Emergency resource data sharing method based on deep learning and storage medium

Abstract

The application provides an emergency resource data sharing method and storage medium based on deep learning, which comprises the following steps: acquiring original resource data sets distributed in different emergency response nodes; performing feature unification processing on the original resource data sets to generate an emergency resource feature set with consistent semantics; inputting the emergency resource feature set into a pre-trained sharing relationship modeling model for correlation analysis processing to generate a sharing correlation feature set reflecting the resource supply-demand matching relationship; generating an emergency resource data sharing rule set based on the sharing correlation feature set; and performing a sharing operation on multi-source emergency resource data according to the emergency resource data sharing rule set. The emergency resource data sharing process can make full use of the complementary information of multi-source data and flexibly adjust the sharing strategy according to the actual scene demand, thereby improving the pertinence of emergency resource allocation and the effectiveness of information transmission and ensuring the accuracy of resource supply-demand matching.

Description

Technical Field

[0001] The present invention relates to the field of data sharing, and more specifically, to an emergency resource data sharing method and storage medium based on deep learning. Background Art

[0002] Emergency resource data sharing refers to a key technology that enables the exchange of resource information between multiple nodes, such as reserves, transportation, and demand, through information exchange during emergency response, to support rapid decision-making and resource allocation. Current emergency resource data sharing methods often rely on direct data transmission or preset fixed rules. For example, material storage lists from reserve nodes are pushed directly to demand nodes, or information is matched according to pre-set rules. However, this approach faces the dual challenges of heterogeneous multi-source data and dynamic demand changes. The data formats and semantic standards of reserve, transportation, and demand nodes vary. Directly transmitted isolated data cannot reflect the implicit associations between material types and demand scenarios, transportation routes, and storage locations. Preset rules cannot adjust priorities based on the supply and demand patterns in the actual data, resulting in insufficiently targeted resource information transmission and difficulty meeting the needs of flexible allocation in emergency scenarios. Summary of the Invention

[0003] In view of this, an embodiment of the present invention provides an emergency resource data sharing method and storage medium based on deep learning.

[0004] According to one aspect of an embodiment of the present invention, a method for sharing emergency resource data based on deep learning is provided, comprising:

[0005] Obtaining a collection of original resource data distributed across different emergency response nodes;

[0006] Performing feature unification processing on the original resource data set to generate an emergency resource feature set with semantic consistency;

[0007] Inputting the emergency resource feature set into a pre-trained shared relationship modeling model for association analysis and processing to generate a shared association feature set reflecting the resource supply and demand matching relationship;

[0008] generating an emergency resource data sharing rule set based on the shared association feature set;

[0009] The multi-source emergency resource data sharing operation is performed according to the emergency resource data sharing rule set.

[0010] According to another aspect of an embodiment of the present invention, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the method provided in the first aspect above are implemented.

[0011] The emergency resource data sharing method based on deep learning provided by the application, by acquiring the original resource data set distributed in different emergency response nodes, converting the heterogeneous storage, circulation and request record into an emergency resource feature set with semantic consistency, the originally isolated multi-source data has a structured basis for associated analysis; further, the pre-trained shared relationship modeling model is used for associated analysis of the feature set, and the implicit matching relationship between the material type and the demand scene, the transportation path and the storage location, the circulation time and the demand time limit is mined, which changes the limitation that the traditional emergency resource sharing only relies on the preset fixed rules; the dynamic sharing rule set generated based on the above associated relationship can automatically adjust the sharing priority according to the supply and demand mode in the actual data, avoiding the defect that the static rule cannot adapt to the dynamic change of the emergency scene; finally, the multi-source data sharing operation is performed according to the rule, and the effective intercommunication of resource information between the storage, transportation and demand nodes is realized. This method converts multi-source data into associated structured features, uses a deep learning model to mine the implicit supply and demand relationship, and generates dynamically adaptive sharing rules, so that the emergency resource data sharing process can not only make full use of the complementary information of multi-source data, but also flexibly adjust the sharing strategy according to the actual scene demand, thereby improving the pertinence of emergency resource allocation and the effectiveness of information transmission, and ensuring the accuracy of resource supply and demand matching.

[0012] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, but not limiting the technical solutions of the application. BRIEF DESCRIPTION OF DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0014] Figure 1 is an architecture schematic diagram of an application scenario provided by the application;

[0015] Figure 2 is a flowchart of an emergency resource data sharing method based on deep learning provided by the application;

[0016] Figure 3 is a structural schematic diagram of a server provided by the embodiment of the application. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] In order to facilitate a clearer understanding of the present invention, we first introduce the application scenarios of the emergency resource data sharing method based on deep learning of the present invention. Figure 1 As shown, the application scenario includes a server 10 and a terminal cluster. The terminal cluster can include one or more terminals. There is no limit on the number of terminals. Each terminal cluster can correspond to an emergency response node. Figure 1 As shown, the terminal cluster may specifically include terminal 1, terminal 2, ..., terminal n; it can be understood that terminal 1, terminal 2, terminal 3, ..., terminal n can all be connected to the server 10 through a network connection, so that each terminal can exchange data with the server 10 through the network connection.

[0019] It is understandable that the server 10 may refer to a device that executes the emergency resource data sharing method based on deep learning provided by the present invention, wherein the server may be an independent physical server, or a server cluster or distributed system composed of at least two physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal may specifically refer to a reserve node terminal, a transport node terminal, or a demand node terminal, but is not limited thereto. Each terminal and server may be directly or indirectly connected via wired or wireless communication. At the same time, the number of terminals and servers may be one or at least two, and the present invention does not impose any restrictions thereto.

[0020] Further, see Figure 2 , is a flowchart of an emergency resource data sharing method based on deep learning provided by an embodiment of the present invention. Figure 2 As shown, this method can be Figure 1 The server in the embodiment is used for execution, wherein the emergency resource data sharing method based on deep learning may include the following steps S100 to S500:

[0021] Step S100: Acquire a set of original resource data distributed on different emergency response nodes.

[0022] The raw resource data set refers to a collection of unprocessed resource-related data from different emergency response nodes. This data contains relevant information about emergency resources in the storage, transportation, and demand stages. The material storage record of a reserve node refers to a detailed record of the material storage situation in the reserve node, where the material identifier is used to uniquely identify the type of material, the storage capacity identifier indicates the storage quantity of the material, and the storage time identifier records the specific time when the material enters the warehouse. The material flow record of a transportation node refers to the relevant record of the material transportation process in the transportation node. The material identifier is also used to identify the material. The starting position identifier indicates the starting point of the material transportation, the target position identifier indicates the end point of the material transportation, and the transportation time identifier records the time required for the material transportation. The material request record of a demand node refers to the record of the material demand situation in the demand node. The material identifier identifies the required material, the required quantity identifier indicates the required quantity of the material, the required location identifier records the location of the required material, and the required time identifier records the time when the material is required.

[0023] As an implementation method, the original resource data set includes material storage records of reserve nodes, material flow records of transportation nodes, and material request records of demand nodes. Step S100 obtains the original resource data set distributed across different emergency response nodes. Specifically, the following steps S110 to S150 can be implemented:

[0024] Step S110: Send a storage data pull instruction to the reserve node, and receive the material storage record containing the material identification, storage capacity identification and storage time identification returned by the reserve node.

[0025] The storage data pull instruction is used to obtain material storage records from a storage node. This instruction requires specifying the target storage node and the type of material storage record to be pulled. In this embodiment, the process of obtaining material storage records is implemented through a series of specific operations. As an implementation method, step S110 sends a storage data pull instruction to the storage node, and receives the material storage record returned by the storage node, which includes the material identifier, storage capacity identifier, and storage time identifier. Specifically, the following steps can be implemented: S111-S115:

[0026] Step S111: Construct a storage data pulling instruction including a storage node identifier and a data type identifier. The storage node identifier is used to locate the target storage node, and the data type identifier is used to specify the type of material storage record to be pulled.

[0027] The reserve node identifier is information capable of uniquely identifying the reserve node, which can be, for example, a geographic location code of the reserve node, a network address, etc., through which the reserve node to which data is to be acquired can be accurately found. The data type identifier is used to explicitly specify the specific type of the material storage record to be pulled, such as classification according to the use, specifications, etc. of the material. When constructing the storage data pulling instruction, the reserve node identifier and the data type identifier are reasonably combined to form a complete instruction information. For example, the reserve node identifier and the data type identifier can be spliced in a set format, such as "reserve node number_data type number", to indicate which type of material storage record is to be acquired from which reserve node.

[0028] Step S112: sending the storage data pulling instruction to the communication interface of the target reserve node through the emergency data transmission protocol.

[0029] The emergency data transmission protocol is a protocol designed for data transmission in emergency scenarios, and the communication interface of the target reserve node is an interface of the reserve node for receiving external instructions and sending data, which complies with communication standards and specifications. When sending the storage data pulling instruction, the constructed instruction is encapsulated according to the format of the emergency data transmission protocol, for example, adding protocol header, protocol tail, and check code, etc. information to ensure the integrity and accuracy of the instruction in the transmission process. Then, the encapsulated instruction is sent to the communication interface of the target reserve node through the network, and the communication interface receives the instruction and parses and processes the instruction.

[0030] Step S113: receiving the response data returned by the target reserve node through the emergency data transmission protocol, the response data containing a material identification field, a storage capacity field, and a warehousing time field.

[0031] After receiving the storage data pulling instruction, the target reserve node extracts the corresponding material storage record from its storage system according to the requirements of the instruction, and encapsulates these records according to the format of the emergency data transmission protocol to form response data returned to the sender. The material identification field in the response data is used to record the unique identification information of the material, the storage capacity field records the storage quantity of the material, and the warehousing time field records the specific time of warehousing of the material. When receiving the response data, the same emergency data transmission protocol as sending the instruction is used for parsing to ensure that the material identification, storage capacity, and warehousing time, etc. information can be accurately acquired.

[0032] Step S114: field parsing and processing of the response data, extracting the material classification code in the material identification field, the specific storage quantity in the storage capacity field, and the warehousing date in the warehousing time field, to generate a material storage record containing the material identification, the storage capacity identification, and the warehousing time identification.

[0033] The field parsing process is a process of splitting and extracting each field in the response data. For the material identification field, the string splitting method can be used to extract the material classification code in the field according to the predefined separator. For the storage capacity field, the data type conversion method can be used to convert the storage quantity information in the field into a suitable data type, such as an integer or a floating point number. For the warehousing time field, the date parsing function can be used to extract the date information in the field and convert it into a unified date format. Through these processes, the response data is converted into a material storage record containing material identification, storage capacity identification and warehousing time identification.

[0034] Step S115: integrity check processing is performed on the material storage record, and the field integrity check rule is used to check whether the material identification, storage capacity identification and warehousing time identification all contain valid data values, and the invalid storage record with missing fields is removed to obtain the complete material storage record.

[0035] The field integrity check rule is a set of rules defined in advance to determine whether each field in the material storage record contains valid data. For example, it can be checked whether the material identification is an empty string, whether the storage capacity identification is a legal numerical value, and whether the warehousing time identification is a valid date format. For the material storage record with missing fields, it is determined as an invalid storage record and is removed. Through the integrity check processing, it can be ensured that the obtained material storage record is complete and valid, and accurate data basis is provided for subsequent processing.

[0036] Step S120: sending a flow data query instruction to the transport node, receiving the material flow record returned by the transport node containing the material identification, the starting position identification, the target position identification and the transport time identification.

[0037] The flow data query instruction is an instruction for obtaining the material flow record from the transport node, which needs to specify the related information of the material flow record to be queried. In this embodiment, when the flow data query instruction is sent to the transport node, it is also encapsulated according to the protocol format to ensure that the instruction can be accurately received and processed by the transport node. After receiving the instruction, the transport node extracts the corresponding material flow record from its storage system and returns these records in the specified format. The material identification in the material flow record is used to identify the transported material, the starting position identification represents the starting point of the material transportation, the target position identification represents the end point of the material transportation, and the transport time identification records the time required for the material transportation. After receiving the returned material flow record, it is parsed and verified to ensure the accuracy and integrity of the record.

[0038] Step S130: Send a data synchronization request instruction to the demand node, and receive a material request record including a material identifier, a required quantity identifier, a required location identifier, and a required time identifier returned by the demand node.

[0039] The request data synchronization instruction is used to obtain material request records from the demand node. The purpose of this instruction is to ensure that the latest material demand information of the demand node is obtained. When sending the request data synchronization instruction, follow the corresponding communication protocol and format to ensure that the instruction can be correctly identified and processed by the demand node. After receiving the instruction, the demand node organizes and encapsulates the material request records it stores and then returns them to the sender. The material identifier in the material request record is used to identify the required materials, the required quantity identifier indicates the required quantity of materials, the required location identifier records the location of the required materials, and the required time identifier records the time when the materials are required. After receiving the returned material request record, it is subjected to data cleaning and verification to ensure the quality and availability of the record.

[0040] Step S140: perform identification alignment processing on the material storage records, material flow records and material request records, unify the inconsistent material identifications in each record into a preset standard material classification code, unify the unmatched location identifications in each record into a preset regional coordinate code, and obtain a set of original resource data with consistent identification.

[0041] Identifier alignment is designed to eliminate inconsistencies in identifiers across different records and improve data consistency and comparability. The pre-set standard material classification code is a unified material classification standard that unifies material identifiers of varying formats and representations. The pre-set regional coordinate code is a unified code used to represent geographic locations, unifying different location identifiers into a standard coordinate format. When performing identifier alignment, the mapping relationship between material identifiers and standard material classification codes, as well as between location identifiers and regional coordinate codes, must be established. This conversion can be achieved by searching pre-established mapping tables. For example, for material identifiers, each material identifier can be matched with a standard material classification code, the corresponding code found, and replaced. For location identifiers, they can be converted into geographic coordinates, and then, based on pre-set regional division rules, these coordinates are converted into regional coordinate codes. Through these processes, a consistent set of raw resource data with consistent identifiers is obtained, providing a unified data foundation for subsequent data analysis and processing.

[0042] Step S150: Perform time dimension calibration on the original resource data set. Based on the demand time identifier, convert the storage time identifier of the storage record into a time difference identifier with the demand time, and convert the transportation time identifier of the flow record into a schedulable time window identifier before the demand time, to obtain the original resource data set aligned with the time dimension.

[0043] The time dimension calibration process is to make the original resource data set consistent and comparable in the time dimension. The demand time identifier is the time point of material demand, and this is used as the benchmark for time dimension calibration. For the storage record's warehousing time identifier, by calculating the difference between the warehousing time and the demand time, it is converted into a time difference identifier with the demand time. This time difference identifier can reflect the storage length of the material before the demand time. For the transportation time identifier of the circulation record, based on the demand time and transportation time, the time window in which the material can be scheduled before the demand time is calculated, and the transportation time identifier is converted into a schedulable time window identifier before the demand time. In this way, different records can be aligned in the time dimension to obtain a time-dimension aligned original resource data set, which helps to more accurately analyze and process emergency resource data.

[0044] Step S200: performing feature unification processing on the original resource data set to generate an emergency resource feature set with semantic consistency.

[0045] Feature unification processing is to process different types of data in the original resource data set, extract representative features, and unify and standardize these features to generate an emergency resource feature set with semantic consistency. The emergency resource feature set is a set of features that can reflect the relevant characteristics of emergency resources, including material type association features corresponding to storage records, transportation path association features corresponding to flow records, and demand scenario association features corresponding to request records. In this embodiment, the purpose of feature unification processing is to eliminate differences in the original data caused by factors such as different data sources and formats, so that different types of data can be analyzed and processed under a unified semantic framework.

[0046] As an implementation method, the emergency resource feature set includes material type association features corresponding to storage records, transportation route association features corresponding to flow records, and demand scenario association features corresponding to request records. In step S200, the original resource data set is subjected to feature unification processing to generate an emergency resource feature set with semantic consistency. Specifically, the following steps S210 to S250 can be implemented:

[0047] Step S210: semantically analyze the material storage records in the original resource data set, extract the association relationship between the storage capacity identifier and the material identifier, and generate material type association features reflecting the storage scale of different material types.

[0048] Semantic parsing involves in-depth analysis of the data in material storage records to understand their inherent semantic information. In this embodiment, the storage capacity identifiers and material identifiers in the material storage records are analyzed to extract the association between them. For example, the sum of the storage capacity identifiers corresponding to each material identifier can be calculated to reflect the storage scale of different material types. The generated material type association feature can be represented by a vector or matrix, where each element represents the storage scale information of a material type. In this way, the data in the material storage records can be converted into more representative and analyzable material type association features.

[0049] Step S220: performing path analysis on the material flow records in the original resource data set, extracting the association relationship between the starting location identifier, the target location identifier and the transportation time identifier, and generating a transportation path association feature reflecting the timeliness of the transportation path.

[0050] Path analysis processing is to analyze the information related to the transportation path in the material flow record to extract valuable features. In this embodiment, the starting position identifier, the target position identifier and the transportation time identifier are analyzed to extract the correlation between them. For example, the transportation time between different starting positions and target positions can be calculated, and the relationship between the transportation time and the starting position and target position can be analyzed. The graph theory method can be used to regard the starting position and target position as nodes in the graph, the transportation path as an edge, and the transportation time as the weight of the edge to construct a transportation path graph. Then, by analyzing the transportation path graph, features reflecting the timeliness of the transportation path are extracted, such as the shortest transportation time, the average transportation time, etc. The generated transportation path correlation features can be represented by a vector or a matrix, in which each element represents the timeliness information of a transportation path.

[0051] Step S230: Perform scenario recognition processing on the material request records in the original resource data set, extract the association relationship between the demand quantity identifier, the demand location identifier and the demand time identifier, and generate demand scenario association features that reflect the urgency and spatial distribution of the demand.

[0052] Scenario recognition processing is to analyze the data in the material request record to identify different demand scenarios. In this embodiment, by analyzing the demand quantity identifier, demand location identifier and demand time identifier, the correlation between them is extracted. For example, by calculating the difference between the demand time and the current time, the urgency of the demand is determined; by performing spatial analysis on the demand location identifier, the spatial distribution of the demand is identified. Specifically, the time difference calculation is performed on the demand time identifier in the material request record, and the time interval value between the demand time identifier and the current time is calculated based on the current time. The demand scenarios are divided into urgency according to the size of the time interval value to generate a demand urgency identifier. The demand location identifier in the material request record is subjected to regional clustering processing, and the demand location identifier is divided into multiple demand-intensive areas through a spatial clustering algorithm to generate a demand spatial distribution identifier. The demand urgency identifier is associated and bound with the demand spatial distribution identifier to generate a demand scenario feature vector reflecting the urgency and spatial distribution of the demand. The demand scenario feature vector is normalized, and the time interval value of the demand urgency identifier and the regional density value of the demand spatial distribution identifier are mapped to standardized values ​​within a preset range through the feature normalization rule, so as to obtain the demand scenario association feature reflecting the demand urgency and spatial distribution.

[0053] As an implementation method, step S230 performs scenario recognition processing on the material request records in the original resource data set, extracts the association relationship between the demand quantity identifier, the demand location identifier, and the demand time identifier, and generates demand scenario association features reflecting the urgency and spatial distribution of the demand. Specifically, the following steps S231 to S235 can be implemented:

[0054] Step S231: Calculate the time difference of the demand time identifier in the material request record, and calculate the time interval between the demand time identifier and the current time based on the current time.

[0055] In this embodiment, the demand time stamp records the specific time of the material demand. By obtaining the current time and comparing it with the demand time stamp, the time interval between them is calculated. This time interval reflects the urgency of the demand; the shorter the time interval, the more urgent the demand. To calculate the time difference, you can use date and time processing functions to convert the demand time stamp and the current time into a unified time format and then calculate the difference. For example, you can convert the time to a timestamp and then calculate the difference between the two timestamps.

[0056] Step S232: Classify the demand scenario into different urgency levels according to the time interval value and generate a demand urgency identifier. After obtaining the time interval value between the demand time and the current time, the demand scenario is classified into different urgency levels according to a preset time interval threshold.

[0057] For example, demand scenarios with time intervals less than a certain threshold can be classified as high urgency, those within a set range as medium urgency, and those greater than a certain threshold as low urgency. Each urgency level is then assigned a unique identifier to generate a demand urgency identifier. This identifier can be used for subsequent data analysis and decision-making.

[0058] Step S233: performing regional clustering processing on the demand location identifiers in the material request record, dividing the demand location identifiers into multiple demand-intensive areas through a spatial clustering algorithm, and generating demand spatial distribution identifiers.

[0059] Spatial clustering algorithms can group demand location identifiers with similar spatial locations into the same cluster. In this embodiment, a spatial clustering algorithm is used to process the demand location identifiers in the material request records, dividing the demand locations into multiple demand-intensive regions. For example, the DBSCAN (density-based spatial clustering application) algorithm can be used, which divides data points into clusters based on the density of the demand locations. Through regional clustering, the spatial distribution of demand can be identified and a demand spatial distribution identifier can be generated. This demand spatial distribution identifier can reflect the concentration of demand in different regions.

[0060] Step S234: associate and bind the demand urgency identifier with the demand spatial distribution identifier to generate a demand scenario feature vector reflecting the demand urgency and spatial distribution.

[0061] After obtaining the demand urgency identifier and demand spatial distribution identifier, they are associated and bound, combining them into a vector to generate a demand scenario feature vector that reflects both demand urgency and spatial distribution. This feature vector comprehensively reflects the characteristics of the demand scenario, combining demand urgency and spatial distribution information. During the association and binding process, the demand urgency identifier and demand spatial distribution identifier are arranged in sequence to form a vector of fixed length.

[0062] Step S235: Normalize the demand scenario feature vector, and map the time interval value of the demand urgency identifier and the regional density value of the demand spatial distribution identifier into standardized values ​​within a preset range through feature normalization rules to obtain demand scenario association features reflecting demand urgency and spatial distribution.

[0063] The feature normalization rule can map the time interval value of the demand urgency identifier and the regional density value of the demand spatial distribution identifier to a preset range, such as [0, 1]. Through normalization, the dimensional differences between different features can be eliminated, making the elements in the feature vector comparable. In this embodiment, the minimum-maximum normalization method can be used to normalize the time interval value of the demand urgency identifier and the regional density value of the demand spatial distribution identifier. Specifically, for each feature value, its minimum value is subtracted, and then divided by the difference between the maximum and minimum values ​​to obtain the normalized feature value. After normalization, a demand scenario association feature reflecting the demand urgency and spatial distribution is obtained, and this feature can be used more accurately for subsequent analysis and processing.

[0064] Step S240: Input the material type-related features, transportation path-related features, and demand scenario-related features into the feature alignment network for dimensional unification processing. Through the feature dimension mapping rules, the storage scale dimension of the material type-related features, the timeliness dimension of the transportation path-related features, and the urgency dimension of the demand scenario-related features are mapped to the same numerical representation range to obtain a dimensionally aligned emergency resource feature set.

[0065] The feature alignment network can map features of different dimensions to the same numerical range. In this embodiment, the feature alignment network can adopt a multi-layer perceptron (MLP) architecture, consisting of an input layer, hidden layers, and an output layer. The input layer receives features associated with material type, transportation route, and demand scenario. The hidden layer performs a nonlinear transformation on the input features, and the output layer outputs dimensionally aligned features. Feature dimension mapping rules are used to map features of different dimensions. They can be mapped to the same numerical range based on their characteristics and distribution. For example, a linear mapping method can be used to linearly transform the storage scale dimension of the material type feature, the timeliness dimension of the transportation route feature, and the urgency dimension of the demand scenario feature, so that they have the same numerical range. Through the processing of the feature alignment network and feature dimension mapping rules, a dimensionally aligned set of emergency resource features is obtained, facilitating subsequent data analysis and processing.

[0066] Step S250: Redundant information is filtered out of the dimensionally aligned emergency resource feature set, and repeated material type identifiers in the storage scale dimension, redundant location coordinate identifiers in the timeliness dimension, and repeated time difference identifiers in the urgency dimension are eliminated through feature importance evaluation rules to obtain an emergency resource feature set with semantic consistency.

[0067] The redundant information filtering process is to remove redundant information from the dimensionally aligned emergency resource feature set and improve the quality and analyzability of the features. The feature importance evaluation rule is a rule used to evaluate the importance of features. It can determine whether a feature is redundant information based on the degree of influence of the feature on the analysis results. In this embodiment, by constructing a feature importance evaluation matrix, the evaluation matrix includes the storage scale change sensitivity index of the material type-related feature, the timeliness fluctuation impact index of the transportation path-related feature, and the urgency response criticality index of the demand scenario-related feature. Each feature item in the dimensionally aligned emergency resource feature set is subjected to indicator quantification processing, and the storage scale change sensitivity value of each material type identifier in the storage scale dimension, the timeliness fluctuation impact value of each location coordinate identifier in the timeliness dimension, and the urgency response criticality value of each time difference identifier in the urgency dimension are calculated. Based on the feature importance evaluation matrix, a first importance threshold is set for the storage scale dimension, a second importance threshold is set for the timeliness dimension, and a third importance threshold is set for the urgency dimension. Duplicate material type identifiers with sensitivity values ​​below the first importance threshold in the storage scale dimension, redundant location coordinate identifiers with impact values ​​below the second importance threshold in the timeliness dimension, and duplicate time difference identifiers with criticality values ​​below the third importance threshold in the urgency dimension are screened out. Feature deletion removes these redundant feature items, retaining core feature items with sensitivity, impact, and criticality values ​​above the corresponding thresholds to obtain a semantically consistent set of emergency resource features.

[0068] As an implementation method, step S250 performs redundant information filtering on the dimensionally aligned emergency resource feature set, and removes duplicate material type identifiers in the storage scale dimension, redundant location coordinate identifiers in the timeliness dimension, and duplicate time difference identifiers in the urgency dimension through feature importance evaluation rules. Specifically, the following steps S251 to S255 can be implemented:

[0069] Step S251: Construct a feature importance evaluation matrix, which includes the storage scale change sensitivity index of the material type-related feature, the timeliness fluctuation impact index of the transportation path-related feature, and the urgency response criticality index of the demand scenario-related feature.

[0070] The feature importance evaluation matrix is ​​a matrix used to evaluate the importance of different features. It contains multiple indicators to measure the degree of influence of different features on the analysis results. In this embodiment, the storage scale change sensitivity index of the material type-related feature is used to measure the degree of influence of the change in the material storage scale on the overall emergency resource situation; the timeliness fluctuation impact index of the transportation path-related feature is used to measure the degree of influence of the fluctuation of the timeliness of the transportation path on the transportation of materials and emergency response; the urgency response criticality index of the demand scenario-related feature is used to measure the criticality of the urgency of the demand scenario to resource allocation and emergency response. When constructing the feature importance evaluation matrix, the weight and value range of each indicator are determined according to the specific business needs and data characteristics. For example, the weight of each indicator can be determined through expert evaluation, historical data statistics, and other methods, and then these indicators can be combined into a matrix.

[0071] Step S252: Perform indicator quantification processing on each feature item in the dimension-aligned emergency resource feature set, calculate the storage scale change sensitivity value of each material type identifier in the storage scale dimension, the timeliness fluctuation impact value of each location coordinate identifier in the timeliness dimension, and the urgency response criticality value of each time difference identifier in the urgency dimension.

[0072] Indicator quantification processing is to perform specific numerical calculations on the importance indicators of feature items in order to more accurately evaluate the importance of features. In this embodiment, the storage scale change sensitivity value of each material type identifier in the storage scale dimension can be determined by calculating the correlation between the rate of change of the material storage scale and other relevant factors. For example, the correlation between the rate of change of the material storage scale and the emergency response effect can be calculated, and the correlation coefficient can be used as the storage scale change sensitivity value. For the timeliness fluctuation impact value of each position coordinate identifier in the timeliness dimension, it can be determined by analyzing the impact of the timeliness fluctuation of the transportation path on the material transportation time and cost. For example, the relationship between the fluctuation of the timeliness of the transportation path and the rate of change of the material transportation cost can be calculated, and the relationship can be used as the timeliness fluctuation impact value. For the urgency response criticality value of each time difference identifier in the urgency dimension, it can be determined by analyzing the relationship between the demand time and the emergency response time. For example, the impact of the difference between the demand time and the emergency response time on the resource allocation effect can be calculated, and the degree of impact can be used as the urgency response criticality value.

[0073] Step S253: Based on the feature importance evaluation matrix, a first importance threshold is set for the storage scale dimension, a second importance threshold is set for the timeliness dimension, and a third importance threshold is set for the urgency dimension.

[0074] The importance threshold is a critical value used to determine whether a feature is redundant information. In this embodiment, corresponding importance thresholds are set for different dimensions based on the indicator weights and value ranges in the feature importance evaluation matrix. For example, for the storage scale dimension, a suitable first importance threshold is determined based on the weight and value range of the storage scale change sensitivity indicator; for the timeliness dimension, a suitable second importance threshold is determined based on the weight and value range of the timeliness fluctuation impact indicator; for the urgency dimension, a suitable third importance threshold is determined based on the weight and value range of the urgency response key indicator. These thresholds can be adjusted according to specific business needs and data characteristics to ensure that redundant features can be accurately screened out.

[0075] Step S254: Filter out duplicate material type identifiers whose sensitivity values ​​in the storage scale dimension are lower than the first importance threshold, redundant location coordinate identifiers whose impact values ​​in the timeliness dimension are lower than the second importance threshold, and duplicate time difference identifiers whose criticality values ​​in the urgency dimension are lower than the third importance threshold.

[0076] After determining the importance threshold of each dimension, the feature items in the dimension-aligned emergency resource feature set are screened. For the storage scale dimension, the storage scale change sensitivity value of each material type identifier is compared with the first importance threshold. If the sensitivity value is lower than the threshold, the material type identifier is determined to be a duplicate material type identifier. For the timeliness dimension, the timeliness fluctuation impact value of each location coordinate identifier is compared with the second importance threshold. If the impact value is lower than the threshold, the location coordinate identifier is determined to be a redundant location coordinate identifier. For the urgency dimension, the urgency response criticality value of each time difference identifier is compared with the third importance threshold. If the criticality value is lower than the threshold, the time difference identifier is determined to be a duplicate time difference identifier. Through this screening method, redundant features that need to be eliminated can be found.

[0077] Step S255: Remove the filtered redundant feature items through feature deletion operation, retain the core feature items whose sensitivity values, influence values ​​and criticality values ​​are all higher than the corresponding thresholds, and obtain an emergency resource feature set with semantic consistency.

[0078] The feature deletion operation is to remove the filtered redundant feature items from the dimensionally aligned emergency resource feature set, and only retain the important core feature items. In this embodiment, the feature deletion operation is implemented by programming, and the duplicate material type identifiers with sensitivity values ​​lower than the first importance threshold in the storage scale dimension, the redundant location coordinate identifiers with influence values ​​lower than the second importance threshold in the timeliness dimension, and the duplicate time difference identifiers with criticality values ​​lower than the third importance threshold in the urgency dimension are deleted from the feature set. After the feature deletion operation, the obtained emergency resource feature set only contains important core feature items, has higher semantic consistency, and is more suitable for subsequent data analysis and processing.

[0079] Step S300: Input the emergency resource feature set into the pre-trained shared relationship modeling model for association analysis and processing to generate a shared association feature set that reflects the resource supply and demand matching relationship.

[0080] The shared relationship modeling model is used to analyze the relationships between emergency resource characteristics. It processes the input emergency resource feature set to generate a shared relationship feature set that reflects the matching relationship between resource supply and demand. In this embodiment, the pre-trained shared relationship modeling model can utilize a deep learning model, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). During the training phase, the model utilizes a large amount of historical data to learn the patterns and regularities in the relationships between emergency resource characteristics. During the association analysis process, the emergency resource feature set is input into the pre-trained shared relationship modeling model. The model analyzes and processes the input features based on the learned patterns and regularities to generate a shared relationship feature set. This shared relationship feature set includes matching relationship features between material type and demand scenario, connection relationship features between transportation route and storage location, and correspondence features between turnover time and demand timeliness. These features can comprehensively reflect the matching relationship between emergency resource supply and demand.

[0081] As an implementation method, the shared association feature set includes matching relationship features between material types and demand scenarios, connection relationship features between transportation routes and storage locations, and corresponding relationship features between circulation time and demand timeliness. In step S300, the emergency resource feature set is input into a pre-trained shared relationship modeling model for association analysis and processing to generate a shared association feature set reflecting the resource supply and demand matching relationship. Specifically, the following steps S310 to S350 can be implemented:

[0082] Step S310: Input the material type association features and demand scenario association features in the emergency resource feature set into the first association analysis layer of the shared relationship modeling model, calculate the matching degree between different material types and demand scenarios through preset matching analysis rules, and generate matching relationship features between material types and demand scenarios.

[0083] The first association analysis layer of the shared relationship modeling model is a layer in the model specifically used to analyze the matching relationship between material types and demand scenarios. The preset matching analysis rules are rules for calculating the matching degree between material types and demand scenarios. The matching degree calculation method can be determined based on the characteristics of the material type association features and the demand scenario association features. In this embodiment, the preset matching analysis rules can consider the degree of matching between factors such as the purpose, specifications, and quantity of the material and factors such as the demand quantity, demand location, and demand time of the demand scenario. For example, the similarity between the material attributes in the material type association features and the demand attributes in the demand scenario association features can be calculated, and the similarity can be used as a measure of matching. After the material type association features and the demand scenario association features are input into the first association analysis layer, the layer will calculate the matching degree between different material types and demand scenarios according to the preset matching analysis rules, and then organize and encode the matching degree information to generate the matching relationship features between the material type and the demand scenario.

[0084] Step S320: Input the transportation path association features in the emergency resource feature set and the storage location information in the material type association features into the second association analysis layer of the shared relationship modeling model, calculate the connection degree between the transportation path and the storage location through the preset connection analysis rules, and generate the connection relationship features between the transportation path and the storage location.

[0085] The second association analysis layer of the shared relationship modeling model is used to analyze the connectivity between transportation paths and storage locations. Pre-set connectivity analysis rules are used to calculate the connectivity between transportation paths and storage locations. The connectivity calculation method can be determined based on the storage location information in the transportation path association features and the material type association features. In this embodiment, the preset connectivity analysis rules can take into account factors such as the transportation path's starting location, the spatial distance between the target location and the storage location, and the transportation time. For example, by calculating the spatial distance between the starting location and the storage location, and the spatial distance between the target location and the demand location identifier in the demand scenario association features, a transportation path connectivity calculation function is constructed based on these distance values. The output value of the connectivity calculation function is negatively correlated with the distance from the starting location to the storage location and the distance from the target location to the demand location. After the storage location information in the transportation path association features and the material type association features are input into the second association analysis layer, the layer calculates the connectivity between the transportation path and the storage location according to the preset connectivity analysis rules. The connectivity information is then organized and encoded to generate a connectivity relationship feature for the transportation path and storage location.

[0086] As an implementation method, step S320 inputs the transportation path association feature in the emergency resource feature set and the storage location information in the material type association feature into the second association analysis layer of the shared relationship modeling model, calculates the connection degree between the transportation path and the storage location through a preset connection analysis rule, and generates the connection relationship feature between the transportation path and the storage location. Specifically, the following steps S321 to S325 can be implemented:

[0087] Step S321: extracting the starting location identifier and the target location identifier from the transport path association feature, and extracting the storage location identifier from the resource type association feature.

[0088] In this embodiment, the transport path association feature includes the transport path's starting and destination location information, and the material type association feature includes the material's storage location information. By parsing and processing the transport path association feature and the material type association feature, the starting location identifier, destination location identifier, and storage location identifier are extracted. These identifiers can include geographic coordinate information, address information, and so on. When extracting identifiers, appropriate extraction methods are used based on the specific data format and storage method. For example, if the data is stored in a table format, the corresponding identifier information can be extracted using column indexing; if the data is stored in text format, the identifier information can be extracted using string matching.

[0089] Step S322: Calculate the spatial distance between the starting location identifier and the storage location identifier, and calculate the spatial distance between the target location identifier and the required location identifier in the required scenario association feature.

[0090] After extracting the starting location identifier, target location identifier, storage location identifier, and required location identifier, the spatial distance value between them is calculated. The spatial distance value can reflect the spatial relationship between the transportation path and the storage location and the required location. In this embodiment, the distance calculation method in the geographic information system (GIS) can be used to calculate the spatial distance value. For example, if the identifier is the coordinate information of the geographic location, the Euclidean distance formula or the spherical distance formula can be used to calculate the distance. Specifically, for the starting location identifier and the storage location identifier, their coordinates are substituted into the distance calculation formula to calculate the spatial distance value between them; for the target location identifier and the required location identifier, their coordinates are also substituted into the distance calculation formula to calculate the spatial distance value between them.

[0091] Step S323: Construct a transport path connectivity calculation function based on the spatial distance value between the starting position and the storage position and the spatial distance value between the target position and the required position. The output value of the connectivity calculation function is negatively correlated with the distance from the starting position to the storage position and negatively correlated with the distance from the target position to the required position.

[0092] The transport path connection degree calculation function is a function for calculating the connection degree between the transport path and the storage location, the input is the spatial distance value between the starting location and the storage location and the spatial distance value between the target location and the demand location, and the output is the connection degree value between the transport path and the storage location. In this embodiment, the output value of the connection degree calculation function is negatively correlated with the distance from the starting location to the storage location, meaning that the farther the distance from the starting location to the storage location, the lower the connection degree; and negatively correlated with the distance from the target location to the demand location, meaning that the farther the distance from the target location to the demand location, the lower the connection degree. The transport path connection degree calculation function can be constructed in a linear combination manner, for example, the spatial distance value between the starting location and the storage location and the spatial distance value between the target location and the demand location are weighted and summed, and then the reciprocal thereof is taken as the connection degree value. The weight can be adjusted according to specific business requirements and data characteristics.

[0093] Step S324: Calculate the connection degree value between each transport path and the corresponding storage location through the connection degree calculation function.

[0094] After the transport path connection degree calculation function is constructed, the spatial distance value between the starting location and the storage location and the spatial distance value between the target location and the demand location of each transport path are substituted into the connection degree calculation function to calculate the connection degree value between each transport path and the corresponding storage location. In this embodiment, batch calculation can be realized through programming, taking all the relevant distance values of the transport paths as input, calling the connection degree calculation function, and obtaining the connection degree value of each transport path. The calculated connection degree value can be used for subsequent analysis and decision-making, for example, selecting the transport path with higher connection degree for material transportation.

[0095] Step S325: Generate the connection relationship feature between the transport path and the storage location according to the size of the connection degree value, which includes the transport path identifier, the storage location identifier and the corresponding connection degree value.

[0096] After calculating the connection degree value between each transport path and the corresponding storage location, the transport path identifier, the storage location identifier and the corresponding connection degree value are combined to generate the connection relationship feature between the transport path and the storage location. The transport path identifier is used to uniquely identify the transport path, the storage location identifier is used to uniquely identify the storage location, and the connection degree value is used to represent the connection degree between the transport path and the storage location. In this embodiment, the transport path identifier, the storage location identifier and the connection degree value can be stored in the form of a table to form a data set containing the connection relationship between the transport path and the storage location. This data set can be used as part of the shared association feature set for subsequent analysis and processing.

[0097] Step S330: Input the flow time information in the emergency resource feature set and the demand timeliness information in the demand scenario association feature into the third association analysis layer of the shared relationship modeling model, calculate the correspondence between the flow time and the demand timeliness through the preset corresponding analysis rules, and generate the corresponding relationship characteristics between the flow time and the demand timeliness.

[0098] The third association analysis layer of the shared relationship modeling model is a layer in the model used to analyze the correspondence between the flow time and the demand timeliness. The preset corresponding analysis rules are rules for calculating the correspondence between the flow time and the demand timeliness. The calculation method of the correspondence can be determined based on the characteristics of the flow time information and the demand timeliness information. In this embodiment, the preset corresponding analysis rules can take into account factors such as the time difference and time overlap between the flow time and the demand timeliness. For example, the absolute value of the time difference between the flow time and the demand timeliness can be calculated and used as a measure of the correspondence. The smaller the time difference, the higher the correspondence. After the flow time information in the emergency resource feature set and the demand timeliness information in the demand scenario association feature are input into the third association analysis layer, the layer will calculate the correspondence between the flow time and the demand timeliness according to the preset corresponding analysis rules, and then organize and encode the correspondence information to generate the correspondence relationship characteristics between the flow time and the demand timeliness.

[0099] Step S340: Input the matching relationship characteristics between material types and demand scenarios, the connection relationship characteristics between transportation routes and storage locations, and the corresponding relationship characteristics between circulation time and demand timeliness into the feature fusion layer of the shared relationship modeling model, and fuse the three types of relationship characteristics through feature fusion rules to generate a shared association feature set that comprehensively reflects the resource supply and demand matching relationship.

[0100] The feature fusion layer of the shared relationship modeling model is a layer in the model that is used to fuse different types of relationship features. The feature fusion rules are rules used to fuse the matching relationship features between material types and demand scenarios, the connection relationship features between transportation routes and storage locations, and the corresponding relationship features between circulation time and demand timeliness. The fusion method can be determined based on the importance and relevance of different features. In this embodiment, the feature fusion rules can use a weighted summation method to assign a weight to each relationship feature, and then perform weighted summation on them to obtain a comprehensive set of shared association features. The weights can be adjusted according to specific business needs and data characteristics. For example, the weight of each relationship feature can be determined through experiments or expert evaluation methods. After the three types of relationship features are input into the feature fusion layer, the layer will fuse these features according to the feature fusion rules to generate a shared association feature set that comprehensively reflects the resource supply and demand matching relationship.

[0101] Step S350: Confidence verification processing is performed on the shared association feature set. The similarity between the current shared association feature set and the historical effective shared association feature set is verified by a historical data verification rule. Abnormal relationship features with a similarity lower than a preset threshold are removed. A final shared association feature set is obtained.

[0102] The confidence verification processing is performed to ensure the reliability and accuracy of the shared association feature set. The abnormal relationship features are removed by comparing with the historical data. The historical data verification rule is a rule for comparing the similarity between the current shared association feature set and the historical effective shared association feature set. The similarity can be calculated according to the properties and values of the features. In this embodiment, the cosine similarity, Euclidean distance, or other methods can be used to calculate the similarity. The preset threshold is a critical value for determining whether the similarity meets the requirements. If the similarity is lower than the preset threshold, the relationship feature is considered abnormal and is removed. When the confidence verification processing is performed, the similarity of each relationship feature is calculated by comparing the current shared association feature set with the historical effective shared association feature set. Then, the abnormal relationship features with a similarity lower than the preset threshold are selected and removed from the shared association feature set. After the confidence verification processing, the final shared association feature set is obtained, which has higher reliability and accuracy.

[0103] Step S400: An emergency resource data sharing rule set is generated based on the shared association feature set.

[0104] The emergency resource data sharing rule set is a set of rules for guiding the sharing of emergency resource data, including matching sharing rules of material types and demand scenarios, connection sharing rules of transportation paths and storage locations, and corresponding sharing rules of flow time and demand time limit. In this embodiment, the process of generating the emergency resource data sharing rule set based on the shared association feature set is to formulate specific sharing rules according to the information in the shared association feature set. For example, according to the matching relationship features of material types and demand scenarios, the matching sharing rules of material types and demand scenarios are formulated to specify which material types should be prioritized to meet which demand scenarios. According to the connection relationship features of transportation paths and storage locations, the connection sharing rules of transportation paths and storage locations are formulated to specify which transportation paths should be connected with which storage locations. According to the corresponding relationship features of flow time and demand time limit, the corresponding sharing rules of flow time and demand time limit are formulated to specify the matching relationship between the flow time and the demand time limit.

[0105] As an implementation method, the emergency resource data sharing rule set includes matching sharing rules for material types and demand scenarios, connecting sharing rules for transportation routes and storage locations, and corresponding sharing rules for flow time and demand timeliness. Step S400 generates the emergency resource data sharing rule set based on the shared association feature set, which can be specifically implemented as the following steps S410-S450:

[0106] Step S410: extract the matching relationship features between the material type and the demand scenario in the shared association feature set, and establish a priority matching rule between the material type and the demand scenario in descending order of matching degree as the matching shared rule between the material type and the demand scenario.

[0107] In this embodiment, matching relationship features between material types and demand scenarios are extracted from a shared association feature set. These features contain information on the degree of matching between different material types and demand scenarios. These features are sorted in descending order of matching degree, with combinations of material types and demand scenarios with high matching degrees placed in front and combinations with low matching degrees placed in the back. Then, a priority matching rule between material types and demand scenarios is established based on the sorting results, that is, material types with high matching degrees are preferentially assigned to corresponding demand scenarios. This priority matching rule can be used as a matching sharing rule between material types and demand scenarios to guide the sharing of emergency resource data.

[0108] Step S420: extracting connection relationship features between the transport path and the storage location from the shared association feature set, and establishing priority connection rules between the transport path and the storage location according to the descending order of connection degree as the connection sharing rules between the transport path and the storage location.

[0109] From the shared association feature set, we extract features that link transport paths and storage locations. These features contain information about the degree of connectivity between different transport paths and storage locations. We then sort these features in descending order of connectivity, prioritizing transport path and storage location combinations with high connectivity and those with low connectivity. We then establish a priority connection rule for transport paths and storage locations based on the sorting results, prioritizing transport paths with high connectivity and their corresponding storage locations. This priority connection rule can be used as a shared rule for connecting transport paths and storage locations, guiding the sharing of emergency resource data.

[0110] Step S430: extracting the correspondence features between the flow time and the demand timeliness in the shared association feature set, and establishing a priority correspondence rule between the flow time and the demand timeliness according to the descending order of the correspondence degree as the corresponding shared rule between the flow time and the demand timeliness.

[0111] The corresponding relationship features of the flow time and the demand timeliness are extracted from the shared association feature set, and these features contain the corresponding degree information of different flow times and demand timeliness. According to the descending order of the corresponding degree, these features are sorted, and the combination of the flow time and the demand timeliness with high corresponding degree is arranged in front, and the combination with low corresponding degree is arranged in back. Then, the priority corresponding rule of the flow time and the demand timeliness is established according to the sorting result, that is, the combination of the flow time and the corresponding demand timeliness with high corresponding degree is preferentially selected for matching. This priority corresponding rule can be used as the corresponding sharing rule of the flow time and the demand timeliness to guide the sharing of the emergency resource data.

[0112] Step S440: Conflict detection processing is performed on the matching sharing rule, the connection sharing rule and the corresponding sharing rule, and a rule conflict detection algorithm is used to identify contradictory provisions about the same material identifier or location identifier in different rules, and the priority order of the conflicting rules is adjusted to eliminate the contradiction.

[0113] The rule conflict detection algorithm is an algorithm for detecting whether there is a conflict between different rules, and can identify contradictory provisions about the same material identifier or location identifier in different rules. In this embodiment, a rule conflict detection table is established, the matching sharing rule, the connection sharing rule and the corresponding sharing rule are filled into the corresponding fields of the rule conflict detection table, and the material identifier and the location identifier involved in each rule are recorded. Then, the rule conflict detection table is traversed to identify different rules with the same material identifier or location identifier. Content comparison processing is performed on the identified conflicting rules to determine whether there is a matching priority conflict for the same material identifier or a connection priority conflict for the same location identifier. If there is a conflict, according to a preset rule priority adjustment strategy, the priority of the matching sharing rule is set to be higher than that of the connection sharing rule and the corresponding sharing rule, the priority of the connection sharing rule is set to be higher than that of the corresponding sharing rule, and the priority order of the conflicting rules is adjusted to eliminate the contradiction. Finally, the adjusted rule conflict detection table is traversed again to ensure that there is no contradictory provision for the same material identifier or location identifier, and the conflict detection processing is completed.

[0114] As an implementation manner, in step S440, the conflict detection processing is performed on the matching sharing rule, the connection sharing rule and the corresponding sharing rule, the rule conflict detection algorithm is used to identify contradictory provisions about the same material identifier or location identifier in different rules, and the priority order of the conflicting rules is adjusted to eliminate the contradiction, which can be implemented as the following steps S441-S446:

[0115] Step S441: A rule conflict detection table is established, which includes a material identifier field, a location identifier field and a rule content field. The rule conflict detection table is a table for recording and detecting rule conflicts, which includes a material identifier field, a location identifier field and a rule content field.

[0116] The Material ID field records the material IDs involved in each rule, the Location ID field records the location IDs involved in each rule, and the Rule Content field records the specific content of each rule. When creating a rule conflict detection table, determine the table size and format based on the number of rules and the length of the fields. For example, you could create a table using spreadsheet software, with the Material ID field, Location ID field, and Rule Content field as column headers, and then fill in the corresponding cells with the relevant information for each rule.

[0117] Step S442: Fill in the matching sharing rules, connection sharing rules and corresponding sharing rules into the corresponding fields of the rule conflict detection table respectively, and record the material identifier and location identifier involved in each rule.

[0118] After creating the rule conflict detection table, organize and analyze the matching shared rules, connection shared rules, and corresponding shared rules. Extract the material and location identifiers involved in each rule, and then enter this information into the corresponding fields of the rule conflict detection table. Also, enter the specific content of each rule into the rule content field. When entering information, ensure its accuracy and completeness to avoid omissions or errors.

[0119] Step S443: traverse the rule conflict detection table to identify different rules with the same material identifier or location identifier.

[0120] After entering all rule information into the rule conflict detection table, the table is traversed to identify different rules with the same material identifier or location identifier. In this embodiment, a loop statement can be used to traverse each row of the rule conflict detection table, comparing the material identifiers and location identifiers of adjacent rows. If different rules with the same material identifier or location identifier are found, these rules are marked as conflicting rules. By traversing the rule conflict detection table, all potentially conflicting rules can be found.

[0121] Step S444: performing content comparison processing on the identified conflicting rules to determine whether there is a conflict in matching priority for the same material identifier or a conflict in connection priority for the same location identifier.

[0122] After the conflicting rules are identified, the contents of the rules are compared to determine whether there is a match priority conflict for the same material identifier or a connection priority conflict for the same location identifier. In this embodiment, the specific contents of the rules can be analyzed to compare the priority provisions of different rules for the same material identifier or location identifier. For example, if one rule provides that a certain material identifier should be assigned to demand scenario A with priority, and another rule provides that the material identifier should be assigned to demand scenario B with priority, there is a match priority conflict for the same material identifier. If one rule provides that a certain location identifier should be connected to transport path C with priority, and another rule provides that the location identifier should be connected to transport path D with priority, there is a connection priority conflict for the same location identifier.

[0123] Step S445: If there are conflicting rules, the priority of the matching shared rule is set to be higher than the connection shared rule and the corresponding shared rule, the priority of the connection shared rule is set to be higher than the corresponding shared rule according to the preset rule priority adjustment strategy, and the priority order of the conflicting rules is adjusted to eliminate the conflict.

[0124] After it is determined that there are conflicting rules, the priority order of the conflicting rules is adjusted to eliminate the conflict according to the preset rule priority adjustment strategy. In this embodiment, the preset rule priority adjustment strategy is to set the priority of the matching shared rule to be higher than the connection shared rule and the corresponding shared rule, and set the priority of the connection shared rule to be higher than the corresponding shared rule. According to this strategy, the priority of the conflicting rules is adjusted. For example, if the matching shared rule conflicts with the connection shared rule, the priority of the matching shared rule is increased; if the connection shared rule conflicts with the corresponding shared rule, the priority of the connection shared rule is increased. By adjusting the priority order, it can be ensured that there is no longer a conflict between different rules.

[0125] Step S446: The adjusted rule conflict detection table is traversed again to ensure that there is no contradictory provision for the same material identifier or location identifier, and the conflict detection process is completed.

[0126] After the priority order of the conflicting rules is adjusted, the adjusted rule conflict detection table is traversed again to ensure that there is no contradictory provision for the same material identifier or location identifier. In this embodiment, the rule conflict detection table is traversed again, the material identifiers and location identifiers of adjacent rows are compared, and it is checked whether there are still conflicting rules. If it is found that there are still conflicting rules, the priority order of the rules needs to be further adjusted until there is no contradictory provision. After the second traversal verification is completed, the conflict detection process is completed, and the adjusted shared rule set is obtained.

[0127] Step S450: perform coverage verification processing on the adjusted shared rule set to ensure that all material types of the reserve nodes, all transportation paths of the transportation nodes, and all demand scenarios of the demand nodes are covered by at least one shared rule through rule coverage verification strategy, and obtain a final emergency resource data shared rule set.

[0128] The rule coverage verification strategy is an algorithm for verifying whether the shared rule set can cover all related information, which can ensure that all material types of the reserve nodes, all transportation paths of the transportation nodes, and all demand scenarios of the demand nodes are covered by at least one shared rule. In this embodiment, the material types of the reserve nodes, the transportation paths of the transportation nodes, and the demand scenarios of the demand nodes are compared with the adjusted shared rule set respectively, and it is checked whether there is uncovered information. If there is uncovered information, the shared rule set needs to be further adjusted and improved, for example, new rules are added or existing rules are modified, to ensure that all information is covered. Through the rule coverage verification strategy, the coverage verification processing is performed on the adjusted shared rule set, and a final emergency resource data shared rule set is obtained, which can effectively guide the sharing of emergency resource data.

[0129] Step S500: perform the sharing operation of multi-source emergency resource data according to the emergency resource data shared rule set.

[0130] In this embodiment, the emergency resource data shared rule set is a rule set generated and verified through the previous steps, which contains the matching shared rule of material type and demand scenario, the connection shared rule of transportation path and storage location, and the corresponding shared rule of flow time and demand time limit. Performing the sharing operation of multi-source emergency resource data according to these rule sets is to share and distribute the material storage records of the reserve nodes, the material flow records of the transportation nodes, and the material request records of the demand nodes according to the rules.

[0131] As an implementation manner, step S500, performing the sharing operation of multi-source emergency resource data according to the emergency resource data shared rule set, can be implemented as steps S510-S550 as follows:

[0132] Step S510: extract the matching shared rule of material type and demand scenario from the emergency resource data shared rule set, and push the material type data with the highest matching degree in the material storage record of the reserve node to the demand node of the corresponding demand scenario according to the matching shared rule.

[0133] In this embodiment, sharing rules for matching material types with demand scenarios are extracted from a set of emergency resource data sharing rules. These rules specify which material types should prioritize which demand scenarios. Then, based on these rules, the material storage records of the reserve nodes are screened to identify the material type data that best matches the demand scenario. This data is then pushed to the demand node for the corresponding demand scenario in accordance with the rules. During data push, accuracy and timeliness are ensured to meet the needs of the demand node.

[0134] Step S520: Extract the connection sharing rules between the transportation path and the storage location from the emergency resource data sharing rule set, and push the transportation path data with the highest connection degree in the material flow record of the transportation node to the reserve node of the corresponding storage location according to the connection sharing rules.

[0135] From the set of emergency resource data sharing rules, shared rules for connecting transportation routes with storage locations are extracted. These rules specify which transportation routes should prioritize which storage locations. Based on these rules, the material flow records at the transportation nodes are screened to identify the transportation routes with the highest degree of connection to the storage locations. These transportation routes with the highest degree of connection are then pushed to the corresponding storage nodes in accordance with the rules. When pushing data, factors such as the timeliness and cost of the transportation routes are considered to ensure that materials are efficiently transported to the storage nodes.

[0136] Step S530: extract the corresponding sharing rules of flow time and demand timeliness from the emergency resource data sharing rule set, and push the flow time data with the highest correspondence in the material flow record of the transportation node to the demand node corresponding to the demand timeliness according to the corresponding sharing rules.

[0137] From the emergency resource data sharing rule set, we extract the corresponding sharing rules for flow times and demand timeliness. These rules specify which flow times should be prioritized for matching with which demand timeliness. Based on these rules, we filter the material flow records at the transportation nodes to identify the flow time data that best matches the demand timeliness. These flow time data with the highest degree of correspondence are then pushed to the demand node corresponding to the demand timeliness, as required by the rules. When pushing data, we ensure that the flow time meets the demand timeliness requirements to improve the efficiency of the emergency response.

[0138] Step S540: Standardize the data formats of the material type data pushed to the demand node, the transportation route data pushed to the reserve node, and the flow time data pushed to the demand node, and unify the data formats of different nodes into a preset standard data format through data format conversion rules.

[0139] Data format standardization ensures data compatibility and exchange between different nodes. Data format conversion rules are used to standardize the material type data pushed to the demand node, the transportation route data pushed to the reserve node, and the transit time data pushed to the demand node into a pre-set standard data format. In this embodiment, data format conversion rules can be formulated based on different data types and node requirements. For example, material type data can be converted to a unified encoding format; transportation route data can be converted to a unified coordinate format; and transit time data can be converted to a unified time format. Data format standardization improves data usability and analyzability.

[0140] Step S550: Perform access rights verification on the shared data in the unified format. Through the permission verification rules, set read-only access rights for the reserve node to the demand data of the demand node, set read and write access rights for the transportation node to the reserve node storage data and the demand node demand data, and set read-only access rights for the demand node to the reserve node storage data and the transportation node flow data, thereby completing the sharing operation of multi-source emergency resource data.

[0141] The authority verification rules are rules used to verify the access rights of different nodes to shared data, which can ensure the security and confidentiality of data. In this embodiment, read-only access rights are set for the reserve node to the demand data of the demand node, which means that the reserve node can only view the demand data of the demand node and cannot modify it; read and write access rights are set for the transport node to the storage data of the reserve node and the demand data of the demand node, which means that the transport node can view and modify the storage data of the reserve node and the demand data of the demand node; read-only access rights are set for the demand node to the storage data of the reserve node and the flow data of the transport node, which means that the demand node can only view the storage data of the reserve node and the flow data of the transport node and cannot modify it. The shared data in the unified format is verified for access rights through the authority verification rules to complete the sharing operation of multi-source emergency resource data, ensuring the security and controllability of data during the sharing process.

[0142] It is understandable that the various algorithms involved in the above-mentioned introductions of the embodiments of the present invention, such as the Euclidean distance algorithm, the cosine distance algorithm, the conflict detection algorithm, etc., can all be learned from the relevant content in the prior art. In order to save space, they will not be expanded too much in the embodiments of the present invention. In addition, when implementing the scheme of the present invention, those skilled in the art can supplement the details according to the common knowledge in this field. For example, according to the common knowledge in this field, normalization can be used to eliminate dimensional conflicts before feature fusion, interpolation can be used to eliminate dimensional differences, and thresholds can be reasonably set based on historical data, experience or business scenario requirements. The model can be trained based on a general model training method, and the number of layers in the model structure can be set based on actual needs, the activation function can be selected, etc. The present invention will no longer provide redundant introductions to the overly detailed implementation process.

[0143] See Figure 3 , is a schematic diagram of the structure of a server provided by an embodiment of the present invention. Figure 3 As shown, the server 10 may include: a processor 1001, a network interface 1004 and a memory 1005. In addition, the server 10 may also include: a user interface 1003, and at least one communication bus 1002. The communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen and a keyboard. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The memory 1005 may optionally be at least one storage device away from the aforementioned processor 1001. As Figure 3 As shown, the memory 1005 as a computer readable storage medium may include an operating system, a network communication module, a user interface module and a device control application. It should be understood that the server 10 described in the embodiment of the present invention can execute the above Figure 2 The description of the multiple deep learning-based emergency resource data sharing methods in the corresponding embodiments will not be repeated here. In addition, the description of the beneficial effects of using the same method will not be repeated here.

[0144] An embodiment of the present invention also provides a computer storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the emergency resource data sharing method based on deep learning in any of the above embodiments.

[0145] It should be noted here that the description of the above storage medium and device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and device embodiments of the present invention, please refer to the description of the method embodiments of the present invention for understanding. The above processor can be at least one of a target application integrated circuit (Application Specific Integrated Circuit, ASIC), a digital signal processor (Digital Signal Processor, DSP), a digital signal processing device (Digital Signal Processing Device, DSPD), a programmable logic device (Programmable Logic Device, PLD), a field programmable gate array (Field Programmable Gate Array, FPGA), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, and a microprocessor. It is understandable that the electronic device that implements the above processor function can also be other, and the embodiments of the present invention are not specifically limited.

[0146] The above-mentioned computer storage medium / memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); it can also be various terminals including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

Claims

1. A method for sharing emergency resource data based on deep learning, characterized in that: include: Obtaining a collection of original resource data distributed across different emergency response nodes; Performing feature unification processing on the original resource data set to generate an emergency resource feature set with semantic consistency; Inputting the emergency resource feature set into a pre-trained shared relationship modeling model for association analysis and processing to generate a shared association feature set reflecting the resource supply and demand matching relationship; generating an emergency resource data sharing rule set based on the shared association feature set; Extracting a matching sharing rule between material types and demand scenarios from the emergency resource data sharing rule set, and pushing the material type data with the highest matching degree in the material storage records of the reserve node to the demand node corresponding to the demand scenario according to the matching sharing rule; Extracting a connection sharing rule between the transport path and the storage location from the emergency resource data sharing rule set, and pushing the transport path data with the highest connection degree in the material flow records of the transport node to the reserve node corresponding to the storage location according to the connection sharing rule; Extracting corresponding sharing rules between flow time and demand timeliness from the emergency resource data sharing rule set, and pushing the flow time data with the highest correspondence in the material flow records of the transportation node to the demand node corresponding to the demand timeliness according to the corresponding sharing rules; Standardize the data formats of material type data pushed to demand nodes, transportation route data pushed to reserve nodes, and flow time data pushed to demand nodes. Use data format conversion rules to unify the data formats of different nodes into a preset standard data format. Access rights verification is performed on the shared data in the unified format. Through the permission verification rules, read-only access rights are set for the reserve node to the demand data of the demand node, read-write access rights are set for the transportation node to the reserve node storage data and the demand node demand data, and read-only access rights are set for the demand node to the reserve node storage data and the transportation node flow data, thereby completing the sharing operation of multi-source emergency resource data.

2. The method for sharing emergency resource data based on deep learning according to claim 1, characterized in that: The original resource data set includes material storage records of reserve nodes, material flow records of transportation nodes, and material request records of demand nodes. The original resource data set distributed across different emergency response nodes is obtained, including: Send a storage data pull instruction to the reserve node, and receive the material storage record containing the material identification, storage capacity identification and storage time identification returned by the reserve node; Sending a flow data query instruction to the transport node, and receiving the material flow record returned by the transport node, which includes the material identification, starting location identification, target location identification, and transport time identification; Send a request data synchronization instruction to the demand node, and receive a material request record containing a material identifier, a required quantity identifier, a required location identifier, and a required time identifier returned by the demand node; Performing identifier alignment processing on the material storage records, the material flow records, and the material request records, unifying inconsistent material identifiers in the records into a preset standard material classification code, and unifying mismatched location identifiers in the records into a preset regional coordinate code, thereby obtaining a set of original resource data with consistent identifiers; The original resource data set is calibrated in time dimension. Based on the demand time identifier, the storage time identifier of the storage record is converted into a time difference identifier with the demand time, and the transportation time identifier of the flow record is converted into a schedulable time window identifier before the demand time, so as to obtain the original resource data set aligned in time dimension.

3. The method for sharing emergency resource data based on deep learning according to claim 1, characterized in that: The emergency resource feature set includes material type association features corresponding to storage records, transportation route association features corresponding to flow records, and demand scenario association features corresponding to request records. The feature unification processing of the original resource data set to generate an emergency resource feature set with semantic consistency includes: Performing semantic parsing on the material storage records in the original resource data set, extracting the association between the storage capacity identifier and the material identifier, and generating material type association features reflecting the storage scale of different material types; Performing path analysis on the material flow records in the original resource data set, extracting the association between the starting location identifier, the target location identifier, and the transportation time identifier, and generating a transportation path association feature that reflects the timeliness of the transportation path; Performing scenario recognition processing on the material request records in the original resource data set, extracting the association relationship between the demand quantity identifier, the demand location identifier, and the demand time identifier, and generating demand scenario association features reflecting the urgency and spatial distribution of the demand; Inputting the material type-related features, the transportation path-related features, and the demand scenario-related features into a feature alignment network for dimensional unification processing, and mapping the storage scale dimension of the material type-related features, the timeliness dimension of the transportation path-related features, and the urgency dimension of the demand scenario-related features to the same numerical representation range through feature dimension mapping rules, thereby obtaining a dimensionally aligned emergency resource feature set; Redundant information filtering is performed on the dimensionally aligned emergency resource feature set. The repeated material type identifiers in the storage scale dimension, the redundant location coordinate identifiers in the timeliness dimension, and the repeated time difference identifiers in the urgency dimension are eliminated through the feature importance evaluation rules to obtain an emergency resource feature set with semantic consistency.

4. The method for sharing emergency resource data based on deep learning according to claim 3, characterized in that: The shared association feature set includes the matching relationship features between material types and demand scenarios, the connection relationship features between transportation routes and storage locations, and the corresponding relationship features between circulation time and demand timeliness. The emergency resource feature set is input into the pre-trained shared relationship modeling model for association analysis and processing to generate a shared association feature set reflecting the resource supply and demand matching relationship, including: Inputting the material type association features and the demand scenario association features in the emergency resource feature set into the first association analysis layer of the shared relationship modeling model, calculating the matching degree between different material types and demand scenarios through preset matching analysis rules, and generating matching relationship features between material types and demand scenarios; Inputting the transport path association feature in the emergency resource feature set and the storage location information in the material type association feature into the second association analysis layer of the shared relationship modeling model, calculating the connection degree between the transport path and the storage location through a preset connection analysis rule, and generating a connection relationship feature between the transport path and the storage location; Inputting the flow time information in the emergency resource feature set and the demand timeliness information in the demand scenario association feature into the third association analysis layer of the shared relationship modeling model, calculating the correspondence between the flow time and the demand timeliness through a preset correspondence analysis rule, and generating a correspondence relationship feature between the flow time and the demand timeliness; Input the matching relationship features between the material type and the demand scenario, the connection relationship features between the transportation route and the storage location, and the corresponding relationship features between the circulation time and the demand timeliness into the feature fusion layer of the shared relationship modeling model, and fuse the three types of relationship features through feature fusion rules to generate a shared association feature set that comprehensively reflects the resource supply and demand matching relationship; The shared association feature set is subjected to confidence verification processing, and the similarity between the current shared association feature set and the historical valid shared association feature set is compared through historical data verification rules, and abnormal relationship features with similarity lower than a preset threshold are eliminated to obtain the final shared association feature set.

5. The method for sharing emergency resource data based on deep learning according to claim 1, characterized in that: The emergency resource data sharing rule set includes matching sharing rules for material types and demand scenarios, connection sharing rules for transportation routes and storage locations, and corresponding sharing rules for flow time and demand timeliness. The generation of the emergency resource data sharing rule set based on the shared association feature set includes: Extracting matching relationship features between material types and demand scenarios from the shared association feature set, and establishing priority matching rules between material types and demand scenarios in descending order of matching degree as matching shared rules between material types and demand scenarios; Extracting connection relationship features between the transport paths and the storage locations from the shared association feature set, and establishing priority connection rules between the transport paths and the storage locations according to descending order of connection degrees as the connection sharing rules between the transport paths and the storage locations; Extracting the corresponding relationship features between the flow time and the demand timeliness from the shared association feature set, and establishing a priority corresponding rule between the flow time and the demand timeliness according to the descending order of the corresponding degree as the corresponding shared rule between the flow time and the demand timeliness; Performing conflict detection on the matching sharing rules, the connection sharing rules, and the corresponding sharing rules, identifying conflicting provisions regarding the same material identifier or location identifier in different rules through a rule conflict detection algorithm, and adjusting the priority order of the conflicting rules to eliminate the conflict; The adjusted sharing rule set is subjected to coverage verification. The rule coverage verification strategy is used to ensure that all material types of reserve nodes, all transportation routes of transportation nodes, and all demand scenarios of demand nodes are covered by at least one sharing rule, thus obtaining the final set of emergency resource data sharing rules.

6. The method for sharing emergency resource data based on deep learning according to claim 2, characterized in that: The step of sending a storage data pull instruction to a storage node and receiving a material storage record containing a material identifier, a storage capacity identifier, and a storage time identifier returned by the storage node includes: Constructing a storage data pull instruction including a storage node identifier and a data type identifier, wherein the storage node identifier is used to locate the target storage node, and the data type identifier is used to specify the type of material storage record to be pulled; Sending the storage data pull instruction to the communication interface of the target storage node through the emergency data transmission protocol; Receiving response data returned by the target reserve node through the emergency data transmission protocol, the response data including a material identification field, a storage capacity field, and a storage time field; Performing field parsing on the response data, extracting the material classification code in the material identification field, the specific storage quantity in the storage capacity field, and the storage date in the storage time field, and generating a material storage record including the material identification, storage capacity identification, and storage time identification; The material storage record is subjected to integrity verification processing. The field integrity verification rules are used to check whether the material identification, storage capacity identification and storage time identification all have valid data values. Invalid storage records with missing fields are eliminated to obtain complete material storage records.

7. The method for sharing emergency resource data based on deep learning according to claim 3, characterized in that: The process of performing scenario recognition processing on the material request records in the original resource data set, extracting the association between the demand quantity identifier, the demand location identifier, and the demand time identifier, and generating demand scenario association features reflecting the urgency and spatial distribution of the demand, includes: Calculate the time difference between the demand time identifier in the material request record and calculate the time interval between the demand time identifier and the current time based on the current time; Dividing the demand scenarios into urgency categories according to the time interval values, and generating demand urgency identifiers; Performing regional clustering processing on the demand location identifier in the material request record, dividing the demand location identifier into multiple demand-intensive areas using a spatial clustering algorithm, and generating a demand spatial distribution identifier; Associating and binding the demand urgency identifier with the demand spatial distribution identifier to generate a demand scenario feature vector reflecting the demand urgency and spatial distribution; The demand scenario feature vector is normalized, and the time interval value of the demand urgency identifier and the regional density value of the demand spatial distribution identifier are mapped to standardized values ​​within a preset range through feature normalization rules to obtain demand scenario association features reflecting demand urgency and spatial distribution.

8. The method for sharing emergency resource data based on deep learning according to claim 4, characterized in that: The step of inputting the transport path association feature in the emergency resource feature set and the storage location information in the material type association feature into the second association analysis layer of the shared relationship modeling model, calculating the connection degree between the transport path and the storage location through a preset connection analysis rule, and generating the connection relationship feature between the transport path and the storage location includes: Extracting the starting location identifier and the target location identifier from the transport path association feature, and extracting the storage location identifier from the material type association feature; Calculating the spatial distance between the starting location identifier and the storage location identifier, and calculating the spatial distance between the target location identifier and the required location identifier in the required scenario association feature; Constructing a transport path connectivity calculation function based on the spatial distance between the starting position and the storage position and the spatial distance between the target position and the required position, wherein an output value of the connectivity calculation function is negatively correlated with the distance from the starting position to the storage position and negatively correlated with the distance from the target position to the required position; Calculating the connection value between each transport path and the corresponding storage location using the connection calculation function; A connection relationship feature between the transport path and the storage location is generated according to the size of the connection degree value, and the connection relationship feature includes a transport path identifier, a storage location identifier and a corresponding connection degree value.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images