于
当i=j时,l
ii=∑
j,j≠ia
ij,当i≠j时,l
ij=-a
ij。A=(a
ij)
N×N,E= {(i,j)|a
ij≠0},其中,A表示网络的邻接矩阵,a
ij表示两个节点间是否有边。当节点i和节点j存在边时,a
ij=1;否则a
ij=0。L=(l
ij)
N×N,L表示网络的拉普拉斯矩阵,拉普拉斯矩阵斜对角线上的元素就是节点i的度。
针对网络中其他节点,构建如下网络状态方程:
S103、根据耦合项和反馈项确定耦合强度矩阵,并获得耦合强度矩阵的最大特征值和最小特征值。
更具体地,耦合强度矩阵的具体表达式为:
B=-L-D
其中,B表示耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络的反馈矩阵,网络的反馈矩阵由反馈系数d构成,且D为N×N维,针对非控制节点,反馈系数为0。
利用线性代数方法计算B=-L-D的特征值,所获得的特征值为:λ
1,λ
2,……,λ
N,且有0>λ
1≥λ
2≥…≥λ
N。
S104、根据最大特征值、最小特征值以及耦合系数确定第一常量和第二常量。
更具体地,根据如下不等式组获取第一常量和第二常量:
α
1<α
2<0
α
1<cλ
N
cλ
1<α
2,
其中,α
1表示第一常量,α
2表示第二常量,且α
1与α
2都是实数。
S105、根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数。
更具体地,根据如下公式确定耦合强度范围:
S106、判断所述控制节点集合中节点数量是否小于预设数量;若判断结果均为是,则转入S107,若判断结果为否,则转入S108;
更具体地,根据控制复杂度要求设置控制节点数量的上限值。
S107、当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量。
更具体地,在执行实施例提供的方法前,可以随机设置标准耦合强度系数,若控制节点集合中节点数量小于预设数量时,可以继续优化控制节点集合,采用如下方式优化控制节点集合:
判断所述耦合强度系数是否大于预设的标准耦合强度系数,若判断结果为是,则利用所述耦合强度系数更新所述标准耦合强度系数,更新后的标准耦合强度系数对应的控制节点集合是目前找到最佳的控制节点集合。
在更新完标准耦合强度系数后,随机更新所述控制节点集合,进入下一轮循环。
S108、输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
更具体地,若控制节点集合中节点数量大于预设数量时,停止优化控制节点集合,并输出标准耦合强度系数对应的控制节点集合,该控制节点集合的耦合强度系数是目前寻找到的最大的耦合强度系数,利用控制节点集合中节点控制网络,可以达到最佳的控制效果。
在本实施例中,构建引入控制节点的网络状态方程,并根据网络状态方程获得最大和最小特征值、第一和第二常量,并根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数,所获得网络耦合强度系数能够更加准确地反应系统的耦合强度,选取耦合强度系数更大的控制节点集合对网络进行控制,实现更好的控制网络。
现分别采用基于图1所示实施例示出的控制方法以及其他两种控制方法获得控制节点集合,执行如下步骤:
S201根据网络的邻接矩阵A等网络拓扑结构参数,计算该模拟网络的传统 比率判据R以及提出的判据ω和v。其中,v为矩阵[Df+βH
T]的最大特征值,
S202分别在R,ω和v最优(最大)的情况下,得到不同的牵引控制节点集。
图2(a)为采用现有的控制方法得到的控制节点集合分布图。如图2(a)所示,节点1、4和7代表R最优情况下的牵引控制集。图2(b)为采用图1所示实施例提供的控制方法得到的控制节点集合分布图。如图2(b)所示,节点4、7和8代表ω最优情况下的牵引控制集。图2(c)为采用现有的其他控制方法得到的控制节点集合分布图;如图2(c)所示,节点2、8和11代表最优的牵引控制集。
如表1所示,在两个牵引控制节点集中,观察R,ω和v的值的变化情况。
表1 R,ω和v的值的变化情况表
| 最优R值 | 最优ω值 | 最优v值 | |
| R | 0.0693 | 0.0679 | 0.0595 |
| ω | 0.1232 | 0.1233 | 0.1231 |
| v | 0.0174 | 0.0126 | -0.0307 |
本发明还尝试在实际系统中进行数据模拟与测试,并取得了可观的效果。这里,我们选择了如下四个真实网络。网络1为ItalyPowerGrid、网络2为PDZBase、网络3为US Air和网络4为Neural。实验中我们将网络结构都看成是无权无向的。网络的结构属性包括网络规模(N)、边的数量(E)、度异质性(H=<k
2>/<k>
2)、度结合性(r)、平均聚类系数(<C>)、平均最短路径(<d>)以及稀疏度(Sparsity),如下表2所示:
表2 网络的结构属性
| N | E | H | r | <C> | <d> | 稀疏度 | |
| 网络1 | 67 | 93 | 1.158 | -0.036 | 0.022 | 6.70 | 3.0×10 -2 |
| 网络2 | 161 | 209 | 2.063 | -0.466 | 0.001 | 5.11 | 1.6×10 -2 |
| 网络3 | 332 | 2126 | 3.464 | -0.208 | 0.625 | 2.74 | 3.9×10 -2 |
| 网络4 | 297 | 2148 | 1.81 | -0.163 | 0.292 | 2.46 | 4.9×10 -2 |
图3(a)描述了网络1在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数。图3(b)描述了网络2在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数。图3(c)描述了网络3在传统比率判据R、牵 引控制点集以及不同的最大优化策略的函数。图3(d)描述了网络4在传统比率判据R、牵引控制点集以及不同的最大优化策略的函数。如图3(a)、图3(b)、图3(c)以及图3(d)所示,R
opt、ω
opt与v
opt分别代表R、ω和v最优化时,牵引节点集在传统比率判据R的结果。纵坐标代表点的牵引点占所有节点的比例。R
opt和ω
opt折线图趋势近乎一样,这表明ω是能够和R一样,描述牵引控制能力。但在图3(a)、图3(b)中,ω和R存在比较小的不一致,对于这四个真实网络,ω和v的最优控制节点不能到达R的最大值(其中,v为根据最优李雅普诺夫指数计算的结果,即最优控制速度),也就是说在实际网络中,R、ω和v的实际应用是不同的。
图4(a)表示网络1本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果,图4(b)表示网络2本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果,图4(c)表示网络3本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果,图4(d)表示网络4本发明提出判据ω在牵引控制点集以及不同的最大优化策略的结果。图4(a)、图4(b)、图4(c)以及图4(d)验证了ω相对于R不同应用情况。在图4(a)中,δ在0.18附近时,ω
opt>R
opt,并达到了R、ω和v的最大值(其中v为根据最优李雅普诺夫指数计算的结果,即最优控制速度),表明只优化单个客观指标是不能到达三个指标R、ω和v最优值的。在这种情况下,ω相较于R,合理的耦合强度表示的范围更大一些。
图5表示在网络1至网络4中交集所占牵引控制节点并集的比例。其中,纵坐标为两个不同标准所确定牵引点集合中,交集所占牵引控制节点并集的比例,其中,空白柱形表示传统比率R和新提出判据ω的交集占牵引控制节点并集的比例,具有斜线条的柱形表示传统比率R与最优控制速度的交集占牵引控制节点并集的比例,具有黑色填充的柱形表示最优控制速度与新提出判据ω的交集占牵引控制节点并集的比例。很明显,传统比率R和新提出判据ω的交集相较于其他,比例最大。在网络3和网络4中,甚至达到了0.7以上,但在网络1和网络2中差异较大。这表明R和ω所使用的场景是不同的,这种差异也同时造成了图4和图5的差异。
由此可见,本发明的新判据ω相对于传统判据R,确实有一定的改善,传统判据R不能精确描述耦合强度范围,在某些情况下,新判据ω所确定的耦合强度范围比传统判据R大。
下面对控制装置进行介绍,其实现原理和技术效果与上述方法原理和技术效果类似,此处不再赘述。
图6为本发明根据一示例性实施例示出的控制装置的结构示意图。如图6所示,本发明提供的控制装置200,包括:获得模块201,用于从网络节点中选取控制节点以获得控制节点集合;构建模块202,用于根据控制节点集合构建网络状态方程;其中,网络状态方程包括耦合项和反馈项,耦合项包括节点耦合系数;确定模块203,用于根据耦合项和反馈项确定耦合强度矩阵,并获得耦合强度矩阵的最大特征值和最小特征值;确定模块203还用于根据最大特征值、最小特征值以及耦合系数确定第一常量和第二常量;确定模块203还用于根据最大特征值、最小特征值、第一常量以及第二常量确定耦合强度系数;判断模块204,用于判断所述控制节点集合中节点数量是否小于预设数量;更新模块205,用于若判断结果为是,则当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量;输出模块206,用于若判断结果为为,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
可选地,确定模块203具体用于:根据如下公式确定耦合强度范围:
其中,R表示耦合强度系数,λ
N表示最大特征值,λ
1表示最小特征值,α
1表示第一常量,α
2表示第二常量,且α
1<α
2<0。
可选地,更新模块206具体用于:从网络的节点中选择一个节点新增至所述控制节点集合,以更新所述控制节点集合。
可选地,耦合强度矩阵的具体表达式为:
A=-L-D
其中,A表示耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络的反馈矩阵。
总之,本申请提供的电子设备可用于执行上述飞机提示方法,其内容和效果可参考方法部分,本申请对此不再赘述。
图7为本发明根据一示例性实施例示出的电子设备的结构示意图。如图7所示,本实施例的电子设备300包括:处理器301以及存储器302,其中,
存储器302,用于存储计算机执行指令;
处理器301,用于执行存储器存储的计算机执行指令,以实现上述实施例中接收设备所执行的各个步骤。具体可以参见前述方法实施例中的相关描述。
可选的,存储器302既可以是独立的,也可以跟处理器201集成在一起。
当存储器302独立设置时,该电子设备300还包括总线303,用于连接所述存储器302和处理器301。
本发明实施例还提供一种计算机可读存储介质,所述计算机可读存储介质中存储有计算机执行指令,当处理器执行所述计算机执行指令时,实现如上所述的控制方法。
最后应说明的是:以上各实施例仅用以说明本发明的技术方案,而非对其限制;尽管参照前述各实施例对本发明进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分或者全部技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本发明各实施例技术方案的范围。
Claims (10)
- 一种控制方法,其特征在于,包括:从网络节点中选取控制节点以获得控制节点集合;根据所述控制节点集合构建网络状态方程;其中,所述网络状态方程包括耦合项和反馈项,所述耦合项包括节点耦合系数;根据所述耦合项和所述反馈项确定耦合强度矩阵,并获得所述耦合强度矩阵的最大特征值和最小特征值;根据所述最大特征值、所述最小特征值以及所述耦合系数确定第一常量和第二常量;根据所述最大特征值、所述最小特征值、所述第一常量以及所述第二常量确定耦合强度系数;判断所述控制节点集合中节点数量是否小于预设数量;若判断结果为是,则当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量;若判断结果为否,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
- 根据权利要求2所述的方法,其特征在于,所述更新所述控制节点集合,具体包括:从网络的节点中选择一个节点新增至所述控制节点集合,以更新所述控制节点集合。
- 根据权利要求1至3任一项所述的方法,其特征在于,所述耦合强度矩阵的具体表达式为:B=-L-D其中,B表示所述耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络 的反馈矩阵。
- 一种控制装置,其特征在于,包括:获得模块,用于从网络节点中选取控制节点以获得控制节点集合;构建模块,用于根据所述控制节点集合构建网络状态方程;其中,所述网络状态方程包括耦合项和反馈项,所述耦合项包括节点耦合系数;确定模块,用于根据所述耦合项和所述反馈项确定耦合强度矩阵,并获得所述耦合强度矩阵的最大特征值和最小特征值;所述确定模块还用于根据所述最大特征值、所述最小特征值以及所述耦合系数确定第一常量和第二常量;所述确定模块还用于根据所述最大特征值、所述最小特征值、所述第一常量以及所述第二常量确定耦合强度系数;判断模块,用于判断所述控制节点集合中节点数量是否小于预设数量;更新模块,用于若判断结果为是,则当所述耦合强度系数大于预设的标准耦合强度系数时,利用所述耦合强度系数更新所述标准耦合强度系数,并更新所述控制节点集合,直至所述控制节点集合中节点数量等于预设数量;输出模块,用于若判断结果为否,则输出标准耦合强度系数对应的控制节点集合,以利用所述控制节点集合中节点控制网络。
- 根据权利要求6所述的装置,其特征在于,所述更新模块具体用于:从网络的节点中选择一个节点新增至所述控制节点集合,以更新所述控制节点集合。
- 根据权利要求5至7任一项所述的装置,其特征在于,所述耦合强度矩阵的具体表达式为:B=-L-D其中,B表示所述耦合强度矩阵,L表示网络的拉普拉斯矩阵,D表示网络的反馈矩阵。
- 一种电子设备,其特征在于,包括:至少一个处理器和存储器;其中,所述存储器存储计算机执行指令;所述至少一个处理器执行所述存储器存储的计算机执行指令,使得所述至少一个处理器执行如权利要求1至4任一项所述的控制方法。
- 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质中存储有计算机执行指令,当处理器执行所述计算机执行指令时,实现如权利要求1至4任一项所述的控制方法。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2019/080965 WO2020199114A1 (zh) | 2019-04-02 | 2019-04-02 | 控制方法、装置、电子设备以及存储介质 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2019/080965 WO2020199114A1 (zh) | 2019-04-02 | 2019-04-02 | 控制方法、装置、电子设备以及存储介质 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020199114A1 true WO2020199114A1 (zh) | 2020-10-08 |
Family
ID=72664428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2019/080965 Ceased WO2020199114A1 (zh) | 2019-04-02 | 2019-04-02 | 控制方法、装置、电子设备以及存储介质 |
Country Status (1)
| Country | Link |
|---|---|
| WO (1) | WO2020199114A1 (zh) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109039745A (zh) * | 2018-08-07 | 2018-12-18 | 东北大学 | 基于社区划分的复杂网络聚类同步的牵制节点选择方法 |
-
2019
- 2019-04-02 WO PCT/CN2019/080965 patent/WO2020199114A1/zh not_active Ceased
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109039745A (zh) * | 2018-08-07 | 2018-12-18 | 东北大学 | 基于社区划分的复杂网络聚类同步的牵制节点选择方法 |
Non-Patent Citations (4)
| Title |
|---|
| JALILI, M. ET AL.: "Optimal pinning controllability of complex networks: Dependence on network structure", PHYSICAL REVIEW E, vol. 91, no. 1, 31 January 2015 (2015-01-31), XP55740938, DOI: 20191220165042X * |
| ZHOU, M. Y. ET AL.: "Diffusion inspires selection of pinning nodes in pinning control", PHYSICA A, vol. 446, 28 November 2015 (2015-11-28), XP029377000, DOI: 20191220164119X * |
| 汪小帆 等 (WANG, XIAOFAN ET AL.): "复杂动态网络控制研究进展 (Recent Progress in Content of Complex Dynamical Networks)", 力学进展 (ADVANCES IN MECHANICS), vol. 38, no. 6, 25 November 2008 (2008-11-25), XP55740932, DOI: 20191220164934Y * |
| 陈关荣 (CHEN, GUANRONG): "复杂动态网络环境下控制理论遇到的问题与挑战 (Problems and Challenges in Control Theory under Complex Dynamical Network Environments)", 自动化学报 (ACTA AUTOMATICA SINICA), vol. 39, no. 4, 30 April 2013 (2013-04-30), XP55740935, DOI: 20191220165319Y * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10733332B2 (en) | Systems for solving general and user preference-based constrained multi-objective optimization problems | |
| Carlini et al. | A fully discrete semi-Lagrangian scheme for a first order mean field game problem | |
| CN115329973B (zh) | 仿真方法、装置、设备及存储介质 | |
| WO2024217241A1 (zh) | 声表面波器件仿真参数的快速拟合方法、系统及相关设备 | |
| CN109063283B (zh) | 一种刚-强度融合约束下的连续体结构可靠性拓扑优化方法 | |
| CN114943199A (zh) | 连通特征确定方法、装置、设备及存储介质 | |
| CN118691082B (zh) | 一种生产线设计变更的风险传播分析方法 | |
| CN108565852A (zh) | 一种三阶段递进的大电网电压稳定评估的故障筛选与排序方法 | |
| CN116614859A (zh) | 一种Ad Hoc网络中关键节点识别方法及系统 | |
| CN115329975A (zh) | 仿真方法、装置、设备及存储介质 | |
| CN115329530A (zh) | 一种基于自适应加点策略的结构混合渐变可靠性评估方法 | |
| Ambjorn et al. | Topology induced first-order phase transitions in lattice quantum gravity | |
| Mallon et al. | Neural level set topology optimization using unfitted finite elements | |
| CN106649939B (zh) | 基于传输线迭代的2d轴对称非线性静磁场模型的求解方法 | |
| CN114707116A (zh) | 一种基于代理模型的索网天线制造误差灵敏度分析方法 | |
| WO2020199114A1 (zh) | 控制方法、装置、电子设备以及存储介质 | |
| CN116434888A (zh) | 基于人工智能的双稳态复合材料结构振动频率优化方法 | |
| CN109951343B (zh) | 控制方法、装置、电子设备以及存储介质 | |
| CN112084702B (zh) | 一种低复杂度的光纤优化设计方法 | |
| CN108345728A (zh) | 一种考虑结构对称性的张拉整体结构非线性找形方法 | |
| CN110137967B (zh) | 一种针对关键节点的大规模电力系统潮流收敛性调整方法 | |
| CN114421483A (zh) | 一种解析式概率潮流计算方法、装置及存储介质 | |
| CN107609274B (zh) | 基于传输线与级别调度法的二维静磁场并行有限元方法 | |
| CN111555311B (zh) | 一种电力系统即插即用的稳定性分析与控制方法 | |
| CN113269449A (zh) | 天然气系统气源供气流量调度的节点级分散方法 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19922352 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 07.03.2022) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19922352 Country of ref document: EP Kind code of ref document: A1 |

