Information flow positive feedback-based dominant path identification method and system in single-source scene

Abstract

The invention discloses an information flow positive feedback-based dominant path identification method and system in a single-source scene, and relates to the technical field of computers. Physical mechanism modeling is carried out through dynamic evolution of electric network equivalence, Kirchhoff's law and positive feedback drive, and the model has clear dynamic explanation on the basis of physical mechanisms such as flow conservation and dynamic evolution of pipeline bearing capacity. Specifically, aiming at network information propagation, a dominant path intelligent identification model is established based on an information flow positive feedback self-ablation mechanism from a mechanism level. According to the method, analysis of an information flow penetration rule is taken as an entry point, and a dominant path during information interaction between a source node and a target node in a target network is quickly and accurately identified in an information flow identification mode. The method has a core support effect on construction of a propagation risk early warning model, implementation of a key node intervention strategy and optimization of an intelligent governance mechanism of a social information system.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and particularly to a method and system for identifying a dominant path in a single-source scenario based on positive feedback of information flow. Background Art

[0002] Identifying a dominant path means screening out the paths that mainly carry information flow during the directional propagation of network information flow. Dominant path identification methods are mainly applied to fields such as precision marketing in social networks and optimization of network structure data.

[0003] The general idea of existing dominant path identification methods is to calculate certain metrics to obtain the importance of links, and then use the method of local optimal addressing to find target nodes. However, these methods have defects in aspects such as insufficient mechanism modeling, neglect of weak connections, and complex interaction modeling.

[0004] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for identifying a dominant path in a single-source scenario based on positive feedback of information flow, aiming to solve the problem that the existing dominant path identification method calculates the importance of links through metrics and identifies the dominant path through local optimal addressing, resulting in insufficient mechanism modeling.

[0006] The technical solution adopted by the present invention to solve the problem is as follows: In a first aspect, an embodiment of the present invention provides a method for identifying a dominant path in a single-source scenario based on positive feedback of information flow, the method comprising: Determine a pair of source nodes and target nodes for information interaction in the target network, and input the network structure data of the target network into a dominant path intelligent recognition model constructed based on a self-ablation mechanism of positive feedback of information flow; Determine the dominant path when the source node and the target node perform information interaction through the dominant path intelligent recognition model; The dominant path intelligent recognition model is used for: According to the network structure data, equivalent the links for information transmission in the target network to the pipes carrying traffic during information transmission in an electrical network, and the traffic flowing through each node in the target network conforms to the law of conservation, and the traffic flowing through each link conforms to Kirchhoff's law; Iteratively update the traffic distribution and traffic carrying capacity of the pipes between the source node and the target node through an evolutionary algorithm until the evolution reaches a steady state, and determine the dominant path according to the traffic distribution results or traffic carrying capacity of each pipe after the evolution reaches the steady state.

[0007] In one embodiment, the type of the target network includes a social network.

[0008] In one embodiment, the traffic carrying capacity of each pipeline is a non-fixed value that varies over time, and the traffic carrying capacity of the pipeline and the traffic in the pipeline change based on a positive feedback mechanism.

[0009] In one embodiment, the evolutionary algorithm is used for: In the first step, based on the traffic carrying capacities of the pipelines between the source node and the target node, trigger the traffic allocation for the current round; In the second step, according to the traffic allocation result of the current round and the pre-established differential equation, adjust the traffic carrying capacities of the pipelines between the source node and the target node; wherein, the differential equation is generated based on the dynamic evolution model established by the positive feedback mechanism, and the differential equation is used to describe the dynamic evolution law of the traffic carrying capacity of the pipeline changing with the traffic. In the third step, determine whether the evolution has reached a steady state. If it has not reached a steady state, continue to execute the first step until the evolution reaches a steady state and stop the iteration.

[0010] In one embodiment, the differential equation includes a first parameter and a second parameter; the first parameter is used to adjust the steady state value of the final pipeline convergence; the second parameter is used to adjust the pipeline convergence speed.

[0011] In one embodiment, according to the traffic allocation result and the pre-established differential equation, adjusting the traffic carrying capacity of each pipeline between the source node and the target node includes: For each pipeline between the source node and the target node, if the calculation result of the differential equation corresponding to the pipeline is negative, indicating that the traffic in the pipeline is less than the maximum traffic carrying capacity, then reduce the traffic carrying capacity of the pipeline; If the calculation result of the differential equation corresponding to the pipeline is positive, indicating that the traffic in the pipeline is greater than the maximum traffic carrying capacity, then increase the traffic carrying capacity of the pipeline; If the traffic carrying capacity of the pipeline after adjustment is zero, but the evolution has not reached a steady state, retain the pipeline, but suspend the transmission activity of the pipeline until the traffic carrying capacity of the pipeline increases during the subsequent evolution process, and then resume the transmission activity of the pipeline.

[0012] In one embodiment, determining the dominant path according to the traffic allocation result or the traffic carrying capacity of each pipeline after the evolution reaches a steady state includes: According to the traffic allocation result or the traffic carrying capacity of each pipeline after the evolution reaches a steady state, and a preset threshold, screen out a number of target links, and make the path composed of all the target links a connected graph; Determine the dominant path according to all the target links.

[0013] In a second aspect, an embodiment of the present invention further provides a dominant path recognition system in a single-source scenario based on positive feedback of information flow. The system includes: A data input module, configured to determine a pair of source nodes and target nodes for information interaction in a target network, and input network structure data of the target network into a dominant path intelligent recognition model constructed based on an information flow positive feedback self-ablating mechanism; The dominant path intelligent recognition model is configured to, according to the network structure data, equivalent the links for information transmission in the target network to pipes carrying traffic during information transmission in an electrical network, and the traffic flowing through each node in the target network conforms to the conservation law, and the traffic flowing through each link conforms to Kirchhoff's law; Iteratively update the traffic distribution and traffic carrying capacity of the pipes between the source node and the target node through an evolutionary algorithm until the evolution reaches a steady state; determine the dominant path when the source node and the target node perform information interaction according to the traffic distribution result or traffic carrying capacity of each pipe after the evolution reaches the steady state.

[0014] In a third aspect, an embodiment of the present invention further provides a terminal, characterized in that the terminal includes a memory and more than one processor; the memory stores more than one program; the program includes instructions for executing the dominant path recognition method in a single-source scenario based on positive feedback of information flow as described in any one of the above; the processor is configured to execute the program.

[0015] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which multiple instructions are stored, characterized in that the instructions are suitable for being loaded and executed by a processor to implement the steps of the dominant path recognition method in a single-source scenario based on positive feedback of information flow as described in any one of the above.

[0016] Advantages of the present invention: The embodiments of the present invention perform physical mechanism modeling through electrical network equivalence, Kirchhoff's law, and positive feedback-driven dynamic evolution. Based on physical mechanisms such as traffic conservation and dynamic evolution of pipe carrying capacity, the model has a clear dynamic explanation. It solves the problem of insufficient mechanism modeling existing in the existing dominant path recognition methods. Specifically, the present invention establishes a dominant path intelligent recognition model based on an information flow positive feedback self-ablating mechanism from the mechanism level for network information propagation. Taking the analysis of the penetration law of information flow as the entry point, through the way of information flow recognition, the dominant path when the source node and the target node in the target network perform information interaction is quickly and accurately recognized. It has a core supporting role in constructing a propagation risk warning model, implementing key node intervention strategies, and optimizing the intelligent governance mechanism of the social information system. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a schematic flowchart of a method for identifying a dominant path in a single-source scenario based on positive feedback of information flow provided by an embodiment of the present invention.

[0019] Figure 2 It is a schematic block diagram of a system for identifying a dominant path in a single-source scenario based on positive feedback of information flow provided by an embodiment of the present invention.

[0020] Figure 3 It is a schematic block diagram of a terminal provided by an embodiment of the present invention. Detailed Embodiments

[0021] The present invention discloses a method and a system for identifying a dominant path in a single-source scenario based on positive feedback of information flow. To make the objectives, technical solutions and effects of the present invention clearer and more definite, the following further describes the present invention in detail with reference to the accompanying drawings and by way of examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0022] Those skilled in the art of the present technology can understand that unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the description of the present invention means the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0023] Those skilled in the art can understand that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this invention belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless specifically defined as here.

[0024] In view of the above-mentioned defects of the prior art, the present invention provides a method for identifying a dominant path in a single-source scenario based on positive feedback of information flow. The method includes: determining a pair of source nodes and target nodes that perform information interaction in a target network, and inputting the network structure data of the target network into a dominant path intelligent recognition model constructed based on the self-ablation mechanism of positive feedback of information flow; determining the dominant path when the source node and the target node perform information interaction through the dominant path intelligent recognition model; the dominant path intelligent recognition model is used for: equivalenting the target network to an electrical network according to the network structure data; wherein, the nodes in the target network correspond to the nodes in the electrical network, the links for information transmission in the target network correspond to the pipes carrying traffic when information is transmitted in the electrical network, and each node conforms to Kirchhoff's law; iteratively updating the traffic distribution and traffic carrying capacity of the pipes between the source node and the target node through an evolutionary algorithm until the evolution reaches a steady state, and determining the dominant path according to the traffic distribution results or traffic carrying capacity of each pipe after the evolution reaches the steady state. The present invention conducts physical mechanism modeling through electrical network equivalence, Kirchhoff's law, and positive feedback-driven dynamic evolution, and is based on physical mechanisms such as traffic conservation and dynamic evolution of pipe carrying capacity, making the model have a clear dynamic interpretation. It solves the problem of insufficient mechanism modeling existing in the existing dominant path identification methods. The present invention establishes a dominant path intelligent recognition model based on the self-ablation mechanism of positive feedback of information flow from the mechanism level for network information propagation. Taking the analysis of the penetration law of information flow as the entry point, the dominant path when a pair of source nodes and target nodes in a target network perform information interaction is quickly and accurately identified through information flow recognition. In practical application scenarios, the present invention can provide a new solution idea for finding dominant paths in various networks, and has a core supporting role in constructing a propagation risk warning model based on network science, implementing key node intervention strategies, and optimizing the intelligent governance mechanism of social information systems.

[0025] As Figure 1 shown, the method specifically includes the following steps: Step S100, determine a pair of source nodes and target nodes that perform information interaction in a target network, and input the network structure data of the target network into a dominant path intelligent recognition model constructed based on the self-ablation mechanism of positive feedback of information flow.

[0026] Specifically, the target network can be any network that needs to identify the dominant path, such as social networks, the Internet of Things, computer networks, railway networks, transportation networks, and so on. First, determine the indexes of the source node and the target node pair for information interaction in the target network, that is, obtain the source node and the target node. The dominant path is composed of the links that carry most of the information flow during information interaction between the source node and the target node. In order to quickly and accurately identify the dominant path, in this embodiment, the network structure data of the target network needs to be input into a pre-constructed intelligent dominant path recognition model. In actual application scenarios, the network structure data can also be called graph structure data, which can be generated by code or obtained by reading the actual network structure data. The network structure data can be stored in a computer using an adjacency matrix or an adjacency list. The information flow penetration has a pattern: for the information transmission between the source node and the target node pair, a small number of reachable paths carry most of the information flow, while the paths formed by most of the edges only transport a small part of the information flow, showing significant dominant path behavior. The intelligent dominant path recognition model is constructed using an information flow positive feedback self-ablation mechanism, and the information flow positive feedback self-ablation mechanism includes an information flow positive feedback mechanism and a self-ablation mechanism. The positive feedback mechanism is reflected in that the traffic carrying capacity of the link is proportional to the traffic in the link. When the traffic is large, the traffic carrying capacity is large; when the traffic is small, the traffic carrying capacity is small. The self-ablation mechanism is an incidental phenomenon in the implementation process of the positive feedback mechanism, which is reflected in that when evolving according to the positive feedback mechanism, there will be a phenomenon that the traffic carrying capacity of some links gradually becomes zero, and this phenomenon is called self-ablation.

[0027] Step S200: Determine the dominant path when the source node and the target node perform information interaction through the intelligent dominant path recognition model; The intelligent dominant path recognition model is used to: according to the network structure data, equivalent the links for information transmission in the target network to the pipes carrying traffic during information transmission in an electrical network, and the traffic flowing through each node in the target network conforms to the conservation law, and the traffic flowing through each link conforms to Kirchhoff's law; through an evolutionary algorithm, iteratively update the traffic distribution and traffic carrying capacity of the pipes between the source node and the target node until the evolution reaches a steady state, and determine the dominant path according to the traffic distribution results or traffic carrying capacity of each pipe after the evolution reaches the steady state.

[0028] Specifically, in this embodiment, the intelligent dominant path recognition model quickly and accurately determines the dominant path through which the traffic flows during the information transfer between the source node and the target node. The functions of the intelligent dominant path recognition model mainly include two parts. The first part is the equivalence of network structure data, and the second part is the evolutionary algorithm.

[0029] For the network structure data equivalence of the first part, it is necessary to introduce pipes to replace the links (also known as edges) in the target network. The thickness of the pipes is used to describe the traffic-carrying capacity in the pipes, and it can change adaptively according to the traffic in the pipes. At the same time, combined with Kirchhoff's law, the target network can be equivalent to an electrical network (i.e., a circuit network or an electrical network): taking the source node as the input node and the target node as the output node, injecting a constant-amplitude traffic or current (such as a traffic with a constant amplitude of 1) into the source node, then the entire electrical network will be powered on, and the traffic will flow in from the source node and flow out from the target node.

[0030] For the evolutionary algorithm of the second part, through the positive feedback mechanism of the traffic-carrying capacity of the pipes adapting to the traffic, the traffic-carrying capacity of the pipes is updated, thereby triggering a new round of traffic distribution evolution. After evolving to the steady state, the dominant path in the information flow propagation process is identified according to the distribution of the traffic or the traffic-carrying capacity. For example, the links with large traffic in the target network may be identified as the dominant paths.

[0031] For example, first, the definition of the information flow is explained, and the target network is determined. , represents the number of nodes included in the target network, represents the number of links included in the target network. A link is a connection between two nodes. Inject a unit traffic at the source node , and then converge and flow out at the target node . Considering all possible path schemes for the traffic from the source node to the target node, among them, except for the source node and the target node, the other nodes passed by the path satisfy that the inflow traffic is equal to the outflow traffic.

[0032] Classify the links in the target network as pipes that can freely expand and contract, which are used to carry the traffic in the information propagation process. First, initialize the traffic-carrying capacity of all pipes to 1. Since the amplitude of the traffic injected from the source node is also 1, the maximum traffic-carrying capacity of the pipes will not be greater than 1.

[0033] To quantify the traffic components in each link, Ohm's law in the electrical network is introduced, that is, the traffic between nodes is caused by the power difference between nodes, and we can get: ; In the formula: is the traffic passing through link ; is a time-varying value, indicating the traffic-carrying capacity of the link connecting node and node . Its value determines the thickness of the link or the pipe; It is the pressure difference or the public opinion difference between nodes.

[0034] Kirchhoff's law: The flow at nodes other than the source node and the target node obeys the conservation of flow in and out. By solving the system of equations of flow conservation for all nodes, the voltage distribution of nodes at each moment can be obtained, and then the flow distribution of the links can be obtained.

[0035] Applying Kirchhoff's law to all nodes in the target network gives: ; In the formula: represents the flow into the source node and the flow out of the target node (or represents the information flow that the source node wants to spread); represents connecting the source node and the nodes directly connected to it (i.e., the direct neighbor nodes), and the flow passing through the link connecting them; represents connecting the target node and the nodes directly connected to it The flow into and out of the intermediate nodes is equal and cancels each other out.

[0036] For a target network containing multiple nodes, the above two equations can be written in matrix form as: ; In the formula: The non-diagonal elements of the matrix are the conductances of the links between node and node , and the diagonal elements ; The column vector has , and , represents the vector describing the magnitude of the net flow in / out of each node, represents the element corresponding to the index of the source node s in , represents the element corresponding to the index of the target node in , represents the element corresponding to the index of other nodes in .

[0037] Analogize the links in the target network to pipelines, and the pipelines have a certain flow-carrying capacity 。The flow allocation and flow carrying capacity of each pipeline are iteratively updated through an evolutionary algorithm. At the initial stage of iteration, the flow carrying capacity of each pipeline is initialized, and the flow allocation is triggered. The flow carrying capacity of each pipeline is adjusted according to the flow classification result and the positive feedback mechanism. In subsequent iteration rounds, a new round of flow allocation is triggered based on the new flow carrying capacity of each pipeline, and the flow carrying capacity of each pipeline is adjusted again according to the flow classification result and the positive feedback mechanism until the evolution reaches a steady state.

[0038] In one implementation, the flow carrying capacity of each pipeline is a non-fixed value that changes over time, and the flow carrying capacity of the pipeline and the flow in the pipeline change based on the positive feedback mechanism.

[0039] Specifically, the positive feedback mechanism of the information flow is specifically reflected as follows: The flow carrying capacity of the pipeline is not a fixed value, but a value that changes over time. The change trend depends on the magnitude of the flow in the pipeline, and the change trend follows the positive feedback mechanism: when the flow is large, the flow carrying capacity of the pipeline increases; when the flow is small, the flow carrying capacity of the pipeline weakens, which in turn leads to a decrease in the flow. In an actual application scenario, taking a pipeline as an example, after the flow allocation is triggered, if the flow in the pipeline is less than the flow carrying capacity of the pipeline, the subsequent flow carrying capacity of the pipeline will decrease; if the flow in the pipeline is greater than the flow carrying capacity of the pipeline, the subsequent flow carrying capacity of the pipeline will increase.

[0040] In one implementation, the evolutionary algorithm is used for: First step, based on the flow carrying capacity of each pipeline between the source node and the target node, trigger the flow allocation of the current round; Second step, adjust the flow carrying capacity of each pipeline between the source node and the target node according to the flow allocation result of the current round and the pre-established differential equation; wherein, the differential equation is generated based on the dynamic evolution model established by the positive feedback mechanism, and the differential equation is used to describe the dynamic evolution law of the flow carrying capacity of the pipeline with the change of the flow. Third step, determine whether the evolution reaches a steady state. If it does not reach a steady state, continue to execute the first step until the evolution reaches a steady state and stop the iteration.

[0041] The traffic distribution basis in this embodiment includes two parts. One part is to distribute traffic according to the traffic-carrying capacity of the adaptive pipeline, that is, the pipeline with a larger traffic-carrying capacity will be allocated more traffic; the other part is to distribute traffic based on Kirchhoff's law, that is, the node traffic is conserved: the total outflow traffic of the source node is equal to the total inflow traffic of the target node, and the inflow traffic of the intermediate node is equal to the outflow traffic. In this embodiment, a dynamic evolution model of the traffic-carrying capacity of the pipeline is pre-constructed, and the traffic-carrying capacity of the pipeline is adjusted through the dynamic evolution model combined with the positive feedback mechanism. In other words, when performing traffic distribution, among the multiple pipelines between the source node and the target node, each has a different traffic-carrying capacity. According to the traffic-carrying capacity of these pipelines, it is determined how the current round of traffic is distributed among each pipeline. For example, a pipeline with a larger traffic-carrying capacity may be allocated more traffic, while a pipeline with a smaller traffic-carrying capacity is allocated relatively less traffic.

[0042] The calculation result of the differential equation of the dynamic evolution model is the derivative. In the initial state, the traffic-carrying capacity of the pipeline will be set to a preset value (for example, the preset value is 1), and the initial derivative will decrease. During the iteration process, since the traffic-carrying capacity of the pipeline changes, the traffic distribution situation will also change. According to the conservation law, when the traffic of some pipelines increases, the traffic of some other pipelines will decrease. According to the positive feedback mechanism, the traffic of some pipelines will continue to increase until it stabilizes, and the traffic of some other pipelines will continue to decrease to the steady state, or even decrease to 0.

[0043] After a finite number of iterations, the evolution will reach a steady state, that is, all indicators of the target network are close to stable and no longer change. When the evolution reaches the steady state, the links with traffic in the pipeline can form a connected graph. In this connected graph, for each node, there are links carrying traffic inflow and links carrying traffic outflow. Except for the source node and the target node, traffic will not stay at nodes connected by only one link.

[0044] In one implementation, the differential equation includes a first parameter and a second parameter; the first parameter is used to adjust the steady-state value at which the pipeline finally converges; the second parameter is used to adjust the pipeline convergence speed.

[0045] The dynamic evolution model of the traffic-carrying capacity of the pipeline can describe the change of the traffic-carrying capacity or conductance between any two adjacent nodes in the target network over time points. The differential equation of the dynamic evolution model can describe the positive feedback mechanism of the traffic-carrying capacity of the pipeline adapting to traffic. This differential equation includes two key parameters: the first parameter , which is used to adjust the steady-state value at which the pipeline finally converges, and the second parameter is used to adjust the pipeline convergence speed. Among them, and The value ranges of all need to be greater than 1. It is determined according to the pressure drop on each link.

[0046] Specifically, the differential equation form of the dynamic evolution model of the flow carrying capacity of the pipeline can be expressed as follows: ; In the formula, represents the change value of based on the differential equation within the time ; represents the change step size; is a time-varying value, representing the flow carrying capacity of the link connecting node and node , or the thickness of the pipeline; represents that it is to the power, used to control the change trend of the differential equation, The parameter represents the parameter for controlling the proportion of the dominant path retained finally.

[0047] In order to make the differential equation have a solution and its solution distribution is convenient for subsequent setting of thresholds to screen the dominant path, it is necessary to ensure and . When is a constant, the equation will have a steady-state solution: ; It should be noted that in this embodiment, the value is not fixed. The steady-state values in this embodiment are not based on the theoretical steady-state values, but on the simulation steady-state values. For , the setting originates from the heterogeneous characteristics of the target network. For example, there are significant differences in the functional attributes (such as bridging edges and redundant edges) and load pressures of different links in the target network.

[0048] Several high-load links connecting the source node and the target node are called the dominant path. In order to effectively distinguish the dynamic behaviors of high-load links and low-load links, this embodiment abandons the traditional static parameterization method and introduces the concept of the time-varying value . Its core basis is the dynamic evolution of the link voltage difference. The link loads in the target network show non-uniform distribution. For example, the hub links carry high traffic and the edge links have low loads. The time-varying value accurately captures the spatio-temporal heterogeneity of the link state through the real-time feedback of the voltage difference, avoiding model distortion caused by fixed parameters.

[0049] Traditional static models are prone to falling into local equilibrium due to path dependence in path selection (such as over-reliance on strong connections). In contrast, the dominant path intelligent recognition model in this embodiment uses time-varying values Introduce a positive feedback mechanism: High-load links enhance their traffic adsorption capacity through capacity expansion, while low-load links are naturally phased out. This process mimics self-organizing regulation in biological systems, such as pheromone attenuation in ant colonies, to ensure that network resources continuously tilt towards high-load links, thereby breaking through the local convergence limit of the greedy algorithm. And based on Kirchhoff's law, the dynamic adjustment of link capacity strictly follows the traffic conservation constraint. When a link exits the previously identified dominant path due to a decrease in load, the traffic it releases will be redistributed to adjacent links according to the principle of minimum resistance, triggering a new round of capacity evolution.

[0050] Specifically, The value is designed as: ; In the formula, is the pressure drop on each link after the inflow of traffic; is in the mapping process from voltage to and the mapping process from to the steady-state value, to avoid canceling out the effect of The coefficient.

[0051] In one implementation, according to the traffic distribution result and a pre-established differential equation, the traffic carrying capacity of each pipeline between the source node and the target node is adjusted, including: For each pipeline between the source node and the target node, if the calculation result of the differential equation corresponding to this pipeline is negative, indicating that the in-pipe flow of this pipeline is less than the maximum traffic carrying capacity, then reduce the traffic carrying capacity of this pipeline; If the calculation result of the differential equation corresponding to this pipeline is positive, indicating that the in-pipe flow of this pipeline is greater than the maximum traffic carrying capacity, then increase the traffic carrying capacity of this pipeline; If the traffic carrying capacity of this pipeline becomes zero after adjustment, but the evolution has not reached the steady state, retain this pipeline, but suspend the transmission activity of this pipeline until the traffic carrying capacity of this pipeline increases during subsequent evolution, then resume the transmission activity of this pipeline.

[0052] Specifically, the implementation basis of the positive feedback mechanism of the information flow is that the thickness of the pipeline can change according to the change of the flow rate in the pipeline. Therefore, the essence of the differential equation in this embodiment is to describe the dynamic evolution law of the flow rate carrying capacity of the pipeline with the change of the flow rate. The positive or negative sign of the calculation result of the differential equation reflects the correlation between the flow rate of the pipeline and the flow rate carrying capacity, and amplifies this relationship through positive feedback: when the flow rate in the pipeline exceeds the flow rate carrying capacity, the flow rate carrying capacity is strengthened; when the flow rate in the pipeline is insufficient, the flow rate carrying capacity is weakened, and finally the system converges to the equilibrium state of the flow rate and the flow rate carrying capacity.

[0053] Taking a pipeline as an example, when the calculated flow rate of the pipeline is less than the maximum flow rate carrying capacity of the pipeline, the calculation result of the differential equation is negative, indicating that the flow rate carrying capacity of the pipeline becomes weaker at the next moment; when the calculated flow rate of the pipeline is greater than the maximum flow rate carrying capacity of the pipeline, the calculation result of the differential equation is positive, indicating that the flow rate carrying capacity of the pipeline becomes stronger at the next moment. The process of the flow rate carrying capacity of the pipeline gradually becoming zero is equivalent to the self-ablation process. However, in order to avoid falling into a local optimal solution, in this embodiment, when the evolution has not reached a steady state, the pipeline with a flow rate carrying capacity of zero will still be retained, and only its transmission activity will be suspended. In the subsequent evolution process, if a local optimal solution is jumped out and the flow rate carrying capacity of such a pipeline rises, its transmission activity can be restored. In other words, some pipelines will be initially identified as non-dominant links, the flow rate carrying capacity gradually becomes zero, and then when a local optimal solution is jumped out, they will be identified as dominant paths again, and the flow rate carrying capacity will rise again.

[0054] In one implementation, after iteration, if the derivative calculated by the differential equation is less than 0, the calculation result of the differential equation is directly set to , where e represents the natural constant.

[0055] Specifically, when is associated with the voltage difference on the link, at a certain moment, if the flow rate on the link is very small, the pressure difference is also correspondingly very small. When , the entire value will become very large, and at this time the derivative calculated by the differential equation will become negative. However, since is set in the range of [0,1], it will be a number in the range of 0 to 1 minus a very large number. At this time, this derivative may exceed its own size and will become negative after 1 round of iteration. Taking the square root of a negative number is illegal and does not conform to reality (in reality, the flow rate carrying capacity of the pipeline will not be less than 0). At this time, artificial normalization is required. Therefore, in the actual application scenario, when, after iteration, if the derivative calculated by the differential equation is less than 0, it is directly set to .

[0056] The core of the self-ablating mechanism of the information flow is to achieve the elimination of redundant pipelines through threshold judgment, ensuring that system resources are concentrated on high-load links. Moreover, the self-ablating mechanism can also strengthen the positive feedback mechanism: the positive feedback mechanism itself will continuously improve the traffic carrying capacity of high-load links and continuously reduce the traffic carrying capacity of low-load links. When a low-load link cannot meet the minimum survival conditions, it is directly cleared to prevent it from occupying system resources, making the traffic more concentrated on high-load links and forming a virtuous cycle. For example, if the traffic of some links is continuously less than the maximum traffic carrying capacity of the pipeline, the traffic carrying capacity of the pipeline will continuously weaken during the entire evolution process until the thickness of the pipeline approaches 0, which is equivalent to the pipeline being self-ablated due to traffic feedback.

[0057] In one implementation, the dominant path is determined according to the traffic distribution result or traffic carrying capacity of each pipeline after the evolution reaches a steady state, including: According to the traffic distribution result or traffic carrying capacity of each pipeline after the evolution reaches a steady state, and a preset threshold, several target links are screened out, and the path composed of all the target links is a connected graph; The dominant path is determined according to all the target links.

[0058] In an actual application scenario, when the indicators in the target network tend to be stable, it means that the evolution reaches a steady state. For example the value, the information level value of the node, the calculated voltage or current value basically no longer change. When the evolution reaches a steady state, the target network is filled with traffic. However, due to the different traffic carrying capacities of each link, some links have a large amount of traffic and some links have a small amount of traffic. Therefore, it is necessary to further screen the dominant path. The screening criteria for the dominant path include the following three: (1) Screening target links based on traffic size: A larger traffic volume of a link indicates that the link makes a greater contribution to traffic distribution and can be considered as part of the dominant path; among them, there is a corresponding relationship among the edges, links, and pipelines in this embodiment; (2) Screening target links based on path connectivity: The path composed of the screened links should ensure that it is a connected subgraph without isolated edges or nodes; (3) Screening target links based on the consideration of weak-load links: Weak-load links with small traffic should not be completely ignored. They may still play a role in some areas of the network (such as some long paths in the network). Therefore, a balance mechanism is needed to avoid completely discarding weak-load links with small traffic while maintaining the connectivity of the path.

[0059] The target links that ultimately form the dominant path are comprehensively determined through the above three screening criteria. Most of the target links are high-load links, but in order to maintain path connectivity, there may also be some low-load links.

[0060] The way to calculate the traffic distribution in the dominant path intelligent recognition model is to use Kirchhoff's law, so it will naturally exclude nodes with only a single link. And at any time, the graph composed of the links with traffic in the network is a connected graph. All in all, due to the addition of Kirchhoff's law, the dominant path intelligent recognition model naturally meets the above first two screening criteria. For the third screening criterion, weak-load links can play a role in helping information diffusion in the connection scenario between clusters. When the connections between clusters are sparse and the interacting nodes are across communities, according to Kirchhoff's law and the conservation law, the traffic will surely flow through these inter-cluster links. At this time, the weak-load links with existing value can be accurately found through threshold screening. In other words, in the connection scenario between clusters, if the weak-load links carry traffic, they can be retained by reasonably setting the threshold.

[0061] For example, when the evolution reaches a steady state, a threshold is set artificially, such as . The links with traffic less than or equal to are regarded as having traffic approximately equal to 0 or equal to 0, and they are disconnected; the links with traffic greater than , but not reaching the threshold corresponding to the high-load links, are regarded as weak-load links that need to be retained.

[0062] Therefore, through information flow recognition in this embodiment, not only high-load links can be recognized, but also weak-load links that need to be retained can be recognized: after iterating to the steady state, if there is traffic in the weak-load links, they are retained as part of the dominant path; if there is no traffic in the weak-load links, it means that they do not help much in information diffusion during information transmission and cannot be regarded as part of the dominant path.

[0063] Based on the above embodiments, the present invention also provides a dominant path recognition system in a single-source scenario based on information flow positive feedback, as Figure 2 shown. The system includes: A data input module 01, configured to determine a pair of source nodes and target nodes for information interaction in the target network, and input the network structure data of the target network into a dominant path intelligent recognition model constructed based on the information flow positive feedback self-ablating mechanism; The dominant path intelligent recognition model 02 is used to, according to the network structure data, equivalent the links for information transmission in the target network to the pipelines carrying traffic during information transmission in an electrical network, and the traffic flowing through each node in the target network conforms to the conservation law, and the traffic flowing through each link conforms to Kirchhoff's law; The flow distribution and flow-carrying capacity of the pipeline between the source node and the target node are iteratively updated through an evolutionary algorithm until the evolution reaches a steady state; according to the flow distribution result or flow-carrying capacity of each pipeline after the evolution reaches a steady state, the dominant path when the source node and the target node perform information interaction is determined.

[0064] Based on the above embodiments, the present invention further provides a terminal, and its principle block diagram can be as Figure 3 shown. The terminal includes a processor, a memory, a network interface, and a display screen connected through a system bus. Among them, the processor of the terminal is used to provide computing and control capabilities. The memory of the terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the terminal is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it realizes the method for identifying the dominant path in a single-source scenario based on positive feedback of information flow. The display screen of the terminal can be a liquid crystal display screen or an electronic ink display screen.

[0065] Those skilled in the art can understand that Figure 3 the principle block diagram shown in

[0066] merely shows the block diagram of some structures related to the solution of the present invention, and does not constitute a limitation on the terminal to which the solution of the present invention is applied. The specific terminal may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0067] 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, storage, database, or other medium used in the embodiments provided by the present invention can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0068] In summary, the present invention discloses a method and system for identifying a dominant path in a single-source scenario based on positive feedback of information flow, which relates to the field of computer technology. The method includes: determining a pair of source node and target node for information interaction in a target network, and inputting the network structure data of the target network into a dominant path intelligent identification model constructed based on the self-ablation mechanism of information flow positive feedback; determining the dominant path when the source node and the target node perform information interaction through the dominant path intelligent identification model; the dominant path intelligent identification model is used for: equivalently converting the target network into an electrical network according to the network structure data; wherein, the nodes in the target network correspond to the nodes in the electrical network, the link for information transmission in the target network corresponds to the pipeline for carrying traffic during information transmission in the electrical network, and each node conforms to Kirchhoff's law; iteratively updating the traffic distribution and traffic carrying capacity of the pipeline between the source node and the target node through an evolutionary algorithm until the evolution reaches a steady state, and determining the dominant path according to the traffic distribution result or traffic carrying capacity of each pipeline after the evolution reaches the steady state. The present invention conducts physical mechanism modeling through electrical network equivalence, Kirchhoff's law, and positive feedback-driven dynamic evolution, and is based on physical mechanisms such as traffic conservation and dynamic evolution of pipeline carrying capacity, so that the model has a clear dynamic interpretation. It solves the problem of insufficient mechanism modeling existing in the existing dominant path identification methods. The present invention establishes a dominant path intelligent identification model based on the self-ablation mechanism of information flow positive feedback from the mechanism level for network information propagation. Taking the analysis of the penetration law of information flow as the starting point, the dominant path when a pair of source node and target node in the target network perform information interaction is quickly and accurately identified through information flow identification. In practical application scenarios, the present invention can provide new solutions for finding dominant paths in various networks, and has a core supporting role in constructing a propagation risk warning model based on network science, implementing key node intervention strategies, and optimizing the intelligent governance mechanism of social information systems.

[0069] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or transformations can be made according to the above description, and all such improvements and transformations shall fall within the protection scope of the appended claims of the present invention.

Claims

1. A method for identifying a dominant path in a single-source scenario based on positive feedback of information flow, characterized in that: The method comprises: Determine a pair of source nodes and target nodes for information interaction in a target network, and input the network structure data of the target network into a dominant path intelligent identification model constructed based on the information flow positive feedback self-ablation mechanism; Determine the dominant path when the source node and the target node perform information interaction by using the dominant path intelligent identification model; The dominant path intelligent identification model is used to: According to the network structure data, the link used for information transmission in the target network is equivalent to a pipe that carries traffic during information transmission in the electrical network, and the traffic flowing through each node in the target network complies with the conservation law, and the traffic flowing through each link complies with Kirchhoff's law; The flow distribution and flow carrying capacity of the pipeline between the source node and the target node are iteratively updated by an evolutionary algorithm until the evolution reaches a steady state, and the dominant path is determined according to the flow distribution results or flow carrying capacity of each pipeline after the evolution reaches a steady state.

2. The method for identifying the dominant path in a single-source scenario based on positive feedback of information flow according to claim 1 is characterized in that: The type of the target network includes a social network.

3. The dominant path identification method in a single-source scenario based on information flow positive feedback according to claim 1 is characterized in that: The flow carrying capacity of each pipeline is a non-fixed value that changes with time, and the flow carrying capacity of the pipeline changes with the flow in the pipeline based on a positive feedback mechanism.

4. The method for identifying the dominant path in a single-source scenario based on positive feedback of information flow according to claim 3 is characterized in that: The evolutionary algorithm is used to: The first step is to trigger the current round of traffic allocation based on the traffic carrying capacity of each pipeline between the source node and the target node; The second step is to adjust the flow carrying capacity of each pipeline between the source node and the target node according to the flow distribution result of the current round and the pre-established differential equation; wherein the differential equation is generated based on the dynamic evolution model established by the positive feedback mechanism, and the differential equation is used to describe the dynamic evolution law of the flow carrying capacity of the pipeline with the change of flow; The third step is to determine whether the evolution has reached a steady state. If not, continue to execute the first step until the evolution reaches a steady state and stop the iteration.

5. The method for identifying the dominant path in a single-source scenario based on positive feedback of information flow according to claim 4 is characterized in that: The differential equation includes a first parameter and a second parameter; the first parameter is used to adjust the steady-state value of the final pipeline convergence; the second parameter is used to adjust the pipeline convergence speed.

6. The method for identifying the dominant path in a single-source scenario based on positive feedback of information flow according to claim 4 is characterized in that: According to the flow distribution result and the pre-established differential equation, the flow carrying capacity of each pipeline between the source node and the target node is adjusted, including: For each pipeline between the source node and the target node, if the calculation result of the differential equation corresponding to the pipeline is a negative number, indicating that the flow rate in the pipeline is less than the maximum flow rate carrying capacity, then the flow rate carrying capacity of the pipeline is reduced; If the calculation result of the differential equation corresponding to the pipeline is a positive number, it means that the flow rate in the pipeline is greater than the maximum flow carrying capacity, then the flow carrying capacity of the pipeline is increased; If the flow carrying capacity of the pipeline is zero after adjustment, but the evolution has not reached a steady state, the pipeline will be retained, but the transmission activities of the pipeline will be suspended until the flow carrying capacity of the pipeline increases in the subsequent evolution process, then the transmission activities of the pipeline will be resumed.

7. The method for identifying the dominant path in a single-source scenario based on positive feedback of information flow according to claim 1 is characterized in that: Determining the dominant path according to the flow distribution result or flow carrying capacity of each pipeline after the evolution reaches a steady state includes: According to the flow distribution results or flow carrying capacity of each pipeline after the evolution reaches a steady state, and a preset threshold, a number of target links are selected, and a path composed of all the target links is made into a connected graph; The dominant path is determined according to all of the target links.

8. A dominant path identification system in a single-source scenario based on positive feedback of information flow, characterized in that: The system comprises: A data input module, used to determine a pair of source nodes and target nodes for information interaction in a target network, and input the network structure data of the target network into a dominant path intelligent identification model constructed based on the information flow positive feedback self-ablation mechanism; The dominant path intelligent identification model is used to, based on the network structure data, equate the links used for information transmission in the target network to pipes that carry traffic during information transmission in the electrical network, and the traffic flowing through each node in the target network complies with the conservation law, and the traffic flowing through each link complies with Kirchhoff's law; The flow distribution and flow carrying capacity of the pipeline between the source node and the target node are iteratively updated through an evolutionary algorithm until the evolution reaches a steady state; according to the flow distribution results or flow carrying capacity of each pipeline after the evolution reaches a steady state, the dominant path when the source node and the target node perform information interaction is determined.

9. A terminal, characterized in that: The terminal includes a memory and one or more processors; the memory stores one or more programs; the program contains instructions for executing the dominant path identification method in a single-source scenario based on positive feedback of information flow as described in any one of claims 1-7; and the processor is used to execute the program.

10. A computer-readable storage medium having a plurality of instructions stored thereon, characterized in that: The instructions are suitable for being loaded and executed by a processor to implement the steps of the dominant path identification method in a single-source scenario based on positive feedback of information flow as described in any one of claims 1-7.

Images