Complex network system maintenance decision-making method and device
By initializing the update of system parameters and environment entropy functions, adaptive control of large-scale complex network systems in uncertain environments is realized, the problem of lack of general adaptive control strategies in the existing technology is solved, and the adaptive reconstruction and stability of network topology is realized.
Patent Information
- Application Number
- CN202510515254.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-23
- Publication Date
- 2025-07-29
AI Technical Summary
The prior art is difficult to provide general adaptive control strategies for large-scale complex network systems in uncertain environments, especially in scenarios such as the Internet of Vehicles. The network microstates are robust and connected at high frequency evolution, and there is a lack of network topology reconstruction algorithms suitable for multiple types of targets.
By initializing system parameters, including estimation of macro variables, target landscape function and environmental entropy function, the acceptance probability is calculated and the environmental entropy function is updated, and adaptive network topology reconstruction is realized, relying only on the macro information of the system without understanding the micro state.
It provides an adaptive network topology reconstruction strategy that can gradually update the environment estimate in unknown environments, achieve the target landscape, and is suitable for network topology reconstruction of multiple targets, maintaining the robustness and connectivity of the network.
Smart Images
Figure CN120387586A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of network system maintenance decision-making, and particularly to a method and device for maintaining and making decisions for a complex network system. Background Art
[0002] Complex network systems such as communication networks and transportation networks are composed of a large number of interacting microscopic units and can be modeled and expressed by means of networks. In an uncertain environment, the environment of the network changes (for example, in the scenario of an Internet of Vehicles, the density of obstacles and the number of connectable nodes change), and these networks often need to continuously adjust their organizational methods to maintain certain functions. In order to make these complex network systems operate in the desired state, a series of complex network system maintenance decision-making methods have been proposed.
[0003] However, traditional complex network system control methods often assume that the dynamic model followed by the system is known, and thus a control strategy is derived from the dynamic model and the control objective. However, for a complex network system operating in an uncertain environment, the dynamic model followed by the system is often unknown. For a complex network system operating in an uncertain environment, compared with the current fixed control strategy, an adaptive strategy that enables the system to self-learn the nature of the environment is more reasonable. However, the current adaptive control method is developed from classical control theory, and the systems it focuses on are mainly small-scale systems often composed of a small number of components. In such small-scale systems, the number of microscopic states of the system is small, and the parameters characterizing the evolution dynamics of the microscopic states (microscopic states) of the system are few, making it easy to learn and adapt. However, for large-scale systems, they have many microscopic parameters, and it is often difficult to observe the complete microscopic state of the system, which leads to difficulties in designing an adaptive control strategy for large-scale complex network systems. The current adaptive control strategies for large-scale complex network systems mainly focus on objectives such as synchronization and often require the observability of each unit of the system.
[0004] For the problem of communication network topology maintenance, taking mobile ad-hoc communication networks such as vehicle-to-everything (V2X) as an example, a communication network will be established between vehicles and transportation infrastructure to improve traffic safety and efficiency. Due to the high-speed movement of vehicles and road switching, the communication links in the V2X network will need to be frequently broken and new connections established, resulting in high-frequency evolution of the microscopic state of the network. In order to ensure the safety and stability of the V2X network, it is necessary to explore the microscopic states that meet the requirements under constraints while the microscopic state of the network evolves at a high frequency, and maintain the macroscopic states such as the robustness and connectivity of the network at a stable level. In particular, when environmental parameters such as the vehicle movement pattern are unknown, there is an urgent need for the network maintenance decision algorithm to be adaptable and able to adapt to unknown environments. However, the robustness and invulnerability indicators of the network are often characterized by indicators such as the area of the percolation curve, the Fielder value, and the natural connectivity. There is a lack of a simple analytical relationship between these indicators and the network topology, resulting in a lack of a general adaptive algorithm for network topology reconstruction suitable for multiple types of objectives. Summary of the Invention
[0005] The objective of this application is to provide a method and device for maintaining decision-making in complex network systems, which can adaptively achieve the target landscape, and provide a general adaptive control strategy for network topology reconstruction suitable for multiple types of objectives; it allows users to use different types of emergent properties as targets, only need to utilize the macroscopic information of the system without understanding the microscopic state of the system, update the estimation of the environmental entropy function, and do not need to know the environmental properties in advance, and can gradually update the estimation of the environment during environmental interaction.
[0006] To achieve the above objective, this application provides the following solutions.
[0007] In the first aspect, this application provides a method for maintaining decision-making in complex network systems, including the following steps.
[0008] Initialize the system parameters of the target complex network system; the system parameters include macroscopic variables, target landscape function, estimation of the environmental entropy function, and adaptation rate; the target landscape function is used to express the preference of the target complex network system for macroscopic states; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system to explore different macroscopic states if it fully accepts environmental perturbations under the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function.
[0009] At each time step, observe the transfer of the macroscopic variables of the target complex network system from the old macroscopic state to the new macroscopic state driven by environmental perturbations, and calculate the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function.
[0010] Determine the transfer decision result of the macroscopic state based on the acceptance probability.
[0011] Update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transition of the macroscopic state at the next time step.
[0012] In a second aspect, the present application provides a decision-making device for maintaining a complex network system, including the following modules.
[0013] A system parameter initialization module for the target complex network system, configured to initialize the system parameters of the target complex network system; the system parameters include macroscopic variables, a target landscape function, an estimation of the environmental entropy function, and an adaptation rate; the target landscape function is used to express the preference of the target complex network system for macroscopic states; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system to explore different macroscopic states if it fully accepts environmental perturbations under the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function.
[0014] An acceptance probability calculation module, configured to observe the transition of the macroscopic variables of the target complex network system from an old macroscopic state to a new macroscopic state driven by environmental perturbations at each time step, and calculate the acceptance probability of the transition from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function.
[0015] A macroscopic state transition decision result determination module, configured to determine the macroscopic state transition decision result based on the acceptance probability.
[0016] An environmental entropy function estimation update module, configured to update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transition of the macroscopic state at the next time step.
[0017] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned complex network system maintenance decision-making method.
[0018] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned complex network system maintenance decision-making method is implemented.
[0019] According to the specific embodiments provided by the present application, the following technical effects are disclosed in the present application:
[0020] The present application provides a method and apparatus for maintaining and making decisions in a complex network system. By estimating the target landscape function and the environmental entropy function, the acceptance probability of the transition from the old macroscopic state to the new macroscopic state is calculated. Based on the acceptance probability, the decision result of the macroscopic state transition is determined to decide whether to accept the transition of the macroscopic state. A general adaptive control strategy suitable for network topology reconstruction of multiple types of targets is provided, which adaptively realizes the target landscape, allows users to use different types of emergent properties as targets, and only requires the macroscopic information of the system without knowing the microscopic state of the system, solving the problem of the lack of a general adaptive control strategy suitable for network topology reconstruction of multiple types of targets; by updating the estimation of the environmental entropy function, it is not necessary to know the environmental properties in advance, and the estimation of the environment can be gradually updated during the environmental interaction, having an adaptive characteristic. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.
[0022] Figure 1 It is an application environment diagram of a method for maintaining and making decisions in a complex network system according to an embodiment of the present application;
[0023] Figure 2 It is a flowchart showing the process of a method for maintaining and making decisions in a complex network system provided by an embodiment of the present application;
[0024] Figure 3 It is a flowchart showing the process of system deployment and operation provided by an embodiment of the present application;
[0025] Figure 4 It is a schematic diagram of the functional modules of a device for maintaining and making decisions in a complex network system provided by an embodiment of the present application;
[0026] Figure 5 It is a schematic diagram of the structure of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present application belong to the scope of protection of the present application.
[0028] Current communication network reconstruction algorithms such as LEACH are often specifically designed for specific objectives such as energy consumption control, and these topology maintenance decision algorithms lack generality. The robustness and invulnerability indicators of a network are often characterized by indicators such as the area of the percolation curve, the Fielder value, and the natural connectivity. There is a lack of a simple analytical relationship between these indicators and the network topology, making it difficult to specifically design an adaptive network reconstruction algorithm for this purpose. In addition, the network often faces complex constraint conditions, such as the geographical distance constraint of the connected edges (nodes pairs with too far geographical distance cannot establish connected edges), and the degree constraint (there is an upper limit on the number of connected edges that a node can establish). These constraints and objectives are variable and specific, making it difficult to apply the previous topology maintenance algorithms.
[0029] In view of the above problems, the present application provides a method and device for maintaining and making decisions for a complex network system.
[0030] To make the above objects, features, and advantages of the present application more obvious and understandable, the following further details the present application in conjunction with the accompanying drawings and specific embodiments.
[0031] The method for maintaining and making decisions for a complex network system provided by the embodiments of the present application can be applied to an application environment as Figure 1 shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set separately, integrated on the server 104, placed on the cloud or other servers. The terminal 102 can send a maintenance decision request to be processed to the server 104. After receiving the maintenance decision request to be processed, the server 104 initializes the system parameters of the target complex network system. At each time step, it observes the transfer of the macroscopic variables of the target complex network system from the old macroscopic state to the new macroscopic state driven by the environmental perturbation, and calculates the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state based on the estimation of the target landscape function and the environmental entropy function. Based on the acceptance probability, it determines the transfer decision result of the macroscopic state; updates the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function. The server 104 can feedback the obtained transfer decision result for the maintenance decision request to be processed to the terminal 102. In addition, in some embodiments, the method for maintaining and making decisions for a complex network system can also be implemented separately by the server 104 or the terminal 102. For example, the terminal 102 can directly process the maintenance decision request to be processed, or the server 104 can obtain the maintenance decision request to be processed from the data storage system and process the maintenance decision request to be processed.
[0032] Among them, the terminal 102 can be but is not limited to various desktop computers, laptop computers, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.
[0033] In an exemplary embodiment, as Figure 2 shown, a method for maintaining a decision of a complex network system is provided. This method is executed by a computer device, and specifically can be executed alone by a computer device such as a terminal or a server, or can be jointly executed by a terminal and a server. In the embodiment of the present application, taking this method applied to Figure 1 the server 104 in it as an example for illustration, it includes the following steps 201 to step 204.
[0034] Step 201, initialize the system parameters of the target complex network system; the system parameters include macroscopic variables, the target landscape function, the estimation of the environmental entropy function, and the adaptation rate; the target landscape function is used to express the preference of the target complex network system for the macroscopic state; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system exploring different macroscopic states if it completely accepts environmental perturbations in the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function.
[0035] Step 202, at each time step, observe the transfer of the macroscopic variables of the target complex network system driven by environmental perturbations from the old macroscopic state to the new macroscopic state, and calculate the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function.
[0036] Step 203, determine the transfer decision result of the macroscopic state based on the acceptance probability.
[0037] Step 204, update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transfer of the macroscopic state in the next time step.
[0038] In view of the current lack of a general adaptive control strategy for network topology reconstruction suitable for multiple types of targets and the lack of an adaptive control strategy for large-scale complex network systems such as communication networks, the above steps 201 to 204 are implemented, enabling the complex network system to achieve a given control target without environmental information and system microstate information. Statistical physics and thermodynamics have accumulated rich experience in analyzing large-scale systems. This application is proposed under the inspiration of the ideas of statistical physics and thermodynamics and mainly focuses on the macroscopic properties (i.e., emergent properties) of the system. In terms of target selection, this application selects the landscape function of the macroscopic variables of interest as the target, which corresponds to the free energy landscape describing the macroscopic properties of materials, proteins, etc., and is suitable for characterizing the macroscopic properties of complex network systems. In particular, for the connectivity and robustness of the network, a series of indicators such as the percolation curve area, Fielder value, and natural connectivity are proposed to characterize the connectivity and robustness of the network. In terms of learning and estimating the environment, this application proposes a simple rule to adaptively estimate the environmental entropy function, which describes the macroscopic state that is more likely to be biased under the free evolution of the system. In terms of the decision-making mechanism, this application gives how to selectively receive the current environmental perturbation based on the target landscape and the estimation of the environment. This application enables complex network systems such as communication networks to achieve the designed target landscape without environmental information and system microstate information.
[0039] (1) Using the landscape function of the system macroscopic variables as the system regulation target, where the system macroscopic variables can be adjusted according to the needs of the method application scenario, such as using indicators such as the percolation curve area, Fielder value, and natural connectivity to characterize the connectivity and robustness of the network topology.
[0040] The macroscopic variables are determined by the user's focus on the target complex network system; the macroscopic variables are the Fielder value, average clustering coefficient, or modularity; the Fielder value is the second smallest eigenvalue of the Laplace matrix, which is used to measure the difficulty of disassembling the complex network system into unconnected subclusters; the Laplace matrix is defined based on the adjacency matrix; the adjacency matrix represents the connection strength between nodes in the target complex network system. When the target complex network system is a vehicle-to-everything (V2X) communication system, each vehicle-mounted device, vehicle-mounted terminal, and communication device in the V2X communication system serves as a node.
[0041] The above step 201 includes the estimation of macroscopic variables, target landscape function, environmental entropy function, and the initialization process of the adaptation rate. The specific initialization process includes the following steps.
[0042] (1) Initialization of macroscopic variables: The user selects the macroscopic variable x of interest according to needs. For the maintenance scenario of a typical communication network, this application selects the second smallest eigenvalue of the Laplace matrix L as the macroscopic variable x, which is called the Fielder value and measures the difficulty of disassembling the target complex network system into unconnected sub-clusters.
[0043] The Laplace matrix L is defined based on the adjacency matrix A. The element A at the i-th row and j-th column of the adjacency matrix ij represents the connection strength between node i and node j. The element on the diagonal of the i-th row of the Laplace matrix is defined as L ii = ∑ j A ij . The off-diagonal elements of the Laplace matrix are defined as L ij = -A ij , where i ≠ j. The user can select other macroscopic variables that can be used to describe the complex network system according to needs. If the clustering characteristics of the complex network system are concerned, the average clustering coefficient can be selected as the macroscopic variable x of interest. If the modularity of the complex network system is concerned, the modularity can be selected as the macroscopic variable x of interest.
[0044] (2) Initialization of the target landscape function: The user sets the target landscape function U design (x) and stores it. According to this target landscape function, the probability that the system is in the macroscopic state x will be proportional to exp(-U design (x)). The target landscape function U design (x) expresses the preference of the complex network system for the macroscopic state x. For example, if the value of the target landscape function U design (x) at x = x0 is low, the complex network system will more frequently appear in the macroscopic state x = x0.
[0045] (3) Initialization of the estimation of the environmental entropy function U env (x) and store it. When the nature of the environment is unknown, it can be initialized as a constant function U env (x) = 0 (here is the identity sign). Here, if the macroscopic variable x is a continuous macroscopic variable, the macroscopic variable needs to be discretized and split into several intervals according to the required precision.
[0046] The estimation of the environmental entropy function U env (x) is used to characterize the difficulty of the system exploring different macroscopic states x if it completely accepts perturbations in the current environment. If the system evolves completely randomly, the macroscopic state x of the system will stabilize at U env(x) The macrostate that obtains the minimum value corresponds to the entropy maximization phenomenon of the second law of thermodynamics. The estimation of the environmental entropy function will change with the evolution of the environment. For example, in the vehicle networking scenario, changes in traffic flow density, density of obstacles, electromagnetic field interference intensity, etc. will all cause changes in the estimation U of the environmental entropy function of the macroscopic variable x of the network topology. env (x) Changes continuously, leading to the offset and fluctuation of the stable point of the system. And this application will stabilize the system macrostate at the target landscape U design (x) by selectively accepting random environmental perturbations.
[0047] Then in step 201, the initialization of the estimation of the environmental entropy function may include: when the environmental properties are unknown, initializing the estimation of the environmental entropy function as a constant function; when the macroscopic variable is a continuous variable, discretizing the macroscopic variable.
[0048] (4) Initialization of the adaptation rate f: The adaptation rate f is a constant greater than 0, which determines the update speed of the estimation U env (x). The smaller the adaptation rate f, the smoother the system runs, but the update speed of U env (x) will slow down, and the user needs to set f to an appropriate value according to the system operation situation.
[0049] (2) The decision-making scheme based on the estimation of the environmental entropy function and the target landscape function, that is, for each environmental perturbation suffered by the system, how to give the most appropriate acceptance probability according to the target and the current environmental estimation. The following gives the specific process.
[0050] After completing the above initialization steps, the system will selectively accept the environmental perturbations at each time step. The "environmental perturbations at each time step" here refers to that within one time step, under the drive of environmental noise, the system changes from the original macrostate (old macrostate) x = x i to the new macrostate x = x j . The "selectively accept" here means that the acceptance probability A(x i →x j ) accepts the transition from the old macrostate x = x i to the new macrostate x = x j . If not accepted, the actuator is called to keep the system in the old macrostate x = x i .
[0051] In this application, consider an application scenario of maintaining a complex network system to illustrate "environmental perturbations at each time step" and "selective acceptance". Consider a complex network system (such as a transportation network, a wireless communication network), and the maintenance of this complex network system faces the following decision-making problems: When an environmental factor causes an edge in the network to be removed, the complex network system can choose not to repair the edge (accept the state transition) or restore the edge (reject the state transition); when the complex network system has the opportunity to establish a new edge, the complex network system can choose to seize the opportunity to establish the edge (accept the state transition) or give up establishing the new edge (reject the state transition); when the complex network system proposes a plan to replace an edge with a new edge, the complex network system can choose to stick to the old edge (reject the state transition) or replace it with the new edge (accept the state transition). How to make a decision for each step of the state transition (such as removing an existing edge or establishing a new edge) starting from the macroscopic goal of the complex network system (represented by the objective landscape function U design (x)), that is, with what probability the complex network system should accept the state transition. Here, our goal is the objective landscape function U design (x), such that the steady-state distribution of a macroscopic variable x of interest in the complex network system (characterizing a certain macroscopic property of the network, which can be set as the clustering coefficient, the average shortest path length, etc.) satisfies p(x) ∝ exp(-U design (x)), that is, the complex network system operates near the minimum value of the objective landscape function U design (x), where p(x) is the steady-state distribution probability of the macroscopic variable. The landscape function here corresponds to the free energy landscape in materials, chemistry, and proteins, and describes the macroscopic properties of the complex network system.
[0052] To make a certain macroscopic property x of interest in the complex network system follow the objective landscape function U design (x), the complex network system needs to make the most appropriate decision for each step of the state transition (such as the establishment or destruction of an edge) (such as whether to accept the establishment or destruction of an edge). To achieve the goal in the operating environment, the decision must depend on the system goal (represented by the objective landscape function U design (x)) and the nature of the environmental perturbation (represented by the estimate of the environmental entropy function U env (x)). This application enables the complex network system to gradually update its estimate of the environment (represented by the estimate of the environmental entropy function U env (x)), so that for any state transition, based on the system's macroscopic state x i before the transition and the system's macroscopic state x j after the transition, an appropriate acceptance probability A(x i → x j ) is given, enabling the system to achieve the objective landscape function U in an unknown environment.design (x).
[0053] Specifically, the complex network system repeatedly performs the following operations at each time step \(t\rightarrow t + 1\):
[0054] Step 301: Use an observer to observe the new macroscopic state \(x = x\) to which the environmental noise drives the system j .
[0055] Step 302: Accept the environmental noise perturbation with probability \(A(x\) i \(\rightarrow x\) j ) and reject the environmental noise perturbation with probability \(1 - A(x\) i \(\rightarrow x\) j ). Here, the acceptance probability is given by the following formula.
[0056]
[0057] where \(A\) t (x i \(\rightarrow x\) j ) is the acceptance probability of the state transition from \(x\) at time \(t\) i to the state \(x\) j ; \(U\) design () is the target landscape function; is the estimate of the environmental entropy function at time \(t\).
[0058] Taking the case of network reconnection in the vehicle - to - everything (V2X) network as an example, when a vehicle discovers a new connectable link during movement, according to the system's macroscopic state \(x\) before the edge removal i and the system state \(x\) after the edge removal j , this application gives that the system should disconnect the original link and access the new link with probability \(A(x\) i \(\rightarrow x\) j ).
[0059] Step 303: Generate a random number \(u\) uniformly distributed in the interval \([0, 1]\). If \(u>A(x\) i \(\rightarrow x\) j ), the system rejects the environmental noise perturbation, and then the actuator of the system takes effect, keeping the system in the original state, that is, \(x(t + 1)=x_i\); if \(u < A(x\) i \(\rightarrow x\) j ), the system accepts the environmental noise perturbation, and then the state transition of the system is \(x(t + 1)=x\) j . Step 303 is to execute the decision given in Step 302, thereby making a state transition \(x\) i ->x jGive a response. For example, in the context of the Internet of Vehicles, when a new connectable link is discovered during vehicle movement, "rejecting environmental noise disturbance" means maintaining the original link, and "accepting environmental noise disturbance" means disconnecting the original link and accessing the new link.
[0060] Then, step 203 may include: generating a random number within the interval [0, 1]; determining the transfer decision result of the macroscopic state based on the random number and the acceptance probability, specifically including: if the random number is greater than the acceptance probability, the transfer decision result of the macroscopic state is that the target complex network system rejects the environmental disturbance, and the macroscopic state of the target complex network system at the next moment is the old macroscopic state; if the random number is not greater than the acceptance probability, the transfer decision result of the macroscopic state is that the target complex network system accepts the environmental disturbance, and the macroscopic state of the target complex network system at the next moment is the new macroscopic state.
[0061] Step 304: Update rule for environmental entropy function estimation: Update the system's estimation U env (x) of the environmental entropy function. The estimation function U env (x) at x = x(t + 1) is updated to the form shown below, and the update formula for the estimation of the environmental entropy function is as shown below.
[0062]
[0063] Where, is the updated estimation of the environmental entropy function at time t + 1, is the estimation of the environmental entropy function at time t, f is the adaptation rate, and x(t + 1) is the macroscopic state at time t + 1.
[0064] Figure 3 in means is reduced by f × exp[U design (x(t + 1)), compared with the estimation of the environmental entropy function at time t The estimation of the environmental entropy function at time t + 1 is reduced by f × exp[U design (x(t + 1)) at x = x(t + 1), but the estimated values at other positions remain unchanged, that is
[0065] The estimation U env (x) of the environmental entropy function remains unchanged at other positions. Here, the larger the adaptation rate f, the entropy function U env(x) The faster it is updated. The entropy function describes the characteristic properties of the environment, which will be used in the calculation of the acceptance probability of the system in step 302. Step 304 enables the system to continuously update the estimate of the environment based on the environmental feedback, so as to adaptively adjust the system behavior. For example, if the occurrence frequency of a certain state x = x0 is too high, according to the update formula in step 304, the value of U env (x) at x = x0 will drop faster, which means that U env (x0) is relatively small. At this time, according to the acceptance probability formula in step 302, the system will reject entering the macroscopic state x = x0 with a higher probability, and will escape from the macroscopic state x = x0 with a higher probability, thus reducing the occurrence frequency of x = x0. This estimation rule establishes a feedback between the environmental feedback (the occurrence frequency of each state) and the system behavior (the acceptance probability of state transition), ensuring that the system achieves the target landscape function.
[0066] Step 5: Update the time t = t + 1.
[0067] The above execution process is jointly implemented by the observer, actuator, and memory of the system. The functions of each subsystem are as follows: The observer of the system is responsible for giving the macroscopic state x of the system at each moment; the actuator of the system is responsible for performing rejection operations on the disturbances of the environment to keep the system state the same as the previous moment; the memory of the system is responsible for storing the target landscape function U design (x) and the estimated value U env (x) of the environmental entropy. This scheme can be used for the maintenance decision of complex systems, and the flowchart of the operation is as Figure 3 shown: This complex network system includes four nodes A, B, C, and D. When the environment is about to remove the link BD in the complex network system, the system needs to make a decision on whether to repair the link BD. This application does not need to know the microstates before and after the environmental disturbance (that is, which nodes exactly had links before, represented by σ Figure 3 in i , σ j represents the microstate before knowing the environmental disturbance, and σ i represents the microstate after knowing the environmental disturbance), but only needs to obtain the value (old macroscopic state) x j of the macroscopic variable x of interest before the environmental disturbance and the value (new macroscopic state) x i after the disturbance, calculate the probability A(x j → x i → x j ) of accepting this state transition based on this, and then use it to decide to accept this state transition (that is, do not repair the link BD, then x(t + 1) = x j ) or reject (repair the link BD, then x(t + 1) = x i ), and correspondingly reduce the estimated value U env of the environmental entropy after completing the action.(x) value at x = x(t + 1) Figure 3 The shaded part of the lower left bar chart in Figure 3 represents the estimated value U of the environmental entropy. env (x) decrease at x = x(t + 1)).
[0068] The present application also provides an application scenario, which applies the above-mentioned complex network system maintenance decision method. Specifically: The complex network system maintenance decision method provided in this embodiment can be applied to the maintenance decision scenario of the vehicle networking system. The maintenance decision scenario of the vehicle networking system includes a request receiving link and a system maintenance decision link; the maintenance decision request to be processed enters the system maintenance decision link from the request receiving link, obtains the corresponding maintenance decision result, and enters the downstream content distribution link. The complex network system maintenance decision method provided in this embodiment belongs to the machine tagging link in the system maintenance decision link. Specifically, in the process of the system maintenance decision link for the maintenance decision request to be processed, the system parameters of the target complex network system can be initialized. At each time step, observe the transfer of the macroscopic variables of the target complex network system driven by the environmental disturbance from the old macroscopic state to the new macroscopic state, and calculate the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state based on the estimation of the target landscape function and the environmental entropy function. Determine the transfer decision result of the macroscopic state based on the acceptance probability; update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function.
[0069] The present application has the following advantages:
[0070] 1) In terms of goal setting, the present application allows users to take different types of emergent properties as goals. Existing complex network system control methods often focus on system synchronization as the goal. The existing topology maintenance decision algorithms for communication networks are specifically set for a few indicators such as energy consumption, lacking generality in goal setting. The advantage of flexible goal setting in the present application comes from step 201. On the one hand, the macroscopic variable x of interest in the method can be flexibly set according to the scenario and user needs. For example, if the clustering characteristics of the complex network system are concerned, the average clustering coefficient can be selected as the macroscopic variable x of interest. If the modularity of the complex network system is concerned, the modularity can be selected as the macroscopic variable x of interest. On the other hand, we choose the landscape function as the goal, which is more suitable for describing the complex network system.
[0071] 2) In terms of the information used, the present application only needs to use the macroscopic information of the system without knowing the microscopic state of the system. This advantage comes from the fact that each execution step of the present application only requires macroscopic state information (such as the clustering coefficient of the network), and does not require the complete microscopic state information of the system (such as the complete connection relationship of the network).
[0072] 3) In terms of adapting to unknown environments, the present application has an adaptive characteristic. Compared with related technologies, the present application does not need to know the nature of the environment in advance and can gradually update the estimation of the environment during environmental interaction, thereby adjusting its own behavior accordingly. This advantage enables the system to autonomously adapt to unknown environments and gradually changing environments. This advantage stems from step 204. According to the update rule of the environmental entropy function estimation given by the method proposed in the present application, the system can update the estimation of the environment in real time based on environmental feedback, so that the system can gradually learn the nature of the environment and enable the system decision-making to achieve the target landscape in this environment.
[0073] Based on the same inventive concept, the embodiments of the present application also provide a complex network system maintenance decision-making device for implementing the complex network system maintenance decision-making method involved above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the complex network system maintenance decision-making device provided below can refer to the limitations on the complex network system maintenance decision-making method in the above text, and will not be elaborated here.
[0074] In an exemplary embodiment, as Figure 4 shown, a complex network system maintenance decision-making device is provided, which includes the following modules.
[0075] The system parameter initialization module T1 of the target complex network system is used to initialize the system parameters of the target complex network system; the system parameters include macroscopic variables, the target landscape function, the estimation of the environmental entropy function, and the adaptation rate; the target landscape function is used to express the preference of the target complex network system for the macroscopic state; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system to explore different macroscopic states if it completely accepts environmental perturbations under the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function.
[0076] The acceptance probability calculation module T2 is used to observe the transfer of the macroscopic variables of the target complex network system from the old macroscopic state to the new macroscopic state driven by environmental perturbations at each time step, and calculate the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function.
[0077] The transfer decision result determination module T3 of the macroscopic state is used to determine the transfer decision result of the macroscopic state based on the acceptance probability.
[0078] The estimation update module T4 of the environmental entropy function is used to update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transfer of the macroscopic state in the next time step.
[0079] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as shown in Figure 5 . The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store system maintenance decision processing data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for maintaining decisions in a complex network system.
[0080] Those skilled in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component layout.
[0081] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, it implements the steps in the above method embodiments.
[0082] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, it implements the steps in the above method embodiments.
[0083] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.
[0084] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random-access memories (ReRAMs), magnetoresistive random-access memories (MRAMs), ferroelectric random-access memories (FRAMs), phase change memories (PCMs), graphene memories, etc. Volatile memories can include random-access memories (RAMs) or external caches, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random-access memory (SRAM) or dynamic random-access memory (DRAM), etc.
[0085] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.
[0086] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.
[0087] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.
Claims
1. A method for maintaining decision-making of a complex network system, characterized in that, The complex network system maintenance decision-making method includes: Initializing the system parameters of the target complex network system; the system parameters include macroscopic variables, the target landscape function, the estimation of the environmental entropy function, and the adaptation rate; the target landscape function is used to express the preference of the target complex network system for the macroscopic state; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system exploring different macroscopic states if it completely accepts environmental perturbations under the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function. At each time step, observe the transfer of the macroscopic variables of the target complex network system driven by environmental perturbations from the old macroscopic state to the new macroscopic state, and calculate the acceptance probability of the transfer from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function. Determine the transfer decision result of the macroscopic state based on the acceptance probability. Update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transfer of the macroscopic state at the next time step.
2. The maintenance decision-making method for a complex network system according to claim 1, wherein The macroscopic variable is determined by the user's focus on the target complex network system; the macroscopic variable is the Fielder value, the average clustering coefficient, or the modularity; the Fielder value is the second smallest eigenvalue of the Laplace matrix and is used to measure the difficulty of disassembling the complex network system into unconnected sub-clusters. The Laplace matrix is defined based on the adjacency matrix. The adjacency matrix represents the connection strength between nodes in the target complex network system.
3. The method for making a maintenance decision for a complex network system according to claim 1, characterized in that The initialization of the estimation of the environmental entropy function includes: When the environmental property is unknown, initialize the estimation of the environmental entropy function as a constant function. When the macroscopic variable is a continuous variable, discretize the macroscopic variable.
4. The maintenance decision-making method for a complex network system according to claim 1, characterized in that The calculation formula of the acceptance probability is as follows: Among them, A t (x i →x j ) is the acceptance probability of the transition from the state x i to the state x j ; U design () is the objective landscape function; is the estimate of the environmental entropy function at time t.
5. The maintenance decision-making method for a complex network system according to claim 1, wherein Determining the transfer decision result of the macroscopic state based on the acceptance probability specifically includes: Generate a random number in the interval [0, 1]. Determine the transfer decision result of the macroscopic state according to the random number and the acceptance probability.
6. The maintenance decision-making method for a complex network system according to claim 5, characterized in that The determination of the transfer decision result of the macroscopic state according to the random number and the acceptance probability specifically includes: If the random number is greater than the acceptance probability, the transfer decision result of the macroscopic state is: the target complex network system rejects the environmental perturbation, and the macroscopic state of the target complex network system at the next moment is the old macroscopic state. If the random number is not greater than the acceptance probability, the transfer decision result of the macroscopic state is: the target complex network system accepts the environmental perturbation, and the macroscopic state of the target complex network system at the next moment is the new macroscopic state.
7. The maintenance decision-making method for a complex network system according to claim 1, wherein The update formula of the estimation of the environmental entropy function is as follows: Among them, is the estimate of the updated environmental entropy function at time t + 1, is the estimate of the environmental entropy function at time t, f is the adaptation rate, and x(t + 1) is the macroscopic state at time t + 1.
8. A maintenance decision-making device for a complex network system, characterized in that, The complex network system maintenance decision-making device includes: The system parameter initialization module of the target complex network system is used to initialize the system parameters of the target complex network system; the system parameters include macroscopic variables, the target landscape function, the estimation of the environmental entropy function, and the adaptation rate; the target landscape function is used to express the preference of the target complex network system for the macroscopic state; the estimation of the environmental entropy function is used to characterize the difficulty of the target complex network system to explore different macroscopic states if it fully accepts environmental perturbations under the current environment; the adaptation rate is used to determine the update speed of the estimation of the environmental entropy function. The acceptance probability calculation module is used to observe the transition of the macroscopic variables of the target complex network system driven by environmental perturbations from the old macroscopic state to the new macroscopic state at each time step, and calculate the acceptance probability of the transition from the old macroscopic state to the new macroscopic state according to the target landscape function and the estimation of the environmental entropy function. The transition decision result determination module of the macroscopic state is used to determine the transition decision result of the macroscopic state based on the acceptance probability. The estimation update module of the environmental entropy function is used to update the estimation of the environmental entropy function according to the adaptation rate to obtain the updated estimation of the environmental entropy function; the updated estimation of the environmental entropy function is used for the transition of the macroscopic state at the next time step.
9. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the complex network system maintenance decision method according to any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the complex network system maintenance decision method according to any one of claims 1-7.