Power communication path determination method and device, storage medium and computer equipment

Abstract

The invention discloses an electric power communication path determination method and device, a storage medium and computer equipment, and the method comprises the steps: obtaining a network topological graph of a target electric power communication network, and determining a starting node and a termination node of a to-be-determined electric power communication path in the network topological graph; the network topological graph is identified to obtain an identification result, and the identification result comprises nodes of the network topological graph, connection lines between the nodes, connection relations between the nodes and node attribute information; the nodes are communication devices in the electric power communication network, connection lines between the nodes are transmission media between the communication devices, the connection relation between the nodes is a transmission medium communication interaction rule, and the node attribute information comprises at least one of bandwidth, time delay, packet loss rate, device type and carried communication traffic; and based on the identification result, searching an optimal power communication path from the starting node to the ending node in the network topological graph.

Description

Technical Field

[0001] This invention relates to the field of power communication technology, and in particular to a method, apparatus, storage medium, and computer equipment for determining power communication paths. Background Technology

[0002] A large-scale power communication network consists of various communication devices and transmission media. With the rapid development of information technology, the role of power communication networks in power systems is becoming increasingly important. As one of my country's important power communication infrastructures, the power communication network undertakes the crucial task of ensuring the safe and stable operation of the power system. To improve the performance and reliability of the power communication network, it is necessary to determine the power communication paths.

[0003] Currently, the shortest path in a power communication network is typically used as the power communication path. However, the shortest path is usually statically determined based on the number of hops in the power communication network. This statically determined shortest path is not guaranteed to ensure stable and smooth communication, and may lead to low resource utilization of the power communication network. Summary of the Invention

[0004] This invention provides a method, apparatus, storage medium, and computer equipment for determining power communication paths, which mainly improves the accuracy of power communication path determination and reduces the low utilization rate of power network resources, thereby enhancing the power communication effect.

[0005] According to a first aspect of the present invention, a method for determining a power communication path is provided, comprising: In response to a signal indicating the determination of a power communication path for a target power communication network, a network topology map of the target power communication network is acquired, and the starting node and ending node of the power communication path to be determined are determined in the network topology map. The network topology is identified to obtain an identification result, wherein the identification result includes the nodes of the network topology, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of bandwidth, latency, packet loss rate, device type, and the amount of communication carried. Based on the identification results, the optimal power communication path from the starting node to the ending node is searched in the network topology map, and the optimal power communication path is taken as the power communication path of the target power communication network.

[0006] Optionally, the step of identifying the network topology map to obtain the identification result includes: The network topology graph is converted into a topology matrix, wherein the nodes in the network topology graph are mapped to the corresponding pixels in the topology matrix, and the lines between the nodes in the network topology graph are mapped to the connection paths between each pixel in the topology matrix. The pixels and connection paths in the topology matrix are encoded using a preset space-filling curve to obtain an encoded topology matrix; The encoded topology matrix is ​​split into a multi-channel topology matrix, wherein the multi-channel topology matrix includes a node connection representation matrix of nodes and connections in the network topology graph, a color information representation matrix of the network topology graph, and a depth information representation matrix of the network topology graph. Obtain a preset feature extraction model, wherein the preset feature extraction model includes a convolutional layer for parallel feature extraction, a batch normalization layer for feature normalization, a pooling layer for feature pooling, and a fully connected layer for feature integration. The node connection representation matrix, the color information representation matrix, and the depth information representation matrix are input into the preset feature extraction model. The convolutional layer performs parallel feature extraction on the node connection representation matrix, the color information representation matrix, and the depth information representation matrix in corresponding channels to obtain parallel features for different channels. The parallel features are then processed sequentially through the batch normalization layer and the pooling layer to obtain the parallel pooling features output by the pooling layer. The parallel pooling features are then input into the fully connected layer for feature integration to obtain node features and connection features in the network topology graph. Based on the node features and the connection features, the nodes, connections between nodes, connection relationships between nodes, and node attribute information in the network topology graph are identified.

[0007] Optionally, based on the node features and the connection features, the nodes, connections between nodes, and node attribute information in the network topology graph are identified, including: Extract the local and global features of the node features to obtain the corresponding node local features and node global features; extract the local and global features of the connection features to obtain the corresponding connection local features and connection global features. The local features and global features of the nodes are fused to obtain node fused features, and the local features and global features of the connections are fused to obtain connection fused features; The node fusion features and the connection fusion features are input into a preset node recognition model to obtain the nodes, connections between nodes, and node attribute information in the network topology diagram.

[0008] Optionally, based on the node features and the connection features, identifying the connection relationships between nodes in the network topology graph includes: Each node in the network topology graph is taken as a target node, and multiple neighboring nodes corresponding to the target node are determined in the network topology graph. The association weight of the target node relative to each of the neighboring nodes is determined respectively. The neighboring nodes whose association weight is greater than a preset weight threshold are taken as the target neighbor nodes. The node features of the target neighbor nodes are aggregated to the node features of the target node to obtain the aggregated features of the target node. The aggregation features of each node in the network topology graph and the connection features are input into a preset topology graph recognition model to identify the connection relationships, thereby obtaining the connection relationships between nodes in the network topology graph.

[0009] Optionally, based on the identification result, searching for the optimal power communication path from the starting node to the ending node in the network topology graph includes: The network topology map is divided into regions to obtain network topology maps corresponding to different regions; Different preset heuristic functions are determined for the network topology maps corresponding to different regions, wherein the preset heuristic function is a function constructed from at least distance and time; Take any region among the different regions as a target region, take the network topology graph corresponding to the target region as the target network topology graph, and take the preset heuristic function corresponding to the target network topology graph as the preset target heuristic function. Determine a preset open list and a preset closed list, and add the starting node in the target network topology diagram to the preset open list; The preset target heuristic function is used to evaluate the estimated cost from the starting node to the ending node in the target network topology graph, and the actual cost between the starting node and each target neighbor node in the target network topology graph is determined. The estimated priority of the target neighbor node is determined based on the estimated cost, the actual priority of the target neighbor node is determined based on the actual cost, and a high-quality neighbor node is selected from the target neighbor nodes based on the estimated priority and added to the preset open list. Based on the estimated and actual priorities of the high-quality neighbor nodes, the total priority of each high-quality neighbor node in the preset open list is determined, and the optimal neighbor node corresponding to the maximum total priority is transferred to the preset closed list. The optimal neighbor node corresponding to the highest total priority in the preset closed list is taken as the new starting node, and the target neighbor node adjacent to the new starting node is taken as the new target neighbor node. Based on the new starting node, the new target neighbor node, and the preset target heuristic function, a new high-quality neighbor node that meets the cost requirements from the starting node to the ending node of the target network topology is searched iteratively. The new optimal neighbor node is then moved to the preset closed list until the ending node of the target network topology is found. Finally, the optimal power communication path is formed based on the starting node, the ending node of the target network topology, and each optimal neighbor node in the preset closed list.

[0010] Optionally, based on the identification result, searching for the optimal power communication path from the starting node to the ending node in the network topology graph includes: Based on the connections between nodes in the network topology diagram, multiple candidate power communication paths are determined from the starting node to the ending node in the network topology diagram, and link information between each node on each candidate power communication path is determined, wherein the link information includes at least one of load and failure rate. Based on the node attribute information, link information between nodes, and connection relationships between nodes of each candidate power communication path, the optimal power communication path is selected on each candidate power communication path.

[0011] Optionally, selecting the optimal power communication path on each candidate power communication path based on the node attribute information of each node on each candidate power communication path, the link information between nodes, and the connection relationship between nodes includes: For each node in each of the candidate power communication paths, the attribute feature vector corresponding to the node attribute information, the link feature vector corresponding to the link information, and the connection feature vector corresponding to the connection relationship are determined respectively. The attribute feature vector, the link feature vector, and the connection feature vector are cross-processed to obtain the path cross feature vector; The path intersection feature vector of each candidate power communication path is input into a preset path prediction model to predict the path and obtain the optimal power communication path.

[0012] According to a second aspect of the present invention, an apparatus for determining a power communication path is provided, comprising: The acquisition unit is configured to acquire a network topology map of the target power communication network in response to a determination signal for a power communication path of the target power communication network, and determine the starting node and ending node of the power communication path to be determined in the network topology map. The identification unit is used to identify the network topology map and obtain the identification result, wherein the identification result includes the nodes of the network topology map, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of bandwidth, latency, packet loss rate, and device type. The search unit is used to search for the optimal power communication path from the starting node to the ending node in the network topology map based on the identification result, and to use the optimal power communication path as the power communication path of the target power communication network.

[0013] According to a third aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described method for determining a power communication path.

[0014] According to a fourth aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described method for determining a power communication path.

[0015] According to the present invention, a method, apparatus, storage medium, and computer device for determining power communication paths, compared with the current method of using the shortest path in the power communication network as the power communication path, the present invention identifies nodes, connections between nodes, connection relationships between nodes, and node attribute information by identifying the network topology of the power communication network, and determines the optimal power communication path by comprehensively analyzing the nodes, connections between nodes, connection relationships between nodes, and node attribute information. This ensures that the determined optimal power communication path can guarantee communication stability and smoothness while satisfying the requirement of reasonable path length, thereby improving the accuracy of power communication path determination and enhancing the utilization rate of communication resources. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 A flowchart of a method for determining a power communication path provided by an embodiment of the present invention is shown; Figure 2 A flowchart of another method for determining a power communication path provided by an embodiment of the present invention is shown; Figure 3 A schematic diagram of the structure of a power communication path determination device provided in an embodiment of the present invention is shown; Figure 4 A schematic diagram of another power communication path determination device provided in an embodiment of the present invention is shown; Figure 5 A schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention is shown. Detailed Implementation

[0017] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.

[0018] Currently, using the shortest path in the power communication network as the power communication path has limitations. Since the shortest path is not necessarily the most stable or smooth path for communication, the accuracy of power communication path determination is low, resulting in low utilization of power network resources.

[0019] To address the aforementioned problems, embodiments of the present invention provide a method for determining a power communication path, such as... Figure 1 As shown, the method includes: 101. In response to the determination signal of the power communication path for the target power communication network, obtain the network topology map of the target power communication network, and determine the starting node and ending node of the power communication path to be determined in the network topology map.

[0020] Each power communication network has its corresponding network topology diagram, which includes nodes corresponding to various communication devices and edges connecting these nodes. Communication devices include power plants, substations, control centers, user terminals, and other communication equipment (such as routers, switches, and repeaters).

[0021] In this embodiment of the invention, after obtaining the network topology map of the target power communication network, the starting node and the ending node of the power communication path to be determined are determined in the network topology map, that is, the communication starting point device and the communication terminal device for power communication.

[0022] 102. Identify the network topology and obtain the identification results, which include the nodes of the network topology, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of the following: bandwidth, latency, packet loss rate, device type, and the amount of communication carried.

[0023] The transmission media include, but are not limited to, optical fiber, twisted pair, cable, power line carrier, microwave communication, etc.; the communication interaction rules of the transmission media include direct connection, indirect connection, multipath connection, etc.; the node attribute information also includes the node's capacity, processing capability, status (such as online / offline), etc. In this embodiment of the invention, after determining the network topology, information such as nodes, connections between nodes, connection relationships between nodes, and node attribute information is extracted from the network topology.

[0024] 103. Based on the identification results, search for the optimal power communication path from the starting node to the ending node in the network topology graph, and use the optimal power communication path as the power communication path of the target power communication network.

[0025] In this embodiment of the invention, after identifying the network topology and obtaining information such as nodes, connections between nodes, connection relationships between nodes, and node attribute information, it is necessary to iteratively search for the optimal power communication path from the starting node to the ending node in the network topology based on the above information. Therefore, step 103 specifically includes: dividing the network topology into regions to obtain network topology maps corresponding to different regions; determining different preset heuristic functions for the network topology maps corresponding to different regions, wherein the preset heuristic function is a function constructed from at least distance and time; taking any region among the different regions as a target region, taking the network topology map corresponding to the target region as the target network topology map, and taking the preset heuristic function corresponding to the target network topology map as the preset target heuristic function; determining a preset open list and a preset closed list, and adding the starting node in the target network topology map to the preset open list; using the preset target heuristic function to evaluate the estimated cost from the target neighbor node adjacent to the starting node in the target network topology map to the ending node, and determining the actual cost between the starting node and each target neighbor node in the target network topology map; and determining the actual cost based on the estimated cost. The estimated priority of the target neighbor node is determined based on the actual cost. A high-quality neighbor node is selected from the target neighbor nodes based on the estimated priority and added to the preset open list. Based on the estimated and actual priorities of the high-quality neighbor nodes, the total priority of each high-quality neighbor node in the preset open list is determined. The optimal neighbor node corresponding to the highest total priority is moved to the preset closed list. The optimal neighbor node corresponding to the highest total priority in the preset closed list is used as the new starting node. The target neighbor nodes adjacent to the new starting node are used as new target neighbor nodes. Based on the new starting node, the new target neighbor nodes, and the preset target heuristic function, a new high-quality neighbor node whose cost from the starting node to the ending node of the target network topology satisfies the requirements is iteratively searched. The new optimal neighbor node is moved to the preset closed list. This process continues until the ending node of the target network topology is found. Finally, the optimal power communication path is constructed based on the starting node, the ending node of the target network topology, and each optimal neighbor node in the preset closed list.

[0026] Specifically, the network topology is divided into regions based on factors such as the function, geographical distribution, or traffic density of nodes in the network topology graph, resulting in network topology graphs corresponding to different regions. Different heuristic function calculation methods are used for nodes within different regions. For example, in high-traffic-density regions, higher weights can be given to link congestion levels, making the preset heuristic function more inclined to select relatively idle links. Conversely, in regions geographically close to the target but with poor link quality, the proportion of straight-line distance in the preset heuristic function can be appropriately reduced to guide the algorithm to explore other more reliable paths. This allows the algorithm to better adapt to the local characteristics of the topology graph during the search process, improving the accuracy and efficiency of path search.

[0027] Furthermore, the preset open list and preset closed list are table structures. An embodiment of the present invention illustrates the search structure for the optimal power communication path. For example, if the starting node is A, the ending node is B, and the target neighbor nodes adjacent to the starting node A are C and D, the node attribute information, connections between nodes, distances between nodes, transmission costs between nodes, connection relationships, and link information of each target neighbor node between the starting node and the ending node are substituted into a preset target heuristic function to obtain the estimated cost of each target neighbor node. Based on this estimated cost, an estimated priority is determined; the smaller the estimated cost, the higher its corresponding estimated priority. Simultaneously, the actual cost from the starting node to each neighbor node is determined, and the actual priority is determined based on this actual cost; the larger the actual cost, the lower its corresponding actual priority. Then, target neighbor nodes with estimated priorities greater than a preset priority threshold (the preset priority threshold is set according to actual needs) are added to the preset open list. The actual priority and estimated priority of the target neighbor nodes in the preset open list are added together to obtain the total priority. Finally, the target neighbor node with the highest total priority is moved to the preset closed list. Furthermore, the target neighbor nodes in the preset closed list are used as new starting nodes, and the nodes adjacent to the new starting node A are used as new target neighbor nodes. The node attribute information, connections, distances, transmission costs, connection relationships, and link information of each node between the new target neighbor node and the termination node are substituted into the preset target heuristic function to obtain the estimated cost of each new target neighbor node. Based on this estimated cost, an estimated priority is determined; the lower the estimated cost, the higher the estimated priority. Simultaneously, the actual cost between the new starting node and each new target neighbor node is determined, and the actual priority is determined based on this actual cost; the higher the actual cost, the lower the actual priority. Then, new target neighbor nodes with estimated priorities greater than a preset priority threshold (set according to actual needs) are added to the preset open list. The actual priority and estimated priority of the new target neighbor nodes in the preset open list are added together to obtain the total priority. Finally, the new target neighbor node with the highest total priority is moved to the preset closed list. Therefore, following the above method, new starting nodes are continuously determined from the preset closed list, and stage searches are continuously performed based on these new starting nodes until the location of the ending node is found. Finally, each node stored in the preset closed list is taken as the optimal neighbor node, and the optimal power communication path is formed by the starting node, the ending node, and the optimal neighbor nodes. That is, in this embodiment of the invention, initializing the open list is a crucial step in finding the optimal path. First, the starting point needs to be added to the open list and marked as visited. Then, the priority of each neighboring node is evaluated based on the distance from the starting point to other nodes and the heuristic function value, and the node with the highest priority is added to the open list.This process repeats continuously until only the destination remains in the open list. Initializing the open list ensures the search starts from the beginning and expands the search range progressively according to priority, ultimately finding the shortest path. Through continuous iteration, the elements in the open list are expanded one by one until the target node is found or the open list is empty. In each iteration, the node with the smallest cost function value is selected as the current node and moved from the open list to the closed list. Simultaneously, all neighboring nodes of the current node are generated, and their cost function values ​​and parent node information are updated. This process continues until the target node is found or no more nodes can be expanded. By cyclically executing the search process, the optimal power communication path can be found efficiently. This optimal power communication path is then used as the power communication path for the target power communication network, and power communication is then performed according to this path. Therefore, this embodiment of the invention identifies nodes, connections between nodes, connection relationships between nodes, and node attribute information by identifying the network topology of the power communication network. It then determines the optimal power communication path by comprehensively analyzing the nodes, connections between nodes, connection relationships between nodes, and node attribute information. This ensures that the determined optimal power communication path can guarantee stable and smooth communication while meeting the requirement of reasonable path length, thereby improving the accuracy of power communication path determination and enhancing the utilization rate of communication resources.

[0028] According to the present invention, a method for determining power communication paths, compared with the current method of using the shortest path in the power communication network as the power communication path, the present invention identifies the nodes, the connections between nodes, the connection relationships between nodes, and the node attribute information by identifying the network topology of the power communication network. Furthermore, it determines the optimal power communication path through comprehensive analysis of the nodes, the connections between nodes, the connection relationships between nodes, and the node attribute information. This ensures that the determined optimal power communication path, while satisfying the requirement of a reasonable path length, also guarantees stable and smooth communication, thereby improving the accuracy of power communication path determination and enhancing the utilization rate of communication resources.

[0029] Furthermore, to better illustrate the process of determining the power communication path described above, as a refinement and extension of the above embodiments, this invention provides another method for determining the power communication path, such as... Figure 2 As shown, the method includes: 201. In response to the power communication path determination signal for the target power communication network, obtain the network topology map of the target power communication network, and determine the starting node and ending node of the power communication path to be determined in the network topology map.

[0030] The graphical representation of communication equipment, nodes, and their connections in a power system visually illustrates the structure, data flow, and layout of key components in the power communication network. Specifically, it obtains the network topology map of the target power communication network and determines the starting and ending nodes of the power communication path based on actual transmission requirements, i.e., it identifies the communication initiation and termination devices in the power communication process.

[0031] 202. Convert the network topology graph into a topology matrix, where nodes in the network topology graph are mapped to corresponding pixels in the topology matrix, and the lines between nodes in the network topology graph are mapped to connection paths between each pixel in the topology matrix.

[0032] 203. Encode the pixels and connection paths in the topology matrix using a preset space filling curve to obtain an encoded topology matrix.

[0033] Specifically, the network topology graph is converted into a two-dimensional matrix (topology matrix). Besides mapping nodes in the network topology graph to pixels at specific locations in the matrix and representing connections as pixel paths, a pre-defined space-filling curve (such as a Z-curve or Hilbert curve) can be introduced to encode the pixel and connection information (connection paths) corresponding to the nodes. This encoding method can preserve the local and global structural information of the topology graph to a certain extent, enabling more effective feature capture in subsequent operations. For example, for network topologies with complex branching structures, pre-defined space-filling curve encoding can ensure relative proximity between adjacent nodes in the two-dimensional matrix, improving feature extraction efficiency and thus improving the efficiency of communication path determination.

[0034] 204. The encoded topology matrix is ​​split into a multi-channel topology matrix, wherein the multi-channel topology matrix includes a node connection representation matrix of nodes and connections in the network topology graph, a color information representation matrix of the network topology graph, and a depth information representation matrix of the network topology graph.

[0035] Specifically, node and connection representation elements are extracted from the encoded topology matrix, and a node and connection representation matrix is ​​constructed from these elements. Color representation elements are also extracted from the encoded topology matrix, and a color information representation matrix is ​​constructed from these elements. Finally, depth representation elements are extracted from the encoded topology matrix, and a depth information representation matrix is ​​constructed from these elements. This embodiment of the invention, by splitting the encoded topology matrix into a multi-channel topology matrix, enhances the representation of information such as nodes and connections in the topology graph, thereby improving feature extraction quality and ultimately increasing the accuracy of communication path determination.

[0036] 205. Obtain a preset feature extraction model, wherein the preset feature extraction model includes a convolutional layer for parallel feature extraction, a batch normalization layer for feature normalization, a pooling layer for feature pooling, and a fully connected layer for feature integration.

[0037] 206. Input the node connection representation matrix, color information representation matrix, and depth information representation matrix into the preset feature extraction model. Perform parallel feature extraction on the node connection representation matrix, color information representation matrix, and depth information representation matrix in the corresponding channels through the convolutional layer to obtain parallel features of different channels. Process the parallel features sequentially through the batch normalization layer and the pooling layer to obtain the parallel pooling features output by the pooling layer. Input the parallel pooling features into the fully connected layer for feature integration to obtain the node features and connection features in the network topology graph.

[0038] In this embodiment of the invention, to improve the feature extraction accuracy of the preset feature extraction model, it is first necessary to train and construct the preset feature extraction model. Based on this, the method includes: constructing a preset initial feature extraction model; obtaining a sample dataset, wherein the sample dataset includes a node connection representation matrix, a color information representation matrix, and a depth information representation matrix in the topology matrix corresponding to a sample power communication network topology map with node features and connection features; dividing the sample dataset into a training set and a test set; training the preset initial feature extraction model using the training set; testing the trained preset initial feature extraction model using the test set; and finally, using the trained preset initial feature extraction model that meets the test conditions as the preset feature extraction model.

[0039] Specifically, during model training, a pre-defined initial feature extraction model is first constructed, followed by the acquisition of a sample dataset. The dataset is ensured to contain all necessary files, and the data is converted to a format understandable by the pre-defined initial feature extraction model. Finally, the model is trained and tested. Specifically, the dataset can be divided first: using random or specific strategies (such as stratified sampling), the sample dataset is divided into training and test sets. The model is then trained using the training set, and tested using the test set to evaluate its performance on unseen data. Metrics such as mCP, precision, and recall on the test set are calculated and recorded. If the model performance does not meet requirements, the training phase can be returned for further iterations or adjustments. This process yields a pre-defined feature extraction model that meets the requirements.

[0040] Furthermore, after constructing the preset feature extraction model, the node connection representation matrix, color information representation matrix, and depth information representation matrix are jointly input into the preset feature extraction model. Convolutional layers in the model perform parallel feature extraction on different channels of the node connection representation matrix, color information representation matrix, and depth information representation matrix. Then, the features extracted from each channel are sequentially input into a batch normalization layer and a pooling layer for feature processing, obtaining the corresponding features output by the pooling layer. Finally, each feature is input into a fully connected layer, which integrates multiple features to obtain the node features and connection features in the network topology graph. This embodiment of the invention, by dividing the topology matrix into multiple sub-matrices and performing feature extraction on the sub-matrices on different channels, enriches the acquisition of feature information, thereby improving the accuracy of feature extraction.

[0041] 207. Based on node features and connection features, identify nodes, connections between nodes, connection relationships between nodes, and node attribute information in a network topology graph.

[0042] In this embodiment of the invention, after extracting node features and connection features, it is necessary to identify information such as nodes, connections between nodes, connection relationships between nodes, and node attribute information based on the above features. Based on this, step 207 specifically includes: extracting local and global features of the node features to obtain corresponding node local features and node global features; extracting local and global features of the connection features to obtain corresponding connection local features and connection global features; fusing the node local features and the node global features to obtain node fusion features; fusing the connection local features and the connection global features to obtain connection fusion features; and inputting the node fusion features and the connection fusion features into a preset node recognition model to obtain the nodes, connections between nodes, and node attribute information in the network topology diagram.

[0043] Specifically, to improve the recognition accuracy of the preset node recognition model, it is first necessary to construct the preset node recognition model. Based on this, the method includes: constructing a preset initial node recognition model; obtaining a sample dataset, wherein the sample dataset includes sample node fusion features and sample node connection features of a sample power communication network topology map with node, node connection, and node attribute information labels; dividing the sample dataset into a training set and a test set; training the preset initial node recognition model using the training set; and testing the trained preset initial node recognition model using the test set; finally, the trained preset initial node recognition model that meets the testing conditions is taken as the preset node recognition model. Specifically, in the model training process, the preset initial node recognition model is first constructed, and then the sample dataset is obtained. Ensure that the dataset contains all necessary files. Convert the data to a format that the preset initial node recognition model can understand, and finally train and test the model. Specifically, the dataset can be divided first: using random or specific strategies (such as stratified sampling) to divide the sample dataset into a training set and a test set. Then, the model is trained using the training set, and the trained model is tested using the test set to evaluate its performance on unseen data. Calculate and record metrics such as mCP, precision, and recall on the test set. If the model performance does not meet the requirements, return to the training phase for further iterations or adjustments. This process yields a pre-defined node recognition model that meets the requirements.

[0044] Furthermore, local information includes the specific characteristics of nodes and the attributes of adjacent connections, while global information reflects the structure and connection patterns of the entire topology. For example, for the node and connection features within a local region, the model can learn its local features to identify the specific structure of that region; while for the entire topology, the model can analyze the global distribution and connection relationships of nodes and connections to determine its overall topology type. This approach, which considers both local and global information, can improve the model's ability to recognize topologies of different scales and complexities. Feature extraction models such as encoders and decoders can be used to extract local and global features of node and connection features. Then, a fusion operation, such as horizontal stitching, is performed on the corresponding local and global features to obtain node fusion features and connection fusion features. Finally, the node fusion features and connection fusion features are directly input into a preset node recognition model, which can directly obtain information about nodes, connections between nodes, and node attributes in the network topology. This embodiment of the invention improves the model's ability to recognize topologies of different scales and complexities by considering both local and global information simultaneously.

[0045] Furthermore, after identifying the nodes, connections between nodes, and node attribute information in the network topology graph, it is also necessary to identify the connection relationships between nodes in the network topology graph. Based on this, step 207 further includes: taking any node in each node of the network topology graph as a target node, determining multiple neighbor nodes corresponding to the target node in the network topology graph; determining the association weight of the target node relative to each of the neighbor nodes, taking the neighbor nodes whose association weight is greater than a preset weight threshold as target neighbor nodes, and aggregating the node features of the target neighbor nodes to the node features of the target node to obtain the aggregated features of the target node; inputting the aggregated features of each node in the network topology graph and the connection features into a preset topology graph recognition model to identify the connection relationships, thereby obtaining the connection relationships between nodes in the network topology graph.

[0046] The preset weight threshold is set according to actual needs. To improve the recognition accuracy of the preset topology graph recognition model, it is first necessary to train and construct the preset topology graph recognition model. Based on this, the method includes: constructing a preset initial topology graph recognition model; obtaining a sample dataset, wherein the sample dataset includes the aggregate features of each node in a sample power communication network topology graph with labels of connection relationships between nodes; dividing the sample dataset into a training set and a test set, using the training set to train the preset initial topology graph recognition model, and using the test set to test the trained preset initial topology graph recognition model, and finally using the trained preset initial topology graph recognition model that meets the test conditions as the preset topology graph recognition model.

[0047] Furthermore, taking a node in the network topology graph as the target node as an example, the association weight of the target node relative to each of its neighboring nodes is calculated based on physical parameters such as link bandwidth, delay, and optical attenuation between the target node and each of its corresponding neighboring nodes. Within each neighboring node, target neighboring nodes with association weights greater than a preset weight threshold are identified. Based on the association weights of the target neighboring nodes, the node features of each target neighboring node are weighted and summed, and the weighted summation result is fused with the node features of the target node to obtain the aggregated features of the target node. Thus, the aggregated features of each node in the network topology graph can be determined as described above. Subsequently, the aggregated features of each node can be directly output as the connection relationships between nodes in the network topology graph. This embodiment of the invention enhances the feature representation of nodes by aggregating node features, thereby improving the accuracy of identifying the connection relationships between points.

[0048] 208. Based on nodes, connections between nodes, connection relationships between nodes, and node attribute information, search for the optimal power communication path from the starting node to the ending node in the network topology graph, and use the optimal power communication path as the power communication path of the target power communication network.

[0049] In this embodiment of the invention, after identifying the nodes, connections between nodes, connection relationships between nodes, and node attribute information in the network topology diagram, it is necessary to optimize the power communication path based on the above information. Therefore, step 208 specifically includes: determining multiple candidate power communication paths from the starting node to the ending node in the network topology diagram based on the connections between nodes in the network topology diagram, and determining the link information between each node on each candidate power communication path, wherein the link information includes at least one of load and failure rate; and selecting the optimal power communication path on each candidate power communication path based on the node attribute information of each node on each candidate power communication path, the link information between nodes, and the connection relationships between nodes. The method for selecting the optimal power communication path on each candidate power communication path includes: for each node in each candidate power communication path, determining the attribute feature vector corresponding to the node attribute information, the link feature vector corresponding to the link information, and the connection feature vector corresponding to the connection relationship; performing cross processing on the attribute feature vector, the link feature vector, and the connection feature vector to obtain a path cross feature vector; and inputting the path cross feature vector of each candidate power communication path into a preset path prediction model for path prediction to obtain the optimal power communication path.

[0050] Specifically, firstly, multiple candidate power communication paths are determined using the starting node and the path's beginning point, and the ending node as the path's ending point. Taking a candidate power communication path as an example, a feature extraction network (such as a CNN) is used to extract the attribute feature vectors corresponding to the node attribute information, the link feature vectors corresponding to the link information, and the connection feature vectors corresponding to the connection relationships for each node on the candidate power communication path. Then, the attribute feature vectors, link feature vectors, and connection feature vectors of the same node are cross-processed to obtain the path cross feature vector for each node. These path cross feature vectors are then fused to obtain the path fused feature vector for the candidate power communication path. Finally, the path fused feature vector is directly input into a preset path prediction model for path prediction to obtain the optimal power communication path. The preset path prediction model is trained using a dataset consisting of path fused feature vectors corresponding to sample candidate paths labeled with power communication paths.

[0051] Specifically, the method for cross-processing the attribute feature vector, link feature vector, and connection feature vector includes: performing feature-level cross-processing on the attribute feature vector, link feature vector, and connection feature vector to obtain a feature cross vector; performing element-level cross-processing on the attribute feature vector, link feature vector, and connection feature vector to obtain an element cross vector; performing low-order cross-processing on the attribute feature vector, link feature vector, and connection feature vector to obtain a low-order cross vector; and combining the feature cross vector, element cross vector, and low-order cross vector to obtain the trend cross feature vector.

[0052] Specifically, to fully utilize the relationships between data, extract more latent features, and simultaneously handle both high-order and low-order processing to make data utilization more efficient and subsequent prediction results more accurate, thus meeting the needs of practical application scenarios, it is necessary to perform cross-processing on attribute feature vectors, link feature vectors, and connection feature vectors. The specific cross-processing method is as follows: if the attribute feature vector is (a1, a2), the link feature vector is (b1, b2), and the connection feature vector is (c1, c2), the specific cross-processing methods include: performing feature-level cross-processing between different feature vectors, i.e., performing a Hadamard product on all elements of the vectors, followed by a convolution transformation under a certain weight w1, to obtain the feature cross vector f(w1 × (a1 × b1 × c1, a2 × b2 × c2)); simultaneously, performing element-level cross-processing on all feature vector data, i.e., performing a Hadamard product on each element of the vectors, assigning different weight values ​​w2 and w3 to each product result, and then performing a linear transformation, to obtain the element cross vector f(w2 × a1 × b1 × c1, w3 × a2 × b2 × c2)). c2); In addition, all feature vectors are subjected to low-order cross processing, and then the result of the cross processing is assigned a weight coefficient w4, and then a linear transformation is performed to obtain the low-order cross vector f(w4(a1,a2,b1,b2,c1,c2)); Finally, the above feature cross vectors, element cross vectors, and low-order cross vectors are combined, such as by horizontal concatenation, to obtain the trend cross feature vector. It should be noted that the above examples are only illustrative and do not limit the embodiments of this application. Thus, by performing cross processing on attribute feature vectors, link feature vectors, and connection feature vectors, different features can be automatically or explicitly combined to generate new feature combinations. These combined features may contain complex nonlinear relationships between the original features, enabling the model to capture more refined and richer information in the data, that is, to make full use of the relationships between various data, extract more latent features, and take into account both high-order and low-order processing, so that the data is utilized more fully and the subsequent communication path determination results are more accurate, meeting the needs of practical application scenarios.

[0053] According to another method for determining power communication paths provided by the present invention, compared with the current method of using the shortest path in the power communication network as the power communication path, the present invention identifies the nodes, the connections between nodes, the connection relationships between nodes, and the node attribute information by identifying the network topology of the power communication network. Then, it determines the optimal power communication path by comprehensively analyzing the nodes, the connections between nodes, the connection relationships between nodes, and the node attribute information. This ensures that the determined optimal power communication path can guarantee communication stability and smoothness while satisfying the requirement of reasonable path length, thereby improving the accuracy of power communication path determination and enhancing the utilization rate of communication resources.

[0054] Furthermore, as Figure 1 In specific implementation, embodiments of the present invention provide a device for determining power communication paths, such as... Figure 3 As shown, the device includes: an acquisition unit 31, an identification unit 32, and a search unit 33.

[0055] The acquisition unit 31 can be used to acquire a network topology map of the target power communication network in response to a determination signal for a power communication path of the target power communication network, and determine the starting node and ending node of the power communication path to be determined in the network topology map.

[0056] The identification unit 32 can be used to identify the network topology map and obtain the identification result, wherein the identification result includes the nodes of the network topology map, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of bandwidth, latency, packet loss rate, and device type.

[0057] The search unit 33 can be used to search for the optimal power communication path from the starting node to the ending node in the network topology map based on the identification result, and use the optimal power communication path as the power communication path of the target power communication network.

[0058] In specific application scenarios, in order to identify network topology maps, such as... Figure 4 As shown, the recognition unit 32 includes a conversion module 321, an encoding module 322, a splitting module 323, an acquisition module 324, a feature extraction module 325, and a recognition module 326.

[0059] The conversion module 321 can be used to convert the network topology graph into a topology matrix, wherein the nodes in the network topology graph are mapped to the corresponding pixels in the topology matrix, and the lines between the nodes in the network topology graph are mapped to the connection paths between each pixel in the topology matrix.

[0060] The encoding module 322 can be used to encode the pixels and connection paths in the topology matrix using a preset space filling curve to obtain an encoded topology matrix.

[0061] The splitting module 323 can be used to split the encoded topology matrix into a multi-channel topology matrix, wherein the multi-channel topology matrix includes a node connection representation matrix of nodes and connections in the network topology graph, a color information representation matrix of the network topology graph, and a depth information representation matrix of the network topology graph.

[0062] The acquisition module 324 can be used to acquire a preset feature extraction model, wherein the preset feature extraction model includes a convolutional layer for parallel feature extraction, a batch normalization layer for feature normalization, a pooling layer for feature pooling, and a fully connected layer for feature integration.

[0063] The feature extraction module 325 can be used to input the node connection representation matrix, the color information representation matrix, and the depth information representation matrix into the preset feature extraction model, and perform parallel feature extraction on the node connection representation matrix, the color information representation matrix, and the depth information representation matrix in corresponding channels through the convolutional layer to obtain parallel features of different channels. The parallel features are then processed sequentially through the batch normalization layer and the pooling layer to obtain the parallel pooling features output by the pooling layer. The parallel pooling features are then input into the fully connected layer for feature integration to obtain node features and connection features in the network topology graph.

[0064] The identification module 326 can be used to identify nodes, connections between nodes, connection relationships between nodes, and node attribute information in the network topology based on the node features and the connection features.

[0065] In specific application scenarios, in order to identify nodes, connections between nodes, and node attribute information in a network topology graph, the identification module 326 can be used to extract local and global features of the node features, corresponding to obtain node local features and node global features; extract local and global features of the connection features, corresponding to obtain connection local features and connection global features; perform fusion processing on the node local features and the node global features to obtain node fusion features; perform fusion processing on the connection local features and the connection global features to obtain connection fusion features; input the node fusion features and the connection fusion features into a preset node identification model to obtain the nodes, connections between nodes, and node attribute information in the network topology graph.

[0066] In specific application scenarios, in order to identify the connection relationships between nodes in a network topology graph, the identification module 326 can be specifically used to take any node in each node of the network topology graph as a target node, determine multiple neighbor nodes corresponding to the target node in the network topology graph; determine the association weight of the target node relative to each of the neighbor nodes, take the neighbor nodes whose association weight is greater than a preset weight threshold as target neighbor nodes, and aggregate the node features of the target neighbor nodes to the node features of the target node to obtain the aggregated features of the target node; input the aggregated features of each node in the network topology graph and the connection features into a preset topology graph identification model to identify the connection relationships, thereby obtaining the connection relationships between nodes in the network topology graph.

[0067] In specific application scenarios, in order to search for the optimal power communication path from the starting node to the ending node in the network topology graph, the search unit 33 includes a partitioning module 331, a determination module 332, an addition module 333, an evaluation module 334, a transfer module 335, and a search module 336.

[0068] The partitioning module 331 can be used to partition the network topology map into regions to obtain network topology maps corresponding to different regions.

[0069] The determining module 332 can be used to determine different preset heuristic functions for network topology maps corresponding to different regions, wherein the preset heuristic function is a function constructed from at least distance and time.

[0070] The determining module 332 can also be used to take any region among the different regions as a target region, take the network topology graph corresponding to the target region as a target network topology graph, and take the preset heuristic function corresponding to the target network topology graph as a preset target heuristic function.

[0071] The adding module 333 can be used to determine a preset open list and a preset closed list, and add the starting node in the target network topology diagram to the preset open list.

[0072] The evaluation module 334 can be used to evaluate the estimated cost from the starting node to the ending node in the target network topology using the preset target heuristic function, and to determine the actual cost between the starting node and each target neighbor node in the target network topology.

[0073] The adding module 333 can also be used to determine the estimated priority of the target neighbor node based on the estimated cost, determine the actual priority of the target neighbor node based on the actual cost, select a high-quality neighbor node from the target neighbor nodes based on the estimated priority, and add the high-quality neighbor node to the preset open list.

[0074] The transfer module 335 can be used to determine the total priority of each high-quality neighbor node in the preset open list based on the estimated priority and actual priority of the high-quality neighbor nodes, and transfer the optimal neighbor node corresponding to the maximum total priority to the preset closed list.

[0075] The search module 336 can be used to take the optimal neighbor node corresponding to the highest total priority in the preset closed list as the new starting node, take the target neighbor node adjacent to the new starting node as the new target neighbor node, and iteratively search for new high-quality neighbor nodes that meet the cost requirements from the starting node to the ending node of the target network topology based on the new starting node, the new target neighbor node, and the preset target heuristic function. The new optimal neighbor node is then transferred to the preset closed list until the ending node of the target network topology is found. Finally, the optimal power communication path is formed based on the starting node, the ending node of the target network topology, and each optimal neighbor node in the preset closed list.

[0076] In specific application scenarios, in order to determine the optimal power communication path, the determining module 332 can also be used to determine multiple candidate power communication paths from the starting node to the ending node in the network topology diagram based on the connections between nodes in the network topology diagram, and determine the link information between each node on each candidate power communication path, wherein the link information includes at least one of load and failure rate; and select the optimal power communication path on each candidate power communication path based on the node attribute information of each node on each candidate power communication path, the link information between nodes, and the connection relationship between nodes.

[0077] In specific application scenarios, in order to determine the optimal power communication path, the determining module 332 can be used to determine, for each node in each candidate power communication path, the attribute feature vector corresponding to the node attribute information, the link feature vector corresponding to the link information, and the connection feature vector corresponding to the connection relationship; perform cross processing on the attribute feature vector, the link feature vector, and the connection feature vector to obtain a path cross feature vector; and input the path cross feature vector of each candidate power communication path into a preset path prediction model for path prediction to obtain the optimal power communication path.

[0078] It should be noted that other corresponding descriptions of the functional modules involved in the power communication path determination device provided in this embodiment of the invention can be found in [reference]. Figure 1 The corresponding description of the method shown will not be repeated here.

[0079] Based on the above, Figure 1 Accordingly, this embodiment of the invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the following steps: In response to a determination signal for a power communication path in a target power communication network, acquiring a network topology map of the target power communication network, and determining the starting node and ending node of the power communication path to be determined in the network topology map; identifying the network topology map to obtain an identification result, wherein the identification result includes nodes of the network topology map, connections between nodes, connection relationships between nodes, and node attribute information, where nodes are communication devices in the power communication network, connections between nodes are transmission media between communication devices, connection relationships between nodes are communication interaction rules of the transmission media, and node attribute information includes at least one of bandwidth, latency, packet loss rate, device type, and carried communication volume; based on the identification result, searching for an optimal power communication path from the starting node to the ending node in the network topology map, and using the optimal power communication path as the power communication path of the target power communication network.

[0080] Based on the above, Figure 1 The method shown and as Figure 3 The embodiment of the device shown in the invention also provides a physical structure diagram of a computer device, such as... Figure 5As shown, the computer device includes: a processor 41, a memory 42, and a computer program stored in the memory 42 and executable on the processor. Both the memory 42 and the processor 41 are mounted on a bus 43. When the processor 41 executes the program, it performs the following steps: In response to a signal indicating the determination of a power communication path for a target power communication network, it acquires a network topology map of the target power communication network and determines the starting and ending nodes of the power communication path to be determined in the network topology map; it identifies the network topology map to obtain an identification result, wherein the identification result includes nodes of the network topology map, connections between nodes, connection relationships between nodes, and node attribute information. Nodes are communication devices in the power communication network, connections between nodes are transmission media between communication devices, connection relationships between nodes are communication interaction rules of the transmission media, and node attribute information includes at least one of bandwidth, latency, packet loss rate, device type, and carried communication volume; based on the identification result, it searches for the optimal power communication path from the starting node to the ending node in the network topology map and uses the optimal power communication path as the power communication path of the target power communication network.

[0081] Through the technical solution of this invention, the present invention identifies nodes, connections between nodes, connection relationships between nodes, and node attribute information by identifying the network topology of the power communication network. It then determines the optimal power communication path by comprehensively analyzing the nodes, connections between nodes, connection relationships between nodes, and node attribute information. This ensures that the determined optimal power communication path can guarantee communication stability and smoothness while meeting the requirement of reasonable path length, thereby improving the accuracy of power communication path determination and enhancing the utilization rate of communication resources.

[0082] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in an order not used herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.

[0083] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for determining a power communication path, characterized in that, include: In response to a signal indicating the determination of a power communication path for a target power communication network, a network topology map of the target power communication network is acquired, and the starting node and ending node of the power communication path to be determined are determined in the network topology map. The network topology is identified to obtain an identification result, wherein the identification result includes the nodes of the network topology, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of bandwidth, latency, packet loss rate, device type, and the amount of communication carried. Based on the identification results, the optimal power communication path from the starting node to the ending node is searched in the network topology map, and the optimal power communication path is taken as the power communication path of the target power communication network.

2. The method according to claim 1, characterized in that, The process of identifying the network topology to obtain the identification result includes: The network topology graph is converted into a topology matrix, wherein the nodes in the network topology graph are mapped to the corresponding pixels in the topology matrix, and the lines between the nodes in the network topology graph are mapped to the connection paths between each pixel in the topology matrix. The pixels and connection paths in the topology matrix are encoded using a preset space-filling curve to obtain an encoded topology matrix; The encoded topology matrix is ​​split into a multi-channel topology matrix, wherein the multi-channel topology matrix includes a node connection representation matrix of nodes and connections in the network topology graph, a color information representation matrix of the network topology graph, and a depth information representation matrix of the network topology graph. Obtain a preset feature extraction model, wherein the preset feature extraction model includes a convolutional layer for parallel feature extraction, a batch normalization layer for feature normalization, a pooling layer for feature pooling, and a fully connected layer for feature integration. The node connection representation matrix, the color information representation matrix, and the depth information representation matrix are input into the preset feature extraction model. The convolutional layer performs parallel feature extraction on the node connection representation matrix, the color information representation matrix, and the depth information representation matrix in corresponding channels to obtain parallel features for different channels. The parallel features are then processed sequentially through the batch normalization layer and the pooling layer to obtain the parallel pooling features output by the pooling layer. The parallel pooling features are then input into the fully connected layer for feature integration to obtain node features and connection features in the network topology graph. Based on the node features and the connection features, the nodes, connections between nodes, connection relationships between nodes, and node attribute information in the network topology graph are identified.

3. The method according to claim 2, characterized in that, Based on the node features and the connection features, the system identifies nodes, connections between nodes, and node attribute information in the network topology graph, including: Extract the local and global features of the node features to obtain the corresponding node local features and node global features; extract the local and global features of the connection features to obtain the corresponding connection local features and connection global features. The local features and global features of the nodes are fused to obtain node fused features, and the local features and global features of the connections are fused to obtain connection fused features; The node fusion features and the connection fusion features are input into a preset node recognition model to obtain the nodes, connections between nodes, and node attribute information in the network topology diagram.

4. The method according to claim 2, characterized in that, Based on the node features and the connection features, the connection relationships between nodes in the network topology graph are identified, including: Each node in the network topology graph is taken as a target node, and multiple neighboring nodes corresponding to the target node are determined in the network topology graph. The association weight of the target node relative to each of the neighboring nodes is determined respectively. The neighboring nodes whose association weight is greater than a preset weight threshold are taken as the target neighbor nodes. The node features of the target neighbor nodes are aggregated to the node features of the target node to obtain the aggregated features of the target node. The aggregation features of each node in the network topology graph and the connection features are input into a preset topology graph recognition model to identify the connection relationships, thereby obtaining the connection relationships between nodes in the network topology graph.

5. The method according to claim 1, characterized in that, Based on the identification results, the optimal power communication path from the starting node to the ending node is searched in the network topology graph, including: The network topology map is divided into regions to obtain network topology maps corresponding to different regions; Different preset heuristic functions are determined for the network topology maps corresponding to different regions, wherein the preset heuristic function is a function constructed from at least distance and time; Take any region among the different regions as a target region, take the network topology graph corresponding to the target region as the target network topology graph, and take the preset heuristic function corresponding to the target network topology graph as the preset target heuristic function. Determine a preset open list and a preset closed list, and add the starting node in the target network topology diagram to the preset open list; The preset target heuristic function is used to evaluate the estimated cost from the starting node to the ending node in the target network topology graph, and the actual cost between the starting node and each target neighbor node in the target network topology graph is determined. The estimated priority of the target neighbor node is determined based on the estimated cost, the actual priority of the target neighbor node is determined based on the actual cost, and a high-quality neighbor node is selected from the target neighbor nodes based on the estimated priority and added to the preset open list. Based on the estimated and actual priorities of the high-quality neighbor nodes, the total priority of each high-quality neighbor node in the preset open list is determined, and the optimal neighbor node corresponding to the maximum total priority is transferred to the preset closed list. The optimal neighbor node corresponding to the highest total priority in the preset closed list is taken as the new starting node, and the target neighbor node adjacent to the new starting node is taken as the new target neighbor node. Based on the new starting node, the new target neighbor node, and the preset target heuristic function, a new high-quality neighbor node that meets the cost requirements from the starting node to the ending node of the target network topology is searched iteratively. The new optimal neighbor node is then moved to the preset closed list until the ending node of the target network topology is found. Finally, the optimal power communication path is formed based on the starting node, the ending node of the target network topology, and each optimal neighbor node in the preset closed list.

6. The method according to claim 1, characterized in that, Based on the identification results, the optimal power communication path from the starting node to the ending node is searched in the network topology graph, including: Based on the connections between nodes in the network topology diagram, multiple candidate power communication paths are determined from the starting node to the ending node in the network topology diagram, and link information between each node on each candidate power communication path is determined, wherein the link information includes at least one of load and failure rate. Based on the node attribute information, link information between nodes, and connection relationships between nodes of each candidate power communication path, the optimal power communication path is selected on each candidate power communication path.

7. The method according to claim 6, characterized in that, The step of selecting the optimal power communication path on each candidate power communication path based on the node attribute information of each node on each candidate power communication path, the link information between nodes, and the connection relationship between nodes includes: For each node in each of the candidate power communication paths, the attribute feature vector corresponding to the node attribute information, the link feature vector corresponding to the link information, and the connection feature vector corresponding to the connection relationship are determined respectively. The attribute feature vector, the link feature vector, and the connection feature vector are cross-processed to obtain the path cross feature vector; The path intersection feature vector of each candidate power communication path is input into a preset path prediction model to predict the path and obtain the optimal power communication path.

8. A device for determining a power communication path, characterized in that, include: The acquisition unit is configured to acquire a network topology map of the target power communication network in response to a determination signal for a power communication path of the target power communication network, and determine the starting node and ending node of the power communication path to be determined in the network topology map. The identification unit is used to identify the network topology map and obtain the identification result, wherein the identification result includes the nodes of the network topology map, the connections between the nodes, the connection relationships between the nodes, and the node attribute information. The nodes are communication devices in the power communication network, the connections between the nodes are the transmission media between the communication devices, the connection relationships between the nodes are the communication interaction rules of the transmission media, and the node attribute information includes at least one of bandwidth, latency, packet loss rate, and device type. The search unit is used to search for the optimal power communication path from the starting node to the ending node in the network topology map based on the identification result, and to use the optimal power communication path as the power communication path of the target power communication network.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images