Transmission resource allocation optimization method, system, device, medium and product

Abstract

This invention discloses a method, system, device, medium, and product for optimizing transmission resource allocation. The method determines the state characteristics of each modal data under the current device state by using the modal feature vectors of multimodal data transmitted in distribution network equipment. It then summarizes the state characteristics of each modal data under the current device state to obtain the overall state of the current equipment. Based on the overall state of the equipment, it determines the current dominant data path and predicts the traffic demand and dominant path probability at a preset future time. The optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment. Based on the information integrity of each modal data as a constraint, it obtains the optimal compression rate for each data channel. Based on the optimal compression rate of each data channel, it determines the traffic flow of the data channel after compression and allocates bandwidth priority in conjunction with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, system, device, medium and product for optimizing transmission resource allocation. Background Technology

[0002] With the deepening of smart grid construction, distribution network equipment condition monitoring technology has become an important means to ensure the safe and stable operation of the power system. Modern distribution network monitoring systems widely adopt multimodal sensing technology, deploying various sensors such as infrared thermal imaging and acoustic fingerprinting to achieve comprehensive perception of the operating status of key equipment such as transformers, circuit breakers, and switchgear. This multimodal data contains rich equipment health information, reflecting the operating characteristics and potential fault signs of equipment from different physical dimensions. However, multimodal monitoring data is characterized by large data volume, high real-time requirements, and dynamically changing transmission demands, posing severe challenges to the data processing capabilities of edge computing nodes and the transmission resources of communication networks. In particular, during the evolution of equipment faults, the importance of different modal data exhibits a dynamic changing pattern; infrared images are most critical at certain times, while acoustic fingerprints or vibration signals dominate at other times. How to accurately predict the changing trends of multimodal data transmission demands under limited network bandwidth conditions and rationally allocate transmission resources to ensure high-quality transmission of critical information has become a core problem that urgently needs to be solved in the field of distribution network smart monitoring, directly affecting the timeliness and accuracy of fault early warning.

[0003] Existing multimodal data transmission methods for power distribution networks suffer from the following limitations. First, traditional methods often employ static bandwidth allocation strategies, pre-allocating resources based on the average flow demand of each modality. This fails to fully consider the dynamic switching characteristics of the dominant mode during equipment state evolution, leading to insufficient bandwidth on the dominant path and idle resources on non-critical paths at critical fault moments, resulting in a contradiction between transmission delay and resource waste. Second, existing compression transmission schemes typically use a uniform compression rate configuration for all modal data, ignoring the differences in characteristics between different modal data and the information redundancy relationships between modes. This fails to guarantee the information quality of the dominant path and also fails to fully utilize the compression potential of non-dominant paths, resulting in low overall bandwidth utilization efficiency. Third, there is a lack of a proactive prediction mechanism for multimodal data transmission demands. The system can only passively adjust based on the current flow status, unable to anticipate upcoming dominant path switching and flow peaks. This causes resource allocation strategies to lag behind actual demand changes, easily leading to the loss of critical information or transmission interruptions in scenarios with rapid fault evolution, severely impacting the accuracy and real-time performance of fault detection. Summary of the Invention

[0004] In view of this, in order to solve the above-mentioned technical problems, the present invention provides a method, system, device, medium and product for optimizing transmission resource allocation.

[0005] The first aspect of this invention provides a method for optimizing transmission resource allocation, comprising:

[0006] Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state;

[0007] The overall state of the current device is obtained by summarizing the state characteristics of each modal data under the current device state;

[0008] The current dominant data path is determined based on the overall status of the device, and the traffic demand and the probability of the dominant path at a future preset time are predicted based on the modal feature vectors of each modal data and the current dominant data path.

[0009] Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and a compression ratio optimization allocation model is constructed according to the information integrity of each modal data as a constraint.

[0010] The compression ratio optimization allocation model is optimized and solved, and the optimal compression ratio of each data channel is obtained based on the optimal solution;

[0011] Based on the optimal compression rate of each data channel, the compressed traffic of the data channel is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution to obtain the bandwidth allocation result of each data channel.

[0012] Preferably, the step of comparing the similarity between each modal feature vector and each preset typical state feature to determine the state features of each modal data under the current device state includes:

[0013] Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector with each preset typical state feature;

[0014] For each of the preset typical state features, the similarity between each modal feature vector and the preset typical state feature is determined based on the weighted result of the feature proximity and change trend between each modal feature vector and the preset typical state feature.

[0015] Based on the similarity, a preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

[0016] Preferably, the step of determining the current dominant data path based on the overall state of the device, and predicting the traffic demand and dominant path probability at a future preset time based on the modal feature vectors of each modal data and the current dominant data path, includes:

[0017] Based on the overall state of the equipment, an expert experience-led path matching the overall state of the equipment is determined. If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path.

[0018] If the similarity between the overall state of the device and the state feature corresponding to the expert experience-dominated path is not greater than a preset similarity threshold, then the data channel with the highest similarity to the overall state of the device is selected from all data channels as the current dominant data channel.

[0019] Based on the modal feature vectors of each modal data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

[0020] Preferably, the objective function corresponding to the optimization objective is:

[0021]

[0022] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. Let Mi be the compression ratio vector of the data channel at time T. For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

[0023] Preferably, the constraints include information loss limitation constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limitation constraints; wherein, the information loss limitation constraints are used to ensure that the information loss degree of each data channel does not exceed the information loss limitation constraints, the dominant path probability adjustment information loss constraints are used to dynamically adjust the upper bound of the information loss degree of non-dominant paths according to the dominant path probability, the inter-modal deducibility control information loss constraints are used to constrain the overall transmission traffic using inter-modal data deducibility, and the compression rate limitation constraints are used to ensure that the compression rate of each data channel is within the allowable range.

[0024] Preferably, the step of determining the compressed traffic of each data channel based on its optimal compression ratio, and allocating bandwidth priority in conjunction with the dominant path probability distribution to obtain the bandwidth allocation result for each data channel includes:

[0025] Based on the optimal compression ratio of each data channel, determine the compressed data flow rate of each data channel;

[0026] Based on the principle of prioritizing dominant paths, and combining the compressed traffic of each data channel with the probability distribution of the dominant path, the bandwidth priority of each data channel is sorted so that the data channel with the higher probability of the dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result of each data channel.

[0027] Secondly, the present invention also provides a transmission resource allocation optimization system, comprising:

[0028] The state feature determination module is used to acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state.

[0029] The device status determination module is used to summarize the status characteristics of each modal data under the current device status to obtain the overall status of the current device.

[0030] The data prediction module is used to determine the current dominant data path based on the overall status of the device, and to predict the traffic demand and the probability of the dominant path at a future preset time based on the modal feature vectors of each modal data and the current dominant data path.

[0031] The allocation model construction module is used to construct a compression ratio optimization allocation model based on the predicted traffic demand and the probability of the dominant path, with the optimization objective of minimizing the total transmission traffic of multimodal data in the distribution network equipment, and with the information integrity of each modal data as a constraint.

[0032] The model solving module is used to find the optimal solution for the compression ratio optimization allocation model and obtain the optimal compression ratio for each data channel based on the optimal solution.

[0033] The bandwidth allocation module is used to determine the compressed traffic of each data channel based on the optimal compression rate of each data channel, and to allocate bandwidth priority in combination with the dominant path probability distribution to obtain the bandwidth allocation result of each data channel.

[0034] Thirdly, the present invention also provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the transmission resource allocation optimization method as described in the first aspect.

[0035] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed, implements the steps of the transmission resource allocation optimization method as described in the first aspect.

[0036] Fifthly, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein, when the program instructions are executed by a computer, the computer performs the steps of the transmission resource allocation optimization method as described in the first aspect.

[0037] As can be seen from the above technical solutions, this invention determines the state characteristics of each modal data under the current equipment state by using the modal feature vector of multimodal data transmitted in the distribution network equipment, and summarizes the state characteristics of each modal data under the current equipment state to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient utilization of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is an application environment diagram of a transmission resource allocation optimization method provided in an embodiment of the present invention;

[0040] Figure 2 A flowchart of a transmission resource allocation optimization method provided in an embodiment of the present invention;

[0041] Figure 3 This is a schematic diagram of a transmission resource allocation optimization system provided in an embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0043] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] The transmission resource allocation optimization method provided in this application embodiment can be applied to, for example, Figure 1The application environment shown is illustrated. Terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be integrated onto server 102, or it can be located in the cloud or on another network server. Terminal 101 or server 102 acquires multimodal data transmitted in the distribution network equipment, extracts the modal feature vectors of each modal data, compares the similarity of each modal feature vector with each preset typical state feature, and determines the state features of each modal data under the current equipment state; summarizes the state features of each modal data under the current equipment state to obtain the overall state of the current equipment; determines the current dominant data path based on the overall state of the equipment, and predicts the traffic demand and dominant path probability at a preset time in the future based on the modal feature vectors of each modal data and the current dominant data path; based on the predicted traffic demand and dominant path probability, with the optimization objective of minimizing the total transmission traffic of multimodal data in the distribution network equipment, and with the information integrity of each modal data as a constraint, constructs a compression ratio optimization allocation model; optimizes and solves the compression ratio optimization allocation model, and obtains the optimal compression ratio of each data channel based on the optimal solution; determines the traffic of the data channel after compression based on the optimal compression ratio of each data channel, and performs bandwidth priority allocation in combination with the dominant path probability distribution to obtain the bandwidth allocation result of each data channel.

[0045] Terminal 101 can be, but is not limited to, various personal computers, laptops, smartphones, and tablets.

[0046] Server 102 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides cloud computing services.

[0047] like Figure 2 As shown in the embodiments of this application, a method for optimizing transmission resource allocation is provided, which is applied to... Figure 1 Taking terminal 101 or server 102 as an example, the explanation includes the following steps S1 to S6. Wherein:

[0048] Step S1: Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity between each modal feature vector and each preset typical state feature, and determine the state features of each modal data under the current equipment state.

[0049] Multimodal data includes heterogeneous data from multiple sources such as images, sound patterns, temperature, current, and voltage from power distribution network equipment. It is collected in real time by sensors and transmitted to edge nodes.

[0050] Edge nodes extract features from the multimodal data channels of each terminal. Assume there are... Each electrical device Equipped One monitoring terminal. The edge computing gateway is in constant... For electrical equipment The Feature extraction is performed on the data uploaded by each monitoring terminal, and the extracted modal feature vector is represented as follows:

[0051] In the formula, For equipment The The raw data from each data channel; For the first Feature extraction function for each data channel; The extracted feature vector; and These are the original data dimension and the feature dimension, respectively.

[0052] The preset typical state features are feature vectors from a library of typical fault modes built based on expert experience, including normal operation, mechanical faults (such as jamming of the operating mechanism), thermal faults (such as overheating of contacts), and insulation faults (such as partial discharge). By determining the similarity between each modal feature vector and each preset typical state feature, the state characteristics of each modal data under the current equipment state are determined, such as normal operation, mechanical faults (such as jamming of the operating mechanism), thermal faults (such as overheating of contacts), and insulation faults (such as partial discharge).

[0053] Step S2: Summarize the state characteristics of each modal data under the current device state to obtain the overall state of the current device.

[0054] If the status characteristics of all data channels are consistent, that is, the device... All All data channels make the same judgment about the state, for example, state characteristics. ,Right now For all If established, then the current overall status of the equipment is as follows: .

[0055] If the states determined by different data channels are inconsistent, the overall state of the device is determined by summarizing the similarity scores, as follows:

[0056]

[0057] In the formula, For indicator functions, The overall state of the device is determined by counting the frequency of each state being determined across all data channels, and selecting the state with the highest frequency as the current overall state of the device. If multiple states have the same frequency, the state with the largest weighted similarity is selected as the final determination result.

[0058] Step S3: Determine the current dominant data path based on the overall status of the equipment, and predict the traffic demand and dominant path probability at a future preset time based on the modal feature vectors of each modal data and the current dominant data path.

[0059] The dominant data path refers to the data mode that most effectively monitors and characterizes the features of power distribution network equipment under different operating conditions. Its selection is closely related to the current state of the equipment and the evolution of the fault. For example, under normal operating conditions of switchgear, the image mode is the dominant path because it can visually display the appearance of the equipment and the status of indicator lights. When the monitoring device detects abnormal noises from the circuit breaker, such as slight friction or impact sounds, but no obvious changes in appearance, the acoustic fingerprint mode becomes the dominant path because acoustic fingerprint characteristics can reveal early signs of mechanical faults. If partial discharge occurs in the contacts accompanied by flashing lights, the image mode becomes dominant again, allowing direct observation of the discharge phenomenon and fault point, providing the most critical diagnostic information.

[0060] Then, based on the modal feature vectors of each modality and the current dominant data path as historical sequence inputs, a multi-layer temporal attention network is used to dynamically weight and fuse the features of each modality, capturing the dependencies and evolution patterns between modalities at different time steps. By introducing a time-aware attention mechanism, the model can focus on the changing trends of the dominant modality at key time nodes, predict the intensity of traffic demand and the probability distribution of the dominant path at multiple future moments, thereby providing a precise basis for the dynamic allocation of transmission resources.

[0061] Step S4: Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the compression rate optimization allocation model is constructed according to the information integrity of each modal data as a constraint.

[0062] In response to the need to ensure the information quality of the dominant path under bandwidth constraints, a compression ratio optimization model is constructed to minimize the amount of data transmitted, i.e., save bandwidth, while ensuring information quality. The objective function corresponding to the optimization objective is:

[0063]

[0064] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. Let Mi be the compression ratio vector of the data channel at time T. For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, The compression ratio of the m-th data channel at time t

[0065] The constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. Among them, the information loss limit constraints are used to ensure that the information loss degree of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints are used to dynamically adjust the upper bound of the information loss degree of non-dominant paths according to the dominant path probability; the inter-modal deducibility control information loss constraints are used to constrain the overall transmission traffic using the inter-modal data deducibility; and the compression rate limit constraints are used to ensure that the compression rate of each data channel is within the allowable range.

[0066] Among them, the information loss constraint, the dominant path probability adjustment information loss constraint, the intermodal derivability control information loss constraint, and the compression rate constraint are as follows:

[0067]

[0068] In the formula, The information loss rate of the data channel under compression. For data channel At any moment The upper bound of information loss. Based on the loss threshold, Let be the probability that the m-th channel is the dominant path at time t (determined by the probability of the dominant path). To relax the parameters, allowing non-dominant pathways to tolerate higher loss levels, The data derivability coefficient is generally between 0.6 and 1.0. The larger the value, the stronger the derivability of information between the modes. For data channel n at time n Information loss degree Let be the compression ratio of data channel n at time t. The upper limit of compression ratio (e.g.) The value is 0.8.

[0069] Specifically, to address the differences in information loss and intermodal information redundancy under different compression ratios, the information loss of multimodal data channels under different compression ratios is fitted based on historical data. The relationship between information loss and compression ratio is modeled as follows:

[0070]

[0071] In the formula, For data channel Compression ratio The information loss degree below, among which The maximum allowable compression ratio depends on the specific algorithm. and For data channel The characteristic parameters are obtained by least-squares fitting of historical data.

[0072] The derivability matrix between modal data channels is defined as follows:

[0073]

[0074] In the formula, Representing modes For data channels The derivability coefficient; For mutual information; Information entropy. When When close to 1, the mode It can better derive the data channel. Information.

[0075] Among them, mutual information and information entropy are based on collected historical multimodal feature datasets, and the probability distribution is calculated by discretizing the feature space using histogram statistics or kernel density estimation. and joint distribution The formulas obtained later are as follows:

[0076]

[0077]

[0078] In the formula, Let x be the probability that the feature variable is x. Let x be the probability that the feature variables are x and y. The probability that it is y.

[0079] Step S5: Find the optimal solution for the compression ratio allocation model and obtain the optimal compression ratio for each data channel based on the optimal solution.

[0080] The Lagrange multiplier method can be used to solve the model. By constructing an augmented Lagrange function, the constrained optimization problem is transformed into an unconstrained problem. Combined with the gradient descent algorithm, the compression ratio variable is iteratively updated. Under the premise of satisfying all constraints, the global optimal solution, i.e., the optimal compression ratio, is obtained. .

[0081] Step S6: Based on the optimal compression rate of each data channel, determine the compressed traffic of the data channel, and combine it with the probability distribution of the dominant path to allocate bandwidth priority and obtain the bandwidth allocation result of each data channel.

[0082] In this process, after obtaining the optimal compression ratio for each data channel, the compression process is performed using the compression ratio. The compressed traffic of the data channel is then determined, and bandwidth priority is allocated based on the probability distribution of the dominant path. This allows the dominant path to obtain higher priority bandwidth resources, while the non-dominant path dynamically adjusts its bandwidth usage within a tolerable range of information loss.

[0083] It should be noted that, in this embodiment, the modal feature vector of multimodal data transmitted in the distribution network equipment is used to determine the state characteristics of each modal data under the current equipment state, and the state characteristics of each modal data under the current equipment state are summarized to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient use of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection.

[0084] In some embodiments, comparing the similarity between each modal feature vector and each preset typical state feature to determine the state features of each modal data in the current device state includes: determining the feature proximity and change trend of each modal feature vector with each preset typical state feature based on each modal feature vector and each preset typical state feature; for each preset typical state feature, determining the similarity between each modal feature vector and the preset typical state feature based on the weighted result of the feature proximity and change trend of each modal feature vector with the preset typical state feature; determining the preset typical state feature with the highest similarity to the modal feature vector based on the similarity, and taking the state feature corresponding to the determined preset typical state feature as the state feature of the modal data in the current device state.

[0085] Among them, for equipment Each data channel Calculate the current features With the Typical state Typical characteristics of this mode similarity Feature proximity is evaluated by the distance between the current feature and the features of each typical state. It directly reflects the static consistency between the current state of the device and the preset mode, which helps to accurately identify the category to which the device belongs when it is in a stable state.

[0086] Incorporating feature change trends into similarity considerations allows for the capture of dynamic evolution information about faults. When the features of a certain modality change significantly within a short period of time, it is possible to capture signs of fault evolution more accurately and promptly.

[0087] Therefore, this similarity is composed of two weighted aspects: feature proximity and trend of change, and is expressed as:

[0088]

[0089] In the formula, For equipment Data Channel At any moment With state Similarity; For equipment The Each data channel at time eigenvectors; For the first Data channel under typical conditions The feature vector of expert experience; These are the weighting coefficients; For data channel The similarity scale parameter; It is a norm 2; For time windows.

[0090] The preset typical state feature with the highest similarity to the modal feature vector is determined by similarity, and this typical state is used as the state determination result under the current modality.

[0091] In some embodiments, determining the current dominant data path based on the overall state of the device, and predicting the traffic demand and dominant path probability at a preset time based on the modal feature vectors of each modal data and the current dominant data path, includes: determining an expert experience dominant path that matches the overall state of the device; if the similarity between the state features corresponding to the overall state of the device and the expert experience dominant path is greater than a preset similarity threshold, then the expert experience dominant path is taken as the current dominant data path; if the similarity between the state features corresponding to the overall state of the device and the expert experience dominant path is not greater than a preset similarity threshold, then the data channel with the highest similarity to the overall state of the device is selected from all data channels as the current dominant data path; and using the modal feature vectors of each modal data and the current dominant data path as a preset multi-layer temporal attention network, enabling the preset multi-layer temporal attention network to predict the traffic demand and dominant path probability distribution at a preset time.

[0092] Among them, the expert experience-led path refers to the key data transmission path pre-defined based on historical operation and maintenance data and domain expert knowledge, which corresponds to the optimal communication channel combination under typical device states. This path is determined during the system initialization phase through offline analysis of a large number of fault cases and normal operating conditions, and stored in the knowledge base of the edge nodes. During operation, if the similarity between the current overall device state and a certain typical state matches, the associated expert experience-led path is directly activated. That is, if the similarity between the overall device state and the state characteristics corresponding to the expert experience-led path is greater than a preset similarity threshold, no dynamic adjustment is required, and this path is directly adopted to ensure transmission efficiency and stability.

[0093] If the similarity between the overall equipment status and the status characteristics corresponding to the expert-driven path is not greater than a preset similarity threshold, it indicates that the current equipment status deviates from typical operating conditions, has potential anomalies, or is in a transition process. In this case, it is necessary to dynamically identify the optimal transmission path. By calculating the real-time similarity between each data channel and the overall equipment status, the channel with the highest similarity is selected as the current dominant data path to enhance the response sensitivity to atypical states.

[0094] Then, targeting the temporal evolution patterns of multimodal data features, this application employs a multi-layer temporal attention network to predict traffic demand and the probability of dominant pathways. This multi-layer temporal attention network can effectively capture cross-channel correlations and temporal dependencies between different modal features by introducing a temporal aliasing mask mechanism. Specifically, the input to the multi-layer temporal attention network is a time window. The model takes a matrix of historical feature sequences (i.e., modal feature vectors concatenated from each modality) as input and marks the currently dominant data pathway within the matrix to enhance its ability to perceive the temporal evolution of key pathways and predict future trends. The traffic demand and dominant path probability distribution at each time point are represented as follows:

[0095]

[0096] In the formula, For the first Hidden layer state; For layer normalization function; The input feature matrix; This is a multi-head self-attention mechanism; It is a feedforward network; For splicing operations; For the number of attention heads; For query matrix; The key matrix; It is a value matrix; For the feature dimension of the attention head; For the first The output of each attention head; For the first The query, key, and value projection matrix of each attention head; For attention function; It is a normalized exponential function; To output the projection matrix; This is a temporal aliasing mask matrix.

[0097] The prediction of traffic demand and the probability of dominant pathways at multiple future time points is expressed as:

[0098]

[0099] In the formula, For the predicted time period Traffic demand vectors for each data channel To predict the length of the time window; The probability distribution of the dominant pathway; The number of network layers; For the Sigmoid function; This is the output layer weight matrix; It is the bias vector; This represents the maximum flow vector for each data channel; This is an element-wise product.

[0100] The dominant path probability is a probability distribution obtained by normalizing the linear combination of the output tensors to the (0,1) interval using the Sigmoid function. It is used to characterize the relative probability that each data channel will become the dominant path in the next time step.

[0101] In some embodiments, the bandwidth allocation results for each data channel are obtained by determining the compressed traffic of each data channel based on the optimal compression ratio of each data channel and by allocating bandwidth priority in combination with the dominant path probability distribution. This includes: determining the compressed traffic of each data channel based on the optimal compression ratio of each data channel; and ranking the bandwidth priority of each data channel based on the dominant path priority principle, in combination with the compressed traffic of each data channel and the dominant path probability distribution, so that the data channel with the higher dominant path probability receives higher bandwidth resources.

[0102] The compressed data traffic is calculated based on the optimized compression ratio, and bandwidth is allocated accordingly. The compressed traffic is:

[0103]

[0104] In the formula, For data channel At any moment The compressed traffic demand, A time discount factor slightly less than 1 can gradually reduce the impact of future traffic on total demand as time moves further away from the present, thus capturing future trends while maintaining focus on current priorities. The length of the time window for flow calculation. Let m be the original flow requirement of channel m of device i at time t.

[0105] This multi-slot traffic prediction mechanism enables more proactive bandwidth allocation, reserving or adjusting resources in advance to effectively avoid transmission bottlenecks caused by sudden traffic spikes. This is especially beneficial in scenarios where changes in device status may cause traffic fluctuations, ensuring smoother and more efficient transmission of critical data. The total bandwidth allocation strategy is based on the dominant path priority principle, expressed as:

[0106]

[0107] In the formula, To be assigned to the device Data Channel At any moment bandwidth; Priority weight coefficient for the dominant pathway; The traversal index for summation is used to traverse all systems in the system. Unit of equipment; Total available bandwidth; Let be the compressed data throughput requirement of the nth data channel of device i at time t. Let be the probability that the nth channel of device i is the dominant channel at time t.

[0108] By prioritizing dominant data channels, the dominant data channel receives higher priority in resource competition, reserving more bandwidth resources for data channels with high dominance probability. Simultaneously, it incorporates future multi-slot traffic demand weighted by a time discount factor. This makes bandwidth allocation forward-looking, overcoming the limitations of traditional methods that rely solely on current traffic. It effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of critical information transmission and guaranteeing high-quality transmission. The significance of this method lies in resolving the contradiction between low bandwidth utilization and compromised critical information quality caused by traditional equal resource allocation. Through differentiated compression and dynamic bandwidth allocation, it maintains the integrity of information on dominant channels even with reduced total bandwidth, significantly improving fault detection accuracy and providing an efficient data transmission guarantee mechanism for intelligent monitoring of distribution networks.

[0109] Based on the same inventive concept, embodiments of this application also provide a transmission resource allocation optimization system for implementing the transmission resource allocation optimization method involved above.

[0110] The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more transmission resource allocation optimization system embodiments provided below can be found in the limitations of the transmission resource allocation optimization method described above, and will not be repeated here.

[0111] like Figure 3 As shown in the figure, this application provides a transmission resource allocation optimization system, including:

[0112] The state feature determination module 100 is used to acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity between each modal feature vector and each preset typical state feature, and determine the state features of each modal data under the current equipment state.

[0113] The device status determination module 200 is used to summarize the status characteristics of each modal data under the current device status to obtain the overall status of the current device.

[0114] The data prediction module 300 is used to determine the current dominant data path based on the overall status of the equipment, and to predict the traffic demand and the probability of the dominant path at a preset time in the future based on the modal feature vectors of each modal data and the current dominant data path.

[0115] The allocation model construction module 400 is used to construct a compression ratio optimization allocation model based on the predicted traffic demand and the probability of the dominant path, with the optimization objective of minimizing the total transmission traffic of multimodal data in the distribution network equipment, and with the information integrity of each modal data as a constraint.

[0116] The model solving module 500 is used to find the optimal solution for the compression ratio optimization allocation model and obtain the optimal compression ratio of each data channel based on the optimal solution.

[0117] The bandwidth allocation module 600 is used to determine the compressed traffic of each data channel based on the optimal compression rate of each data channel, and to allocate bandwidth priority in combination with the probability distribution of the dominant path to obtain the bandwidth allocation result of each data channel.

[0118] In some embodiments, the state feature determination module 100 is configured to:

[0119] Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector to each preset typical state feature;

[0120] For each preset typical state feature, the similarity between each modal feature vector and the preset typical state feature is determined by weighting the feature proximity and change trend between each modal feature vector and the preset typical state feature.

[0121] Based on the similarity, the preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

[0122] In some embodiments, the data prediction module 300 is used for:

[0123] Based on the overall state of the equipment, an expert experience-led path that matches the overall state of the equipment is determined. If the similarity between the state features corresponding to the overall state of the equipment and the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path.

[0124] If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-driven path is not greater than the preset similarity threshold, then the data channel with the highest similarity to the overall state of the equipment is selected from all data channels as the current dominant data channel.

[0125] Based on the modal feature vectors of each modality data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which then predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

[0126] In some embodiments, the objective function corresponding to the optimization objective is:

[0127]

[0128] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. For data channel M i The compression ratio vector at time T, For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

[0129] In some embodiments, the constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. The information loss limit constraints ensure that the information loss of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints dynamically adjust the upper bound of the information loss of non-dominant paths based on the dominant path probability; the inter-modal deducibility control information loss constraints utilize inter-modal data deducibility to constrain the overall transmission traffic; and the compression rate limit constraints ensure that the compression rate of each data channel is within the allowable range.

[0130] In some embodiments, the bandwidth allocation module 600 is used for:

[0131] Based on the optimal compression ratio of each data channel, determine the compressed data throughput of each data channel;

[0132] Based on the principle of prioritizing dominant paths, and combining the compressed traffic and the probability distribution of dominant paths for each data channel, the bandwidth priority of each data channel is ranked so that the data channel with the higher probability of dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result for each data channel.

[0133] This application embodiment determines the state characteristics of each modal data under the current equipment state by using the modal feature vector of multimodal data transmitted in the distribution network equipment, and summarizes the state characteristics of each modal data under the current equipment state to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient use of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection.

[0134] like Figure 4 As shown, this application embodiment provides an electronic device. The electronic device 10 includes a memory 20 and a processor 30. The memory 20 stores a computer program. When the computer program is executed by the processor 30, the processor 30 performs the following steps:

[0135] Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state;

[0136] The overall state of the current device is obtained by summarizing the state characteristics of each modal data under the current device state;

[0137] The current dominant data path is determined based on the overall status of the equipment. The traffic demand and the probability of the dominant path at a future preset time are predicted based on the modal feature vectors of each modal data and the current dominant data path.

[0138] Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and a compression ratio optimization allocation model is constructed according to the information integrity of each modal data as a constraint.

[0139] The optimal compression ratio allocation model is optimized and solved, and the optimal compression ratio of each data channel is obtained based on the optimal solution.

[0140] Based on the optimal compression rate of each data channel, the compressed traffic of the data channel is determined, and bandwidth priority is allocated in combination with the probability distribution of the dominant path to obtain the bandwidth allocation result of each data channel.

[0141] In some embodiments, when the computer program is executed by the processor 30, the processor 30 further performs the following steps:

[0142] Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector to each preset typical state feature;

[0143] For each preset typical state feature, the similarity between each modal feature vector and the preset typical state feature is determined by weighting the feature proximity and change trend between each modal feature vector and the preset typical state feature.

[0144] Based on the similarity, the preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

[0145] In some embodiments, when the computer program is executed by the processor 30, the processor 30 further performs the following steps:

[0146] Based on the overall state of the equipment, an expert experience-led path that matches the overall state of the equipment is determined. If the similarity between the state features corresponding to the overall state of the equipment and the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path.

[0147] If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-driven path is not greater than the preset similarity threshold, then the data channel with the highest similarity to the overall state of the equipment is selected from all data channels as the current dominant data channel.

[0148] Based on the modal feature vectors of each modality data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which then predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

[0149] In some embodiments, the objective function corresponding to the optimization objective is:

[0150]

[0151] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. For data channel Mi The compression ratio vector at time T, For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

[0152] In some embodiments, the constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. The information loss limit constraints ensure that the information loss of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints dynamically adjust the upper bound of the information loss of non-dominant paths based on the dominant path probability; the inter-modal deducibility control information loss constraints utilize inter-modal data deducibility to constrain the overall transmission traffic; and the compression rate limit constraints ensure that the compression rate of each data channel is within the allowable range.

[0153] In some embodiments, when the computer program is executed by the processor 30, the processor 30 further performs the following steps:

[0154] Based on the optimal compression ratio of each data channel, determine the compressed data throughput of each data channel;

[0155] Based on the principle of prioritizing dominant paths, and combining the compressed traffic and the probability distribution of dominant paths for each data channel, the bandwidth priority of each data channel is ranked so that the data channel with the higher probability of dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result for each data channel.

[0156] This application embodiment determines the state characteristics of each modal data under the current equipment state by using the modal feature vector of multimodal data transmitted in the distribution network equipment, and summarizes the state characteristics of each modal data under the current equipment state to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient use of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection.

[0157] This application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed, it performs the following steps:

[0158] Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state;

[0159] The overall state of the current device is obtained by summarizing the state characteristics of each modal data under the current device state;

[0160] The current dominant data path is determined based on the overall status of the equipment. The traffic demand and the probability of the dominant path at a future preset time are predicted based on the modal feature vectors of each modal data and the current dominant data path.

[0161] Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and a compression ratio optimization allocation model is constructed according to the information integrity of each modal data as a constraint.

[0162] The optimal compression ratio allocation model is optimized and solved, and the optimal compression ratio of each data channel is obtained based on the optimal solution.

[0163] Based on the optimal compression rate of each data channel, the compressed traffic of the data channel is determined, and bandwidth priority is allocated in combination with the probability distribution of the dominant path to obtain the bandwidth allocation result of each data channel.

[0164] In some embodiments, when a computer program is executed, it also performs the following steps:

[0165] Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector to each preset typical state feature;

[0166] For each preset typical state feature, the similarity between each modal feature vector and the preset typical state feature is determined by weighting the feature proximity and change trend between each modal feature vector and the preset typical state feature.

[0167] Based on the similarity, the preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

[0168] In some embodiments, when a computer program is executed, it also performs the following steps:

[0169] Based on the overall state of the equipment, an expert experience-led path that matches the overall state of the equipment is determined. If the similarity between the state features corresponding to the overall state of the equipment and the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path.

[0170] If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-driven path is not greater than the preset similarity threshold, then the data channel with the highest similarity to the overall state of the equipment is selected from all data channels as the current dominant data channel.

[0171] Based on the modal feature vectors of each modality data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which then predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

[0172] In some embodiments, the objective function corresponding to the optimization objective is:

[0173]

[0174] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. For data channel M i The compression ratio vector at time T, For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

[0175] In some embodiments, the constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. The information loss limit constraints ensure that the information loss of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints dynamically adjust the upper bound of the information loss of non-dominant paths based on the dominant path probability; the inter-modal deducibility control information loss constraints utilize inter-modal data deducibility to constrain the overall transmission traffic; and the compression rate limit constraints ensure that the compression rate of each data channel is within the allowable range.

[0176] In some embodiments, when a computer program is executed, it also performs the following steps:

[0177] Based on the optimal compression ratio of each data channel, determine the compressed data throughput of each data channel;

[0178] Based on the principle of prioritizing dominant paths, and combining the compressed traffic and the probability distribution of dominant paths for each data channel, the bandwidth priority of each data channel is ranked so that the data channel with the higher probability of dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result for each data channel.

[0179] This application embodiment determines the state characteristics of each modal data under the current equipment state by using the modal feature vector of multimodal data transmitted in the distribution network equipment, and summarizes the state characteristics of each modal data under the current equipment state to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient use of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection.

[0180] This application provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer performs the following steps:

[0181] Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state;

[0182] The overall state of the current device is obtained by summarizing the state characteristics of each modal data under the current device state;

[0183] The current dominant data path is determined based on the overall status of the equipment. The traffic demand and the probability of the dominant path at a future preset time are predicted based on the modal feature vectors of each modal data and the current dominant data path.

[0184] Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and a compression ratio optimization allocation model is constructed according to the information integrity of each modal data as a constraint.

[0185] The optimal compression ratio allocation model is optimized and solved, and the optimal compression ratio of each data channel is obtained based on the optimal solution.

[0186] Based on the optimal compression rate of each data channel, the compressed traffic of the data channel is determined, and bandwidth priority is allocated in combination with the probability distribution of the dominant path to obtain the bandwidth allocation result of each data channel.

[0187] In some embodiments, when program instructions are executed by a computer, the computer also performs the following steps:

[0188] Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector to each preset typical state feature;

[0189] For each preset typical state feature, the similarity between each modal feature vector and the preset typical state feature is determined by weighting the feature proximity and change trend between each modal feature vector and the preset typical state feature.

[0190] Based on the similarity, the preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

[0191] In some embodiments, when program instructions are executed by a computer, the computer also performs the following steps:

[0192] Based on the overall state of the equipment, an expert experience-led path that matches the overall state of the equipment is determined. If the similarity between the state features corresponding to the overall state of the equipment and the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path.

[0193] If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-driven path is not greater than the preset similarity threshold, then the data channel with the highest similarity to the overall state of the equipment is selected from all data channels as the current dominant data channel.

[0194] Based on the modal feature vectors of each modality data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which then predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

[0195] In some embodiments, the objective function corresponding to the optimization objective is:

[0196]

[0197] In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. For data channel M i The compression ratio vector at time T, For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

[0198] In some embodiments, the constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. The information loss limit constraints ensure that the information loss of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints dynamically adjust the upper bound of the information loss of non-dominant paths based on the dominant path probability; the inter-modal deducibility control information loss constraints utilize inter-modal data deducibility to constrain the overall transmission traffic; and the compression rate limit constraints ensure that the compression rate of each data channel is within the allowable range.

[0199] In some embodiments, when program instructions are executed by a computer, the computer also performs the following steps:

[0200] Based on the optimal compression ratio of each data channel, determine the compressed data throughput of each data channel;

[0201] Based on the principle of prioritizing dominant paths, and combining the compressed traffic and the probability distribution of dominant paths for each data channel, the bandwidth priority of each data channel is ranked so that the data channel with the higher probability of dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result for each data channel.

[0202] This application embodiment determines the state characteristics of each modal data under the current equipment state by using the modal feature vector of multimodal data transmitted in the distribution network equipment, and summarizes the state characteristics of each modal data under the current equipment state to obtain the overall state of the current equipment. Based on the overall state of the equipment, the current dominant data path is determined, and the traffic demand and dominant path probability at a preset time in the future are predicted. Based on the predicted traffic demand and dominant path probability, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and the optimal compression rate of each data channel is obtained based on the information integrity of each modal data as a constraint. This enables efficient use of bandwidth resources while ensuring the transmission quality of key information. Based on the optimal compression rate of each data channel, the traffic of the data channel after compression is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution, thereby reserving more bandwidth resources for data channels with high dominant probability. This effectively predicts and responds to future traffic fluctuations, ensuring the continuity and stability of key information transmission, and ensuring high-quality transmission of key information. Through differentiated compression and dynamic bandwidth allocation, the information integrity of the dominant path can still be maintained even when the total bandwidth is reduced, significantly improving the accuracy of fault detection.

[0203] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, electronic devices, computer storage media, and computer program products described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0204] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification, claims and accompanying drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products or devices.

[0205] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0206] In the several embodiments provided by this invention, it should be understood that the disclosed systems, electronic devices, computer storage media, computer program products, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0207] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0208] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0209] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the methods described in the various embodiments of the present invention through a computer device (which may be a personal computer, a server, or a network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0210] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for transmission resource allocation optimization, the method comprising: include: Acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state; The overall state of the current device is obtained by summarizing the state characteristics of each modal data under the current device state; The current dominant data path is determined based on the overall status of the device, and the traffic demand and the probability of the dominant path at a future preset time are predicted based on the modal feature vectors of each modal data and the current dominant data path. Based on the predicted traffic demand and the probability of the dominant path, the optimization objective is to minimize the total transmission traffic of multimodal data in the distribution network equipment, and a compression ratio optimization allocation model is constructed according to the information integrity of each modal data as a constraint. The compression ratio optimization allocation model is optimized and solved, and the optimal compression ratio of each data channel is obtained based on the optimal solution; Based on the optimal compression rate of each data channel, the compressed traffic of the data channel is determined, and bandwidth priority allocation is performed in combination with the dominant path probability distribution to obtain the bandwidth allocation result of each data channel.

2. The method of claim 1, wherein, The step of comparing the similarity between each modal feature vector and each preset typical state feature to determine the state features of each modal data under the current device state includes: Based on each modal feature vector and each preset typical state feature, determine the feature proximity and change trend of each modal feature vector with each preset typical state feature; For each of the preset typical state features, the similarity between each modal feature vector and the preset typical state feature is determined based on the weighted result of the feature proximity and change trend between each modal feature vector and the preset typical state feature. Based on the similarity, a preset typical state feature with the highest similarity to the modal feature vector is determined, and the state feature corresponding to the determined preset typical state feature is used as the state feature of the modal data in the current device state.

3. The method of claim 1, wherein, The step of determining the current dominant data path based on the overall state of the device, and predicting the traffic demand and dominant path probability at a preset time based on the modal feature vectors of each modal data and the current dominant data path, includes: Based on the overall state of the equipment, an expert experience-led path matching the overall state of the equipment is determined. If the similarity between the overall state of the equipment and the state features corresponding to the expert experience-led path is greater than a preset similarity threshold, then the expert experience-led path is taken as the current dominant data path. If the similarity between the overall state of the device and the state feature corresponding to the expert experience-dominated path is not greater than a preset similarity threshold, then the data channel with the highest similarity to the overall state of the device is selected from all data channels as the current dominant data channel. Based on the modal feature vectors of each modal data and the current dominant data path, a preset multi-layer temporal attention network is constructed, which predicts the traffic demand and the probability distribution of the dominant path at a preset time in the future.

4. The transmission resource allocation optimization method according to claim 1, characterized in that, The objective function corresponding to the optimization objective is: In the formula For compression ratio vector, , Let be the compression ratio vector of data channel 1 at the initial moment. For data channel M i The compression ratio vector at time T, For matrix transpose, For the number of data channels, For the total time, For a moment Time discount factor, For data channel At any moment Traffic demand, Let be the compression ratio of the m-th data channel at time t.

5. The transmission resource allocation optimization method according to claim 1, characterized in that, The constraints include information loss limit constraints, dominant path probability adjustment information loss constraints, inter-modal deducibility control information loss constraints, and compression rate limit constraints. Specifically, the information loss limit constraints ensure that the information loss of each data channel does not exceed the information loss limit constraint; the dominant path probability adjustment information loss constraints dynamically adjust the upper bound of the information loss of non-dominant paths based on the dominant path probability; the inter-modal deducibility control information loss constraints utilize inter-modal data deducibility to constrain the overall transmission traffic; and the compression rate limit constraints ensure that the compression rate of each data channel is within the allowable range.

6. The transmission resource allocation optimization method according to claim 1, characterized in that, The process of determining the compressed traffic of each data channel based on its optimal compression ratio, and allocating bandwidth priority according to the dominant path probability distribution to obtain the bandwidth allocation result for each data channel includes: Based on the optimal compression ratio of each data channel, determine the compressed data flow rate of each data channel; Based on the principle of prioritizing dominant paths, and combining the compressed traffic of each data channel with the probability distribution of the dominant path, the bandwidth priority of each data channel is sorted so that the data channel with the higher probability of the dominant path receives higher bandwidth resources, thus obtaining the bandwidth allocation result of each data channel.

7. A transmission resource allocation optimization system, characterized in that, include: The state feature determination module is used to acquire multimodal data transmitted in the power distribution network equipment, extract the modal feature vectors of each modal data, compare the similarity of each modal feature vector with each preset typical state feature, and determine the state features of each modal data under the current equipment state. The device status determination module is used to summarize the status characteristics of each modal data under the current device status to obtain the overall status of the current device. The data prediction module is used to determine the current dominant data path based on the overall status of the device, and to predict the traffic demand and the probability of the dominant path at a future preset time based on the modal feature vectors of each modal data and the current dominant data path. The allocation model construction module is used to construct a compression ratio optimization allocation model based on the predicted traffic demand and the probability of the dominant path, with the optimization objective of minimizing the total transmission traffic of multimodal data in the distribution network equipment, and with the information integrity of each modal data as a constraint. The model solving module is used to find the optimal solution for the compression ratio optimization allocation model and obtain the optimal compression ratio for each data channel based on the optimal solution. The bandwidth allocation module is used to determine the compressed traffic of each data channel based on the optimal compression rate of each data channel, and to allocate bandwidth priority in combination with the dominant path probability distribution to obtain the bandwidth allocation result of each data channel.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the transmission resource allocation optimization method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the steps of the transmission resource allocation optimization method as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, wherein when the program instructions are executed by a computer, the computer performs the steps of the transmission resource allocation optimization method as described in any one of claims 1-6.

Images