Carrier portrait evaluation system and method based on multi-dimensional data fusion

Abstract

The invention relates to the technical field of carrier evaluation, in particular to a carrier portrait evaluation system and method based on multi-dimensional data fusion, and the system comprises a data source obtaining module which is used for collecting a plurality of carrier related data sources in real time; the data preprocessing module is used for carrying out normalization preprocessing on the collected data; the multi-dimensional data fusion module is used for constructing a dynamic carrier portrait by adopting a multi-modal fusion algorithm and a hierarchical data modeling technology so as to provide a basis for evaluation; the scene performance evaluation module is used for carrying out comprehensive evaluation on the carriers in a plurality of business scenes, carrying out comprehensive scoring on the plurality of carriers through a multi-index evaluation method, and obtaining a comprehensive score ranking of the plurality of carriers; and the carrier relationship detection module is used for analyzing the association relationship between the preferred carrier and other carriers. According to the invention, the accuracy of evaluation is improved, and the adaptability of the system to a complex business environment is enhanced, so that an enterprise can make a more scientific decision when selecting a carrier.

Description

Technical Field

[0001] The present invention relates to the technical field of carrier evaluation, and in particular to a carrier portrait evaluation system and method based on multi-dimensional data fusion. Background Art

[0002] With the rapid development of the logistics industry, enterprises are facing increasingly complex challenges when choosing carriers. In order to maintain their advantages in the fierce market competition, enterprises must choose carriers that can operate stably and efficiently in various business scenarios. However, traditional carrier evaluation methods usually rely on single-dimensional data (such as historical performance or qualifications) for analysis, lacking in-depth integration and analysis of multi-dimensional data, and it is difficult to fully reflect the actual performance of carriers in different scenarios. In addition, with the expansion of the scale of carriers and the expansion of their business scope, the relationships between carriers (such as cross-shareholders, overlapping management, historical cooperative relationships, etc.) have become increasingly complex, and the risk of potential conflicts of interest has also increased accordingly.

[0003] The existing carrier evaluation system usually adopts a simple weighted scoring or data analysis based on a single dimension, which cannot fully consider the multi-dimensional characteristics of the carrier (such as qualifications, transportation capacity, historical cooperation records, etc.) and the comprehensive impact of these characteristics in different business scenarios. This leads to the evaluation results often deviating from reality, making enterprises face greater uncertainty in the process of carrier selection.

[0004] Most current evaluation methods cannot simulate the actual performance of carriers in complex business scenarios, and it is difficult to predict the performance of carriers in scenarios such as multimodal transport, temporary route planning, and special cargo transportation. This lack of scenario-based evaluation makes it difficult for companies to fully understand the potential problems and risks of carriers in actual operations, which may lead to failed cooperation or increased operating costs.

[0005] In large-scale, multi-level supply chain networks, the complex relationships between carriers have become an important source of potential risks. However, traditional methods usually only rely on manual or simple database queries, and are unable to comprehensively and automatically identify potential relationships and conflicts of interest between carriers. This not only increases the management difficulty of enterprises, but also reduces the transparency and security of the supply chain to a certain extent. Summary of the invention

[0006] Based on the above objectives, the present invention provides a carrier portrait evaluation system and method based on multi-dimensional data fusion.

[0007] The carrier profile evaluation system based on multi-dimensional data fusion includes: Data source acquisition module: used to collect multiple carrier-related data sources in real time, including carrier qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data; Data preprocessing module: used to perform normalization preprocessing on the collected data; Multi-dimensional data fusion module: It fuses the pre-processed multi-source data, uses multi-modal fusion algorithm and hierarchical data modeling technology to build a dynamic carrier portrait, providing a basis for evaluation; Scenario-based performance evaluation module: Based on multi-dimensional carrier portraits, comprehensive evaluation of carriers in multiple business scenarios is conducted, including multimodal transport, temporary route planning, and special cargo transportation. Based on simulation technology and historical data analysis, a multi-index evaluation method is used to evaluate the performance of carriers in different business scenarios. The multi-index includes coordination and integration capability indicators, real-time decision-making capability indicators, and safety management capability indicators. Based on the evaluation results of multiple indicators, a comprehensive score is given; a comprehensive score is given to multiple carriers through a multi-index evaluation method to obtain a comprehensive score ranking of multiple carriers; Carrier relationship detection module: select the carrier with the highest comprehensive score as the preferred carrier, analyze the relationship between the preferred carrier and other carriers, ensure that there is no potential conflict of interest between the preferred carrier and other carriers, use correlation analysis technology based on graph database to detect potential correlations between shareholders, management and business cooperation, and identify implicit correlations through correlation network analysis technology to help companies avoid cooperation risks.

[0008] Optionally, the data source acquisition module includes: Multi-channel data interface submodule: establishes real-time connections with multiple external data sources, including industry regulatory databases, industry association platforms, enterprise management systems, third-party market analysis tools and logistics platforms, and obtains and updates the carrier's qualification data, performance data, scale data, capacity data, historical cooperation years data and cooperation amount data in real time through the data interface; Data push and synchronization submodule: Through the subscription and push mechanism, it automatically receives updated information from external data sources and synchronizes the data to the evaluation system to ensure the real-time and accuracy of each data collection.

[0009] Optionally, the multidimensional data fusion module specifically includes: Data fusion submodule: Integrate the pre-processed multi-source heterogeneous data, adopt a method combining feature-level fusion and data-level fusion, extract key features from different data sources through feature-level fusion, and uniformly map the key features to a shared feature space; then use the data-level fusion method to comprehensively integrate the data in the unified feature space to generate a high-dimensional, unified-format carrier data set that fully retains the information characteristics and inherent relationships of each data source; Hierarchical data modeling submodule: Receives the carrier data set from the multimodal data fusion submodule and uses hierarchical data modeling technology to divide the data into different levels according to data attributes and business logic, including: Basic layer: covers static attribute data of carriers, including qualifications and scale; Behavior layer: covers the carrier’s dynamic operational data, including performance, capacity, historical cooperation years, and cooperation amount; Comprehensive layer: establishes an association between the basic layer and the behavior layer. The static data of the basic layer has a direct impact on the dynamic data of the behavior layer, reflecting the association between the basic layer and the behavior layer; At each level, a hierarchical and well-structured carrier data model is constructed so that each layer of data can independently describe the information of a specific dimension and reflect the inherent connection of the data through inter-layer correlation, thus forming a comprehensive and dynamic carrier portrait and providing accurate and rich data support for subsequent scenario-based performance evaluation.

[0010] Optionally, the feature-level fusion is used to extract key features from different data sources and unify the key features into the same dimension, which is a shared feature space. The data set of each data source is represented as ,in, Index for data source, each data source Include dimensional features, and the most important features (key features) are extracted through the PCA feature selection algorithm, which is expressed as: ,in, For data source The extracted features are then reduced to form a unified feature representation, and then the extracted key features are merged: .

[0011] The data-level fusion is based on the feature-level fusion, and forms a unified data representation by integrating the feature matrices of different data sources. The data-level fusion uses a weighted average method to combine each The feature matrix of is fused according to its weight, expressed as: ,in, For data source The weight of ,By adjusting the weights, the influence of different data sources on the fusion results can be controlled.

[0012] Optionally, the hierarchical data modeling submodule receives the fused carrier data set from the multimodal data fusion submodule, uses the hierarchical data modeling technology to organize different types of data into layers, and establishes associations between layers to form a unified carrier portrait, specifically including: Base layer modeling: The base layer data includes static attributes of carriers. The data of this layer is modeled as a feature matrix: ,in, Represents a static property, Indicates Item static properties; Behavior layer modeling: Behavior layer data includes dynamic attributes of carriers. Dynamic attributes have time series characteristics and are captured through dynamic models. Behavior layer data is formed into behavioral data vectors through time series analysis. ,in, Represents a moment in a time series: , using an autoregressive model to capture the time evolution of the data: ,in, It's time behavioral characteristics, is the regression coefficient, is the error term, is the lag order; The comprehensive layer is used to establish an association between the basic layer and the behavior layer through the weight coefficient To represent the association between the base layer and the behavior layer: ,in, The comprehensive layer represents the dynamic part of the behavior layer. The comprehensive layer integrates the data of the basic layer and the behavior layer to generate a complete carrier profile. The model of the comprehensive layer is: ,in, It is a nonlinear function that combines the features of the base layer and the behavioral layer to generate a comprehensive carrier profile .

[0013] Optionally, the scenario-based performance evaluation module includes: Select a business scenario, including , , , the simulation environment is constructed based on the selected business scenarios. In each selected scenario, the simulation environment is constructed based on historical data and the current environment (such as traffic flow, weather conditions, cargo characteristics, etc.). The simulation environment is constructed as follows: Multidimensional data integration: Profiling carriers The characteristic data (scale, capacity, and historical cooperation years) in the scenario are combined with the specific parameters in the scenario (route complexity, cargo requirements) to form the input parameter set of the scenario. ; Simulation calculation: Based on the input parameter set , using agent-based simulation to simulate the performance of carriers in each business scenario. The output generated by the simulation calculation includes transportation time ,cost , Security Parameters, simulation output is expressed as: ,in, It's a scene The simulation output results are: The accident rate during transportation and cargo integrity rate To measure.

[0014] Optionally, the simulation output The transportation time, cost and safety parameters in the project are quantitatively analyzed through a multi-index evaluation method, including: Coordination and integration capability index: measures the carrier's coordination and integration capabilities between different modes of transport in multimodal transport scenarios, and is quantified by calculating transport time, transshipment efficiency, and transport cost parameters; Real-time decision-making capability indicator: In the temporary route planning scenario, the carrier's decision-making speed and flexibility in responding to emergencies are evaluated, based on the analysis of historical emergency response time and route adjustment success rate data; Safety management capability index: In special cargo transportation scenarios, the carrier's ability to manage cargo safety during transportation is evaluated and quantified by analyzing cargo integrity rate and safety accident rate; Comprehensive score calculation submodule: Based on the multi-indicator evaluation results, the weighted average method is used to comprehensively score the indicators in each scenario.

[0015] Optionally, the coordination and integration capability index is calculated as: ; in, For multimodal transport , For multimodal transport , Calculated as the ratio of each transshipment operation time to the total transportation time: ,in, It is The time of the refitting operation, For the number of changes, for The weight coefficient measures the importance of transportation time in the overall coordination ability. for The weight coefficient measures the importance of dressing efficiency in coordination ability. for The weight coefficient measures the importance of transportation costs in the evaluation; The real-time decision-making capability index is calculated as: ; in, for The weight coefficient measures the importance of response time. for The weight coefficient measures the importance of the successful adjustment rate and the response time The time difference for responding to emergencies; is the route adjustment success rate, calculated as: ; The safety management capability index is calculated as: , is the cargo integrity rate weight coefficient, is the accident rate weight coefficient; Among them, the simulation results Depend on The simulation results show that; Comprehensive score calculation submodule: The indicators calculated by the multi-indicator evaluation submodule Comprehensive scoring, combined with carrier profile , according to the needs of the enterprise, set the weight of each indicator to flexibly adjust the scoring results and comprehensively score Expressed as: ;in, To correspond to the weight of the scene, by adjusting the weight, the multi-dimensional carrier portrait The different feature weights in the are reflected in the comprehensive score.

[0016] Optionally, the carrier relationship detection module specifically includes: Relationship data acquisition submodule: obtains shareholder information, management information and historical business cooperation records related to the carrier; Graph database construction submodule: Based on the acquired association relationship data, the graph database technology is used to build the association graph between carriers. In the association graph, each node Represents an entity, including a carrier, shareholder, or management individual, and the edges represent the relationship between them. Represents two nodes and If the relationship between ,express and There is a relationship between them, and the construction of the graph database is expressed as: ,in, is a node set, is the edge set; Relationship analysis submodule: Utilizes the relationship graph in the graph database and adopts the relationship network analysis technology to detect the potential relationship between the preferred carrier and other carriers. The specific analysis includes: Shareholder relationship detection: By traversing the shareholder nodes in the graph database, determine whether the preferred carrier and other carriers have common shareholders. and If they are directly connected through a shareholder node V shareholder, then it is considered that there is a shareholder association; Management layer intersection detection: By analyzing the management layer nodes in the graph database, detect whether there is a management layer intersection between the preferred carrier and other carriers. With two carrier nodes at the same time and If it is directly connected, it means that there is overlap in management layers; Business cooperation association detection: By analyzing the historical business cooperation records between carriers, determine whether there is a cooperation relationship. If there is a cooperation history, mark the relationship in the graph database; Implicit association identification submodule: Using the association network in the graph database, through graph traversal and shortest path search algorithm, identify the implicit association between the preferred carrier and other carriers. Implicit association refers to the relationship that is indirectly associated through multi-level nodes, such as shareholder A is associated with carrier B, and management B is associated with carrier C. In the graph database, the shortest path algorithm is used to search and If the path length is less than the set threshold L, it is marked as an implicit association.

[0017] The carrier portrait evaluation method based on multi-dimensional data fusion is implemented by the above-mentioned carrier portrait evaluation system based on multi-dimensional data fusion, and includes the following steps: S1: Real-time collection of multi-source data related to multiple carriers, including carriers’ qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data, and pre-processing of the collected data; S2: Fuse the pre-processed multi-source data, adopt a method combining feature-level fusion and data-level fusion, and build a dynamic carrier profile through multimodal fusion algorithm and hierarchical data modeling technology to provide a basis for evaluation; S3: Based on multi-dimensional carrier portraits, comprehensive evaluation is conducted on carriers in multiple business scenarios, including multimodal transport, temporary route planning, and special cargo transportation. Based on simulation technology and historical data analysis, a multi-index evaluation method is used to calculate the coordination and integration capability index, real-time decision-making capability index, and safety management capability index, and a comprehensive score is given based on the evaluation results of each index. S4: Analyze the relationships between carriers, use graph database-based correlation analysis technology to detect potential correlations in shareholder relationships, management overlaps, and business cooperation, and identify implicit relationships between carriers through correlation network analysis technology to ensure that there are no potential conflicts of interest.

[0018] Beneficial effects of the present invention: The present invention, based on multi-dimensional data fusion technology, deeply integrates multi-source data such as the qualifications, performance, scale, capacity, historical cooperation years, etc. of the carrier through a method combining feature-level fusion and data-level fusion, and constructs a dynamic carrier portrait. Through layered data modeling technology, the information of static attributes (qualification, scale) and dynamic behaviors (capacity, performance) are organically combined to generate a hierarchical and structured carrier portrait, which can comprehensively and real-time reflect the overall performance of the carrier, provide an accurate data basis for subsequent scenario-based performance evaluation, and effectively improve the accuracy of the evaluation and the reliability of decision support.

[0019] The present invention introduces an agent-based simulation algorithm, which generates key evaluation parameters (such as transportation time, cost, and safety) by simulating the actual performance of carriers in specific business scenarios (such as multimodal transport, temporary route planning, and special cargo transportation). Through a multi-indicator evaluation method, combined with the multi-dimensional data output by the simulation, it is possible to quantitatively evaluate the performance of carriers in different business scenarios, and flexibly adjust the weight of each indicator according to the needs of the enterprise, thereby generating a comprehensive integrated score. This method not only improves the accuracy of the evaluation, but also enhances the system's adaptability to complex business environments, allowing enterprises to make more scientific decisions when selecting carriers.

[0020] The present invention realizes the automatic detection of potential conflicts of interest between the preferred carrier and other carriers through the correlation analysis technology based on the graph database, and constructs and analyzes the shareholder relationship, management overlap, business cooperation and implicit association relationship between the carriers by using the graph traversal and path search algorithm, so as to ensure that there is no conflict of interest in the carrier finally selected. This process can effectively avoid the potential risks caused by association relationships in the selection of carriers, improve the transparency and security of cooperation, and safeguard the long-term cooperation of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A schematic diagram of functional modules of an evaluation system according to an embodiment of the present invention; Figure 2 Schematic diagram of the evaluation method flow in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0024] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0025] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0026] like Figure 1 As shown in the figure, the carrier profile evaluation system based on multi-dimensional data fusion includes: Data source acquisition module: used to collect multiple carrier-related data sources in real time, including carrier qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data; Data preprocessing module: used to normalize the collected data; data normalization submodule: normalize the data from different sources and scale the data to the same scale range for subsequent multidimensional data fusion and evaluation analysis. The normalization process uses the MinMax normalization method, which is expressed as: ,in, Represents the original data, represents the normalized data, and They represent the minimum and maximum values ​​in the data set respectively. After normalization, all data will be scaled to the interval [0,1] to ensure that data from different data sources can be fused and analyzed at the same scale.

[0027] Multi-dimensional data fusion module: It fuses the pre-processed multi-source data, uses multi-modal fusion algorithm and hierarchical data modeling technology to build a dynamic carrier portrait, providing a basis for evaluation; Scenario-based performance evaluation module: Based on multi-dimensional carrier portraits, comprehensive evaluation of carriers in multiple business scenarios is conducted. Business scenarios include multimodal transport, temporary route planning, and special cargo transportation. Based on simulation technology and historical data analysis, a multi-index evaluation method is used to evaluate the performance of carriers in different business scenarios. The multiple indicators include coordination and integration capability indicators (the carrier's coordination and integration capabilities between different modes of transport), real-time decision-making capability indicators (the carrier's ability to quickly formulate and adjust transportation routes in emergencies), and safety management capability indicators (the carrier's safety management capabilities when transporting special cargo). Multimodal transport refers to the scenario where goods need to be transported in combination by multiple modes of transport (such as sea, rail, and road); temporary route planning refers to the ability to require carriers to re-plan transportation routes based on real-time road conditions or temporary instructions in temporary or emergency events, which places high demands on the carrier's flexibility and real-time decision-making capabilities; comprehensive scoring is conducted based on the evaluation results of multiple indicators; comprehensive scoring of multiple carriers is conducted through the multi-index evaluation method to obtain a comprehensive score ranking of multiple carriers; Carrier relationship detection module: select the carrier with the highest comprehensive score as the preferred carrier, analyze the relationship between the preferred carrier and other carriers, ensure that there is no potential conflict of interest between the preferred carrier and other carriers, use correlation analysis technology based on graph database to detect potential correlations between shareholders, management and business cooperation, and identify implicit correlations through correlation network analysis technology to help companies avoid cooperation risks.

[0028] The data source acquisition module includes: Multi-channel data interface submodule: establishes real-time connections with multiple external data sources, including industry regulatory databases, industry association platforms, enterprise management systems, third-party market analysis tools and logistics platforms. Through data interfaces, it obtains and updates the carrier's qualification data, performance data, scale data, capacity data, historical cooperation years data and cooperation amount data in real time; Data push and synchronization submodule: Through the subscription and push mechanism, it automatically receives updated information from external data sources and synchronizes the data to the evaluation system to ensure the real-time and accuracy of each data collection.

[0029] The multidimensional data fusion module specifically includes: Data fusion submodule: Integrate the pre-processed multi-source heterogeneous data, adopt a method combining feature-level fusion and data-level fusion, extract key features from different data sources through feature-level fusion, and uniformly map the key features to a shared feature space; then use the data-level fusion method to comprehensively integrate the data in the unified feature space to generate a high-dimensional, unified-format carrier data set that fully retains the information characteristics and inherent relationships of each data source; Hierarchical data modeling submodule: Receives the carrier data set from the multimodal data fusion submodule and uses hierarchical data modeling technology to divide the data into different levels according to data attributes and business logic, including: Basic layer: covers static attribute data of carriers, including qualifications and scale; Behavior layer: covers the carrier’s dynamic operational data, including performance, capacity, historical cooperation years, and cooperation amount; Comprehensive layer: establishes an association between the basic layer and the behavior layer. The static data of the basic layer has a direct impact on the dynamic data of the behavior layer, reflecting the association between the basic layer and the behavior layer; At each level, a hierarchical and well-structured carrier data model is constructed so that each layer of data can independently describe the information of a specific dimension and reflect the inherent connection of the data through inter-layer correlation, thus forming a comprehensive and dynamic carrier portrait and providing accurate and rich data support for subsequent scenario-based performance evaluation.

[0030] Feature-level fusion is used to extract key features from different data sources and unify the key features into the same dimension, which is a shared feature space. The data set of each data source is represented as ,in, Index for data source, each data source Include dimensional features, and the most important features (key features) are extracted through the PCA feature selection algorithm, which is expressed as: ,in, For data source The extracted features are then reduced to form a unified feature representation, and then the extracted key features are merged: , which generates a fused feature matrix through feature concatenation or concatenation operations ; PCA is a dimensionality reduction technique that maps high-dimensional data to low-dimensional space through linear transformation while retaining as much data variance as possible. The specific steps include: Compute the covariance matrix of the data; Perform eigenvalue decomposition on the covariance matrix to obtain eigenvectors and eigenvalues; Sort by eigenvalue and select the first few main eigenvectors as new basis vectors; Use these basis vectors to project the original data into a low-dimensional space to obtain the reduced-dimensional data.

[0031] Data-level fusion is based on feature-level fusion. It integrates the feature matrices of different data sources to form a unified data representation. It is assumed that each data source is normalized and mapped to the same numerical range to ensure the uniformity of the data. Data-level fusion uses a weighted average method to map each The feature matrix of is fused according to its weight, expressed as: ,in, For data source The weight of ,By adjusting the weights, we control the impact of different data sources on the fusion results, and ensure that the most important data sources occupy a larger proportion in the overall fusion.

[0032] The hierarchical data modeling submodule receives the fused carrier data set from the multimodal data fusion submodule, uses hierarchical data modeling technology to organize different types of data into layers, and establishes associations between layers to form a unified carrier portrait, including: Base layer modeling: The base layer data includes static attributes of carriers. The data of this layer is modeled as a feature matrix: ,in, Represents a static property, Indicates Item static properties; Behavior layer modeling: Behavior layer data includes dynamic attributes of carriers. Dynamic attributes have time series characteristics and are captured through dynamic models. Behavior layer data is formed into behavioral data vectors through time series analysis. ,in, Represents a moment in a time series: , using an autoregressive model to capture the time evolution of the data: , represents the behavior trend of the carrier in the historical cooperation process, among which, It's time behavioral characteristics, is the regression coefficient, is the error term, is the lag order; The comprehensive layer is used to establish associations between the basic layer and the behavioral layer. The static data of the basic layer has a direct impact on the dynamic data of the behavioral layer. For example, the qualifications of the carrier May affect its capacity , this influence can be expressed by the weight coefficient To represent the association between the base layer and the behavior layer: ,in, The comprehensive layer represents the dynamic part of the behavior layer. The comprehensive layer integrates the data of the basic layer and the behavior layer to generate a complete carrier profile. The model of the comprehensive layer is: ,in, It is a nonlinear function that combines the features of the base layer and the behavioral layer to generate a comprehensive carrier profile , is a nonlinear mapping function used to combine the base layer and behavior layer data. The form is: ,in is the ReLU activation function, is the weight matrix, is bias.

[0033] It is a description of specific dynamic behavior, combined with the influence of the base layer. P is an overall portrait after integrating all base layer and behavior layer data, containing more extensive multi-dimensional information.

[0034] The scenario-based performance evaluation module includes: Select a business scenario, including , , , the simulation environment is constructed based on the selected business scenarios. In each selected scenario, the simulation environment is constructed based on historical data and the current environment (such as traffic flow, weather conditions, cargo characteristics, etc.). The simulation environment is constructed as follows: Multidimensional data integration: Profiling carriers The characteristic data (scale, capacity, and historical cooperation years) in the scenario are combined with the specific parameters in the scenario (route complexity, cargo requirements) to form the input parameter set of the scenario. ; Simulation calculation: Based on the input parameter set , using agent-based simulation to simulate the performance of carriers in each business scenario. The output generated by the simulation calculation includes transportation time ,cost , Security Parameters, simulation output is expressed as: ,in, It's a scene The simulation output results include transportation time, cost, and safety parameters. The accident rate during transportation and cargo integrity rate To measure.

[0035] Agent-based simulation uses the behavior and interactions of individuals (agents) to simulate the overall dynamics of the system. Each agent represents an entity in the system (carrier, vehicle, cargo). By defining the behavior rules, interaction mechanisms, and decision-making processes of these agents, the simulation system simulates complex business scenarios and generates key output parameters, as follows: 1. Agent definition: In the simulation, we first need to define each agent and their attributes and behavior rules. For the carrier profile evaluation system, the agents include: Carrier agent: represents a specific carrier, including attributes such as capacity, qualifications, performance, and historical cooperation records.

[0036] Vehicle Agent: A vehicle used by a carrier, containing attributes such as vehicle type, capacity, speed, and maintenance status.

[0037] Freight forwarder: represents the goods being transported, including attributes such as weight, volume, transportation requirements (such as temperature control), and the value of the goods.

[0038] Environmental agents: represent the external environment during transportation, such as road networks, weather conditions, and traffic conditions.

[0039] 2. Behavioral rule definition: Each agent sets behavioral rules based on its attributes and business scenario requirements. The behavioral rules determine how the agent makes decisions and interacts with other agents.

[0040] Rules of conduct for carrier agents: selecting optimal routes, dispatching vehicles, and responding to emergencies (traffic jams, route changes).

[0041] Rules of conduct for vehicle agents: perform transportation tasks according to the carrier's dispatch, follow traffic regulations, and handle equipment failures.

[0042] Rules of conduct for freight forwarders: specify transportation requirements, including transportation temperature, arrival time, and track their own status (such as whether it is damaged during transportation).

[0043] Behavioral rules of environmental agents: influence the behavior of other agents, including reduced vehicle speeds due to deteriorating traffic conditions and increased transportation risks due to bad weather.

[0044] 3. Simulation process: By running the simulation model in selected business scenarios (such as multimodal transport, ad hoc route planning, and special cargo transportation), agents interact with other agents according to their behavioral rules and perform tasks.

[0045] Here are the steps: 3.1 Initialize the simulation environment: According to the scenario selection and simulation submodule settings, load the feature data in the multi-dimensional carrier portrait P and initialize the attributes and environment parameters of each agent.

[0046] 3.2 Agent Interaction: Agents interact in the simulation environment. For example, the carrier agent decides whether to change the transportation route based on the current traffic conditions (provided by the environment agent); the vehicle agent performs specific transportation tasks and updates the status of the cargo agent in real time.

[0047] 3.3 Simulation Iteration: As the simulation time step advances, each agent continuously interacts according to the behavioral rules to complete the transportation task. The simulation system records the key indicators of each time step, such as transportation time, transportation cost and safety.

[0048] 3.4 End of simulation and result output: After the simulation is completed, the system summarizes the performance of each agent in the entire simulation process, generates and outputs key parameters.

[0049] Calculation of key output parameters: Transportation time ( ): Calculate the start time of each transport task and end time , the transportation time for each task is: ; The total transportation time is the sum of all transportation task times: ; Shipping costs :Transportation costs include fuel costs, vehicle maintenance costs, and labor costs. Let the unit cost of each transportation task be : ; The total transport cost is the sum of all task costs: ; Security :Safety is measured by the accident rate and cargo integrity rate during transportation. and accident rate Respectively expressed as: ; ; The comprehensive safety score is calculated using the following formula: , is the weight coefficient of cargo integrity rate, which is 0.6. is the accident rate weight coefficient, which is taken as 0.4.

[0050] Simulation Output The transportation time, cost and safety parameters in the project are quantitatively analyzed through a multi-index evaluation method, including: Coordination and integration capability index: measures the carrier's coordination and integration capabilities between different modes of transport in multimodal transport scenarios, and is quantified by calculating transport time, transshipment efficiency, and transport cost parameters; Real-time decision-making capability indicator: In the temporary route planning scenario, the carrier's decision-making speed and flexibility in responding to emergencies are evaluated, based on the analysis of historical emergency response time and route adjustment success rate data; Safety management capability index: In special cargo transportation scenarios, the carrier's ability to manage cargo safety during transportation is evaluated and quantified by analyzing cargo integrity rate and safety accident rate; Comprehensive score calculation submodule: Based on the multi-indicator evaluation results, the weighted average method is used to comprehensively score the indicators in each scenario.

[0051] The coordination and integration capability index is calculated as follows (multimodal transport scenario): ; in, For multimodal transport , For multimodal transport , Calculated as the ratio of each transshipment operation time to the total transportation time: ,in, It is The time of the refitting operation, is the number of transshipments. The transshipment efficiency depends on the operation time of each transport node in the simulation, including the start and end time of transshipment. The relevant data are recorded in the multimodal transport simulation data. for The weight coefficient measures the importance of transportation time in the overall coordination ability and is set to 0.5. for The weight coefficient of , which measures the importance of changing efficiency in coordination ability, is 0.3. for The weight coefficient, which measures the importance of transportation cost in the evaluation, is 0.2; The real-time decision-making capability index is calculated as (temporary route planning scenario): ; in, for The weight coefficient, which measures the importance of response time, is 0.7. for The weight coefficient measures the importance of the successful adjustment rate and is set to 0.3. The response time is the time difference for emergency response, calculated as: ,in, Is the carrier responsible for the event The response time, It is an event Time of occurrence; is the route adjustment success rate, calculated as: ; The safety management capability index is calculated as follows (special cargo transportation scenario): ; Among them, the simulation results Depend on The simulation results show that is the weight coefficient of cargo integrity rate, which is 0.6. is the accident rate weight coefficient, which is 0.4; Comprehensive score calculation submodule: The indicators calculated by the multi-indicator evaluation submodule Comprehensive scoring, combined with carrier profile , according to the needs of the enterprise, set the weight of each indicator to flexibly adjust the scoring results and comprehensively score Expressed as: ;in, To correspond to the weight of the scene, by adjusting the weight, the multi-dimensional carrier portrait The different feature weights in the are reflected in the comprehensive score.

[0052] , , They are all corresponding weights and will be adjusted according to actual conditions.

[0053] Combined with multi-dimensional carrier portraits: In the entire scenario-based performance evaluation process, each step of the scenario selection and simulation submodule and the multi-indicator evaluation submodule is combined with the multi-dimensional carrier portraits. The base layer and behavior layer data contained in the portrait are closely related through the input parameter set are integrated into the simulation model to generate evaluation results that are closely related to actual business needs.

[0054] The carrier relationship detection module specifically includes: Relationship data acquisition submodule: obtains shareholder information, management information and historical business cooperation records related to the carrier; Graph database construction submodule: Based on the acquired association relationship data, the graph database technology is used to build the association graph between carriers. In the association graph, each node Represents an entity, including a carrier, shareholder, or management individual, and the edges represent the relationship between them. Represents two nodes and If the relationship between ,express and There is a relationship between them, and the construction of the graph database is expressed as: ,in, is a node set, is the edge set; Relationship analysis submodule: Utilizes the relationship graph in the graph database and adopts the relationship network analysis technology to detect the potential relationship between the preferred carrier and other carriers. The specific analysis includes: Shareholder relationship detection: By traversing the shareholder nodes in the graph database, determine whether the preferred carrier and other carriers have common shareholders. and If they are directly connected through a shareholder node V shareholder, then it is considered that there is a shareholder association; ; Management layer intersection detection: By analyzing the management layer nodes in the graph database, detect whether there is a management layer intersection between the preferred carrier and other carriers. With two carrier nodes at the same time and If it is directly connected, it means that there is overlap in management layers; ; Business cooperation association detection: By analyzing the historical business cooperation records between carriers, determine whether there is a cooperation relationship. If there is a cooperation history, mark the relationship in the graph database; ; Implicit association identification submodule: Using the association network in the graph database, through graph traversal and shortest path search algorithm, the implicit association between the preferred carrier and other carriers is identified. Implicit association refers to the relationship that is indirectly associated through multi-level nodes, such as shareholder A is associated with carrier B, and management B is associated with carrier C. In the graph database, the shortest path algorithm (Dijkstra algorithm) is used to search and If the path length is less than the set threshold L, it is marked as an implicit association.

[0055] The shortest path algorithm is as follows: Initialization: Set the distance of the preferred carrier node V to 0, and the distances of other nodes to infinity; Traversal update: Starting from the node with the smallest distance, update the distance of its adjacent nodes until all nodes are visited; Output the shortest path: The final path length is the shortest path value.

[0056] Based on the analysis results of shareholder relationships, management overlap, and implicit connections, a comprehensive judgment is made as to whether there is a conflict of interest between the preferred carrier and other carriers.

[0057] If the analysis shows that the preferred carrier has shareholder relationships, overlapping management, business cooperation relationships, or implicit relationships with other carriers, the search will be conducted in descending order according to the aforementioned score ranking. The shareholder relationships, overlapping management, business cooperation relationships, and implicit relationships among the above-mentioned other carriers will be considered during the ranking search. After excluding shareholder relationships, overlapping management, business cooperation relationships, and implicit relationships, the carrier with the highest score can be selected as the final carrier.

[0058] As another implementation mode of the present invention, when constructing a carrier portrait, the carrier relationship detection may also be performed. If it is detected that a carrier has a shareholder relationship, overlapping management, business cooperation relationship, or implicit relationship with other carriers, all mutually related carriers will be directly excluded, and no further scenario-based performance evaluation will be performed.

[0059] like Figure 2 As shown, the present invention also proposes a carrier portrait evaluation method based on multi-dimensional data fusion, which is implemented by the above-mentioned evaluation system and includes the following steps: S1: Real-time collection of multi-source data related to multiple carriers, including carriers’ qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data, and pre-processing of the collected data; S2: Fuse the pre-processed multi-source data, adopt a method combining feature-level fusion and data-level fusion, and build a dynamic carrier profile through multimodal fusion algorithm and hierarchical data modeling technology to provide a basis for evaluation; S3: Based on multi-dimensional carrier portraits, comprehensive evaluation is conducted on carriers in multiple business scenarios, including multimodal transport, temporary route planning, and special cargo transportation. Based on simulation technology and historical data analysis, the coordination and integration capability index, real-time decision-making capability index, and safety management capability index are calculated using a multi-index evaluation method, and a comprehensive score is given based on the evaluation results of each index. S4: Analyze the relationships between carriers, use graph database-based correlation analysis technology to detect potential correlations in shareholder relationships, management overlaps, and business cooperation, and identify implicit relationships between carriers through correlation network analysis technology to ensure that there are no potential conflicts of interest.

[0060] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0061] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. The carrier profile evaluation system based on multi-dimensional data fusion is characterized by: include: Data source acquisition module: used to collect multiple carrier-related data sources in real time, including carrier qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data; Data preprocessing module: used to perform normalization preprocessing on the collected data; Multi-dimensional data fusion module: It fuses the pre-processed multi-source data, uses multi-modal fusion algorithm and hierarchical data modeling technology to build a dynamic carrier portrait, providing a basis for evaluation; Scenario-based performance evaluation module: Based on multi-dimensional carrier portraits, comprehensive evaluation of carriers in multiple business scenarios, including multimodal transport, temporary route planning, and special cargo transportation, and using simulation technology and historical data analysis, a multi-index evaluation method is used to evaluate the performance of carriers in different business scenarios. The multi-index includes coordination and integration capability indicators, real-time decision-making capability indicators, and safety management capability indicators. Comprehensive scoring is performed based on the evaluation results of multiple indicators; Comprehensively score multiple carriers through a multi-index evaluation method to obtain a comprehensive score ranking of multiple carriers; Carrier relationship detection module: select the carrier with the highest comprehensive score as the preferred carrier, analyze the relationship between the preferred carrier and other carriers, ensure that there is no potential conflict of interest between the preferred carrier and other carriers, use correlation analysis technology based on graph database to detect potential correlations among shareholders, management and business cooperation, and identify implicit correlations through correlation network analysis technology.

2. The carrier portrait evaluation system based on multi-dimensional data fusion according to claim 1 is characterized in that: The data source acquisition module includes: Multi-channel data interface submodule: establishes real-time connections with multiple external data sources, including industry regulatory databases, industry association platforms, enterprise management systems, third-party market analysis tools and logistics platforms, and obtains and updates the carrier's qualification data, performance data, scale data, capacity data, historical cooperation years data and cooperation amount data in real time through the data interface.

3. The carrier portrait evaluation system based on multi-dimensional data fusion according to claim 1, characterized in that: The multidimensional data fusion module specifically includes: Data fusion submodule: Integrate the pre-processed multi-source heterogeneous data, adopt a method combining feature-level fusion and data-level fusion, extract key features from different data sources through feature-level fusion, and uniformly map the key features to a shared feature space; then use the data-level fusion method to comprehensively integrate the data in the unified feature space to generate a carrier data set; Hierarchical data modeling submodule: Receives the carrier data set from the multimodal data fusion submodule and uses hierarchical data modeling technology to divide the data into different levels according to data attributes and business logic, including: Basic layer: covers static attribute data of carriers, including qualifications and scale; Behavior layer: covers the carrier’s dynamic operational data, including performance, capacity, historical cooperation years, and cooperation amount; Comprehensive layer: establishes an association between the basic layer and the behavior layer. The static data of the basic layer has a direct impact on the dynamic data of the behavior layer, reflecting the association between the basic layer and the behavior layer; At each level, a hierarchical and clearly structured carrier data model is constructed so that each layer of data can independently describe the information of a specific dimension and reflect the internal connection of the data through inter-layer associations, thus forming a carrier portrait.

4. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 3 is characterized in that: The feature-level fusion is used to extract key features from different data sources and unify the key features into the same dimension, which is a shared feature space. The data set of each data source is represented as ,in, Index for data source, each data source Include dimensional features, and the most important features are extracted through the PCA feature selection algorithm, which is expressed as: ,in, For data source The extracted features are then reduced to form a unified feature representation, and the extracted key features are merged: ; The data-level fusion is based on the feature-level fusion, and forms a unified data representation by integrating the feature matrices of different data sources. The data-level fusion uses a weighted average method to combine each The feature matrix of is fused according to its weight, expressed as: ,in, For data source The weight of ,By adjusting the weights, the influence of different data sources on the fusion results can be controlled.

5. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 4 is characterized in that: The hierarchical data modeling submodule receives the fused carrier data set from the multimodal data fusion submodule, uses the hierarchical data modeling technology to organize different types of data in layers, and establishes associations between layers to form a unified carrier portrait, specifically including: Base layer modeling: The base layer data includes static attributes of carriers. The data of this layer is modeled as a feature matrix: ,in, Represents a static property, Indicates Item static properties; Behavior layer modeling: Behavior layer data includes dynamic attributes of carriers. Dynamic attributes have time series characteristics and are captured through dynamic models. Behavior layer data is formed into behavioral data vectors through time series analysis. ,in, Represents a moment in a time series: , using an autoregressive model to capture the time evolution of the data: ,in, It's time behavioral characteristics, is the regression coefficient, is the error term, is the lag order; The comprehensive layer is used to establish an association between the basic layer and the behavior layer through the weight coefficient To represent the association between the base layer and the behavior layer: ,in, The comprehensive layer represents the dynamic part of the behavior layer. The comprehensive layer integrates the data of the basic layer and the behavior layer to generate a complete carrier profile. The model of the comprehensive layer is: ,in, It is a nonlinear function that combines the features of the base layer and the behavioral layer to generate a comprehensive carrier profile .

6. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 5, characterized in that: The scenario-based performance evaluation module includes: Select a business scenario, including , , , the simulation environment is constructed based on the selected business scenario. In each selected scenario, the simulation environment is constructed based on historical data and the current environment. The simulation environment is constructed as follows: Multidimensional data integration: Profiling carriers The feature data in the scene is combined with the specific parameters in the scene to form the input parameter set of the scene. ; Simulation calculation: Based on the input parameter set , using agent-based simulation to simulate the performance of carriers in each business scenario. The output generated by the simulation calculation includes transportation time ,cost , Security Parameters, simulation output is expressed as: ,in, It's a scene The simulation output results are: The accident rate during transportation and cargo integrity rate To measure.

7. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 6, characterized in that: The simulation output The transportation time, cost and safety parameters in the project are quantitatively analyzed through a multi-index evaluation method, including: Coordination and integration capability index: measures the carrier's coordination and integration capabilities between different modes of transport in multimodal transport scenarios, and is quantified by calculating transport time, transshipment efficiency, and transport cost parameters; Real-time decision-making capability indicator: In the temporary route planning scenario, the carrier's decision-making speed and flexibility in responding to emergencies are evaluated, based on the analysis of historical emergency response time and route adjustment success rate data; Safety management capability index: In special cargo transportation scenarios, the carrier's ability to manage cargo safety during transportation is evaluated and quantified by analyzing cargo integrity rate and safety accident rate; Comprehensive score calculation submodule: Based on the multi-indicator evaluation results, the weighted average method is used to comprehensively score the indicators in each scenario.

8. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 7, characterized in that: The coordination and integration capability index is calculated as: ； in, For multimodal transport , For multimodal transport , Calculated as the ratio of each transshipment operation time to the total transportation time: ,in, It is The time of the refitting operation, For the number of changes, for The weight coefficient measures the importance of transportation time in the overall coordination ability. for The weight coefficient measures the importance of dressing efficiency in coordination ability. for The weight coefficient measures the importance of transportation costs in the evaluation; The real-time decision-making capability index is calculated as: ； in, for The weight coefficient measures the importance of response time. for The weight coefficient measures the importance of the successful adjustment rate and the response time is the time difference for emergency response, Adjust the success rate for the route; The safety management capability index is calculated as: , is the cargo integrity rate weight coefficient, is the accident rate weight coefficient; Among them, the simulation results Depend on The simulation results show that; Comprehensive score calculation submodule: The indicators calculated by the multi-indicator evaluation submodule Comprehensive scoring, combined with carrier profile , according to the needs of the enterprise, set the weight of each indicator to flexibly adjust the scoring results and comprehensively score Expressed as: ;in, To correspond to the weight of the scene, by adjusting the weight, the multi-dimensional carrier portrait The different feature weights in are reflected in the comprehensive score.

9. The carrier profile evaluation system based on multi-dimensional data fusion according to claim 1, characterized in that: The carrier relationship detection module specifically includes: Relationship data acquisition submodule: obtains shareholder information, management information and historical business cooperation records related to the carrier; Graph database construction submodule: Based on the acquired association relationship data, the graph database technology is used to build the association graph between carriers. In the association graph, each node Represents an entity, including a carrier, shareholder, or management individual, and the edges represent the relationship between them. Represents two nodes and If the relationship between ,express and There is a relationship between them, and the graph database is represented as: ,in, is a node set, is the edge set; Association analysis submodule: Utilizes the association graph in the graph database and uses association network analysis technology to detect various associations between the preferred carrier and other carriers. The specific analysis includes: Shareholder relationship detection: By traversing the shareholder nodes in the graph database, determine whether the preferred carrier and other carriers have common shareholders. and If they are directly connected through a shareholder node V shareholder, then it is considered that there is a shareholder association; Management layer intersection detection: By analyzing the management layer nodes in the graph database, detect whether there is a management layer intersection between the preferred carrier and other carriers. With two carrier nodes at the same time and If it is directly connected, it means that there is overlap in management layers; Business cooperation association detection: By analyzing the historical business cooperation records between carriers, determine whether there is a cooperation relationship. If there is a cooperation history, mark the relationship in the graph database; Implicit association identification submodule: Using the association network in the graph database, through graph traversal and shortest path search algorithm, identify the implicit association between the preferred carrier and other carriers. Implicit association refers to the relationship that is indirectly associated through multi-level nodes. In the graph database, the shortest path algorithm is used to search and If the path length is less than the set threshold L, it is marked as an implicit association.

10. A carrier portrait evaluation method based on multidimensional data fusion, implemented by a carrier portrait evaluation system based on multidimensional data fusion according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Real-time collection of multi-source data related to multiple carriers, including carriers’ qualification data, performance data, scale data, capacity data, historical cooperation years data, and cooperation amount data, and pre-processing of the collected data; S2: Fuse the pre-processed multi-source data, adopt a method combining feature-level fusion and data-level fusion, and build a dynamic carrier profile through multimodal fusion algorithm and hierarchical data modeling technology to provide a basis for evaluation; S3: Based on multi-dimensional carrier portraits, comprehensive evaluation is conducted on carriers in multiple business scenarios, including multimodal transport, temporary route planning, and special cargo transportation. Based on simulation technology and historical data analysis, a multi-index evaluation method is used to calculate the coordination and integration capability index, real-time decision-making capability index, and safety management capability index, and a comprehensive score is given based on the evaluation results of each index. S4: Analyze the relationships between carriers, use graph database-based correlation analysis technology to detect potential correlations in shareholder relationships, management overlaps, and business cooperation, and identify implicit relationships between carriers through correlation network analysis technology to ensure that there are no potential conflicts of interest.

Images