Intelligent big data logistics operation system

Abstract

The invention discloses an intelligent big data logistics operation system, relates to the technical field of data processing, and solves the problem that it is difficult to store multi-source logistics data by using a block chain. A vehicle resource availability rate, a human resource availability rate and a storage resource availability rate are difficult to analyze, so that the transport capacity resource state is evaluated in real time; a goods transportation demand peak value is difficult to analyze by using a deep learning model; a shared transport capacity platform is difficult to establish to manage transport capacity resources, and a logistics operation scheme is established by using an optimization algorithm; the cost effectiveness of the logistics operation scheme is difficult to analyze; and various results are difficult to display visually. According to the invention, logistics data is stored through the block chain, and vehicle, manpower and storage resource states are analyzed so as to evaluate the transport capacity resource state; through predicting cargo transportation demands, combining an optimization algorithm and a shared transport capacity platform, transport capacity resources are efficiently configured, a logistics operation scheme is dynamically optimized, cost benefits are analyzed, and finally a result is visually displayed by using a web front end.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of data processing, and specifically relates to an intelligent big data logistics operation system. BACKGROUND

[0002] With the development of emerging technologies such as big data, artificial intelligence and blockchain, the logistics industry has an increasingly urgent demand for intelligent, data-driven and secure and reliable operation modes. By integrating deep learning algorithms to achieve demand prediction and resource optimization scheduling, using blockchain technology to ensure the authenticity and non-tamperability of data, and with the help of visualization technology to provide real-time monitoring and decision support for logistics operation management, the overall efficiency and quality of logistics operation are improved, and the digital transformation and intelligent development of the logistics industry are promoted.

[0003] The existing intelligent big data logistics operation system has the following problems: it is difficult to use blockchain to store multiple source logistics data; it is difficult to analyze the vehicle resource availability, human resource availability and warehouse resource availability to real-time evaluate the state of the transport resource; it is difficult to use a deep learning model to analyze the peak of goods transportation demand; it is difficult to build a shared transport platform to manage transport resources and use optimization algorithms to establish a logistics operation scheme; it is difficult to analyze the cost-effectiveness of the logistics operation scheme; it is difficult to visualize various results. SUMMARY

[0004] To solve the above problems existing in the prior art, the first aspect of the present application provides an intelligent big data logistics operation system, comprising the following modules:

[0005] Data acquisition and processing module: real-time acquisition of multiple source logistics data and preprocessing respectively, construction of structured spatio-temporal data set; use of blockchain technology to store structured spatio-temporal data set;

[0006] Data analysis module: based on the structured spatio-temporal data set, analyze and calculate the vehicle resource availability, human resource availability and warehouse resource availability; based on the above analysis results, real-time evaluate the state of the transport resource; based on the demand prediction model constructed by deep spatio-temporal graph neural network, analyze the peak of goods transportation demand;

[0007] Operation management module: manage transport resources by building a shared transport platform based on blockchain; combine the state of the transport resource and the peak of goods transportation demand, and use optimization algorithms to establish a logistics operation scheme;

[0008] Optimization management module: optimize the operation management scheme by analyzing the cost-effectiveness of the logistics operation scheme;

[0009] Visual display module: dynamically display the state of the transport resource, the peak of goods transportation demand, the logistics operation scheme, and the cost-effectiveness result through the Web front-end interactive dashboard.

[0010] Preferably, the structured spatio-temporal data set is stored by using blockchain technology, including the following steps:

[0011] Integrating with the Hyperledger Fabric consortium chain platform, a data collection, transmission and storage full-process on-chain architecture is built;

[0012] Through the Apache Kafka distributed message queue, high-concurrency logistics data flow is buffered, and a data preprocessing buffer layer is established;

[0013] The buffered structured spatio-temporal data set is packaged as a blockchain transaction proposal;

[0014] The transaction proposal is sent to the endorsement node cluster, and the signature endorsement is obtained after the chain code smart contract verifies the validity of the transaction logic; the transaction with endorsement is packaged into a candidate block in a determined order;

[0015] The candidate block is broadcast to all Peer nodes, and after the read-write set conflict verification and endorsement strategy verification, it is written into the distributed ledger in sequence;

[0016] The encrypted hash value or metadata is stored in the blockchain ledger to form a traceable logistics data storage chain.

[0017] Preferably, the vehicle resource availability rate, the human resource availability rate and the warehouse resource availability rate are calculated based on the structured spatio-temporal data set, including the following steps:

[0018] The formula for calculating the vehicle resource availability rate is:

[0019]

[0020] Wherein, Q 可 is the number of available vehicles, Q 总 is the total number of vehicles, Z i,max is the vehicle load of the i-th available vehicle, Z i is the current load of the i-th available vehicle; α1, α2 and α3 are weight coefficients, and the sum is 1; T i is the available time of the i-th available vehicle, T c总 is the total available time of all vehicles;

[0021] The formula for calculating the human resource availability rate is:

[0022]

[0023] Wherein, P 总 is the total number of people, P 在岗 is the number of people on duty; G j (t) is the real-time task queue length of workstation j, G max is the maximum carrying task amount, wj is the task priority weight on the workstation j, m is the number of workstations; β1, β2 and β3 are weight coefficients, and the sum is 1; T 在岗 is the total working time of the on-duty staff, T p总 is the total working time;

[0024] The formula for calculating the warehouse resource availability rate is:

[0025]

[0026] Among them, A 总 is the total warehouse capacity, A 可 is the current available warehouse capacity, H k is the total amount of the kth kind of goods entering and leaving the warehouse in the time period, H P,k is the average occupancy value of the kth kind of goods in the warehouse capacity in the time period, T Z is the time period, n is the number of goods types, v k is the priority weight of the kth kind of goods; γ1, γ2 and γ3 are weight coefficients, and the sum is 1; S 可用 is the number of normally available devices, S 总 is the total number of warehouse facility devices.

[0027] Preferably, based on the above analysis results, the real-time evaluation of the transport resource state includes the following steps:

[0028] According to the vehicle resource availability rate, the human resource availability rate and the warehouse resource availability rate, the resource coefficient y is calculated by using the weighted average summation formula;

[0029] The formula for calculating the available transport capacity percentage is:

[0030] Among them, D is the real-time order quantity, order_max is the maximum allowable order quantity, O is the target logistics service area meteorological environment coefficient, J V is the average real-time driving speed to the target logistics service area, J is the average historical driving speed to the target logistics service area;

[0031] According to the available transport capacity percentage, the available transport capacity threshold is set, and the transport resource state is divided into:

[0032] When Y≥85%, the transport resource state is normal;

[0033] When 50%≤Y<85%, the transport resource state is in a warning state;

[0034] When Y<50%, the transport resource state is in an emergency state.

[0035] Preferably, the demand prediction model based on the deep spatio-temporal graph neural network analyzes the peak of freight transportation demand, including the following steps:

[0036] Divide the target logistics service area into grid cells, and each grid serves as a graph node; based on the actual road network or logistics path, a weighted adjacency matrix is constructed;

[0037] Map the structured spatio-temporal data set to the corresponding grid node according to the preset time window length, forming a spatio-temporal graph sequence;

[0038] Construct a deep spatio-temporal graph neural network model containing a graph convolution layer and a time series modeling layer; collect multi-source logistics data at the current time, extract freight transportation volume and order volume features, and map them to the grid node;

[0039] Input the real-time graph structure into the trained deep spatio-temporal graph neural network model, and output the predicted value of freight transportation demand in each grid cell in the future specified time window;

[0040] For each grid cell, calculate the average value of freight transportation demand in the same time window in the past set number of days to obtain the historical same period reference value;

[0041] When the predicted value of freight transportation demand is greater than or equal to twice the product of the historical same period reference value, it is determined that the grid has a peak of freight transportation demand.

[0042] Preferably, by building a shared transport capacity platform based on blockchain, the transport capacity resources are managed, including the following steps:

[0043] Construct a dynamic transport capacity matching layer, build a decentralized transport capacity registration pool based on Hyperledger Fabric consortium chain, and real-time structured spatio-temporal data set;

[0044] Use the smart contract deployed on the blockchain to manage the transport capacity transaction process; the smart contract includes encrypted bid generation, carrier qualification file hash verification, and transport contract automatic execution logic;

[0045] The reached transport capacity transaction terms and the smart contract execution state are stored as transaction results in the blockchain distributed ledger.

[0046] Preferably, in combination with the state of transport capacity resources and the peak of freight transportation demand, an optimization algorithm is used to establish a logistics operation scheme, including the following steps:

[0047] Construct a multi-objective optimization model, wherein the objective function is to minimize the total operation cost; the decision variables include a set of vehicle scheduling paths, a dynamic allocation matrix of human resources, and a warehouse resource allocation scheme; the transport capacity resource state is the core constraint;

[0048] If the transport resource state is a normal state, a genetic algorithm is used to solve a cost-optimal scheduling scheme;

[0049] If the transport resource state is a warning state, a shared transport platform is activated, and a blockchain smart contract automatically matches third-party transport;

[0050] If the transport resource state is an emergency state, dynamic path re-planning is started; peak grid resources are preferentially allocated according to grid priority weights;

[0051] The demand peak grid is clustered according to spatial proximity, the future period road network state is predicted based on a deep spatio-temporal graph neural network, and the path weight is dynamically adjusted; the external transport qualification is verified through the blockchain shared transport platform, and an intelligent contract is signed;

[0052] An output verifiable operation scheme is output, a structured decision package including a vehicle scheduling path, a human resource scheduling table and a warehouse allocation scheme is generated, and the decision package hash value is stored in the blockchain.

[0053] Preferably, the optimization management module comprises the following steps:

[0054] According to the decision variables in the logistics operation scheme, cost factors including transportation cost, warehouse cost, human cost and shared transport cost are extracted; the total cost is obtained by statistical summation;

[0055] The cost benefit value formula is:

[0056] Wherein, y before is the resource coefficient before optimization of the logistics operation scheme, y after is the resource coefficient after optimization of the logistics operation scheme, cb is the total cost; T before and T after are the average delivery time of orders before and after optimization; ω1 and ω2 are weight coefficients;

[0057] If the cost benefit value is less than the cost benefit threshold, the current logistics operation scheme is re-optimized; otherwise, the scheme is stored in the blockchain.

[0058] Compared with the prior art, the present application has the following advantages:

[0059] The present application stores structured spatio-temporal data sets by using blockchain technology, realizes the non-tamperability and traceability of logistics data, improves the authenticity and security of data, solves the problem of easy tampering and difficult tracing of logistics data in the prior art, and provides more reliable data support for logistics operation decision-making.

[0060] The application can more accurately analyze the peak value of goods transportation demand by analyzing the vehicle resource availability, human resource availability and warehouse resource availability, and comprehensively evaluating the state of transport resources, so as to provide a basis for the reasonable allocation and scheduling of transport resources, and effectively avoid the idle or excessive tension of transport resources.

[0061] The application can realize dynamic management and optimal allocation of transport resources, improve the utilization efficiency of transport resources and reduce operation cost by building a shared transport platform based on blockchain, combining the state of transport resources and the peak value of goods transportation demand, and establishing a logistics operation scheme by using an optimization algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present 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 present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0063] Fig. 1 The system module schematic diagram of the present application is shown in the figure.

[0064] Fig. 2 The method flowchart of the present application is shown in the figure. DETAILED DESCRIPTION

[0065] The technical solutions of the present application will be described in detail below with reference to the embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0066] Please refer to Figs. 1-2 The first aspect embodiment of the present application provides an intelligent big data logistics operation system, which comprises the following modules:

[0067] Data acquisition and processing module: real-time acquisition of multi-source logistics data and separate preprocessing, construction of structured spatio-temporal data set; using blockchain technology to store structured spatio-temporal data set;

[0068] Data analysis module: based on structured spatio-temporal data set to analyze and calculate vehicle resource availability, human resource availability and warehouse resource availability; comprehensively evaluate the state of transport resources in real time based on the above analysis results; based on the demand prediction model constructed by deep spatio-temporal graph neural network, analyze the peak value of goods transportation demand;

[0069] An operation management module: by building a shared transport platform based on blockchain, managing transport resources; combining the state of transport resources and the peak of freight transportation demand, using optimization algorithms to establish a logistics operation scheme;

[0070] An optimization management module: by analyzing the cost-effectiveness of the logistics operation scheme, optimizing the operation management scheme;

[0071] A visual display module: through the Web front-end interactive dashboard, dynamically displaying the state of the transport resources, the peak of freight transportation demand, the logistics operation scheme, and the cost-effectiveness results.

[0072] Specifically, through Internet of Things sensors such as GPS positioning sensors, vehicle state sensors, warehouse inventory sensors, etc., real-time collection of vehicle location, running state, warehouse goods inventory, and other logistics data. At the same time, order information, transportation plan, and other data are obtained from the logistics information system. For example, the GPS sensor sends the vehicle's latitude and longitude coordinates, speed, and other information to the system every 30 seconds; the warehouse sensor detects the number of goods on the shelf and transmits the data to the system through the wireless communication module. The collected data is cleaned to remove noise data such as abnormal values caused by sensor failure. Unstructured data such as text form order description and image form goods appearance inspection data are converted into structured data. For example, through natural language processing (NLP) technology, key information such as goods name, destination, customer requirements, etc. is extracted from the order text and stored in the corresponding fields of the database; for image data, image recognition techniques such as deep learning algorithms can be used to identify the packaging type of the goods, whether it is damaged, etc. to extract structured features that can be used for subsequent analysis. The processed data is organized according to time and space dimensions. In the time dimension, the exact timestamp of each data point is recorded; in the space dimension, the vehicle location, warehouse facility location, etc. are geographically positioned.

[0073] For example, for vehicle transportation data, the latitude and longitude of each vehicle at each time point, whether the goods are loaded, the destination, etc. are recorded. For warehouse data, combined with the geographic location coordinates of the warehouse, the storage location and storage state of the goods in the warehouse are recorded, and a unified spatio-temporal data table is constructed.

[0074] The hash value of the structured spatio-temporal data set is stored in the blockchain. The distributed ledger feature of the blockchain ensures the data's immutability. For example, the system generates a hash value for each day's spatio-temporal data set through a hash algorithm, and then stores these hash values as transaction records in the blockchain network. When verifying the integrity of the data, the hash value of the current data can be recalculated and compared with its record in the blockchain. If they are consistent, it means that the data has not been tampered with.

[0075] Various data collection techniques such as sensor network communication technology, crawler technology for collecting public logistics information, and data preprocessing algorithms such as statistical analysis algorithm for data cleaning, machine learning algorithm for data conversion are used. With the help of database management systems such as relational database MySQL, NoSQL database MongoDB, etc. to build spatio-temporal data set, and integrate with blockchain technology, through programming interface API to write data hash value into blockchain.

[0076] Through the data analysis module, the availability of vehicle resources, human resources and warehouse resources is analyzed and calculated to evaluate the state of transport resources in real time. Based on the demand prediction model constructed by deep spatio-temporal graph neural network, the peak of freight transportation demand is analyzed.

[0077] The blockchain technology is used to build a shared transport platform to ensure the credibility and data security of the platform. The distributed ledger of blockchain can record the transaction records of transport resources such as vehicle rental, transportation task allocation, etc. so that each transaction can be traced and cannot be tampered with. The transport resources of different logistics enterprises are integrated on the platform, including vehicles, ships, air transportation, etc. A resource directory is established to record the type, capacity, location, status, etc. of each transport resource.

[0078] According to the real-time state of transport resources such as vehicle availability, warehouse space, etc. and the predicted peak of freight transportation demand, optimization algorithms are used to develop logistics operation schemes. When developing logistics operation schemes, multiple objectives are considered, such as minimizing transportation cost, shortest transportation time, highest service quality such as on-time delivery rate of goods, etc. A multi-objective optimization model is established to meet these constraints simultaneously. The transportation task is divided into multiple sub-tasks such as loading, transportation, unloading, etc. and the task is allocated according to the characteristics of transport resources. At the same time, by analyzing the cost-effectiveness of logistics operation scheme, the operation management scheme is optimized.

[0079] Finally, a web frontend interactive dashboard is used to dynamically display the state of the transportation resources, the peak of the cargo transportation demand, the logistics operation scheme, and the cost-benefit results. The mainstream frontend frameworks such as React, Vue.js, or Angular are chosen to build the web frontend interactive dashboard. These frameworks have rich component libraries and good community support, which can efficiently build responsive user interfaces. According to the types of information to be displayed, the dashboard is divided into multiple display areas. For example, one area is set up to display an overview of the state of transportation resources such as vehicle, manpower, and warehouse resource availability and comprehensive state, another area is set up to display a distribution map of the peak of cargo transportation demand, and another area is set up to display detailed information of the logistics operation scheme such as vehicle dispatching path, manpower resource scheduling table, and warehouse allocation scheme, and finally an area is set up to display charts and data tables of cost-benefit results. Suitable visualization components are selected for each display area. For example, chart libraries are used to create column charts, line charts, pie charts, etc. to display the change trend and proportion of resource availability, cost-benefit, etc. Map components are used to display the geographical distribution of the peak of cargo transportation demand. Table components are used to display detailed data of the logistics operation scheme and specific values of cost-benefit, etc.

[0080] Backend API interfaces are developed to provide the required logistics operation data to the frontend dashboard. These interfaces can be based on RESTful architecture or GraphQL technology, returning data in JSON or XML format. The latest data is periodically obtained from the backend through timed polling, or real-time data pushing is achieved through technologies such as WebSocket, allowing the frontend dashboard to update the display content in a timely manner.

[0081] The data obtained from the backend is converted into the format required by the frontend visualization components. For example, time series data is converted into a time axis format that the chart library can recognize, and geographic coordinate data is converted into a marker point format required by the map component. According to the display needs, the data is aggregated and filtered. For example, the vehicle resource availability is summarized by time period to calculate the daily, weekly, or monthly availability; according to the user's selection of a specific area or time range, the corresponding cargo transportation demand peak data is filtered and displayed.

[0082] The comprehensive state of the transportation resources is displayed using intuitive graphical elements such as dashboard charts, progress bars, etc. including normal, warning, and emergency. For example, green represents normal state, yellow represents warning state, and red represents emergency state, with specific comprehensive state evaluation result values marked next to the graphics. The availability of vehicle resources, manpower resources, and warehouse resources is displayed in the form of column charts or pie charts. In the chart, different colors or patterns distinguish different resource types, and specific availability percentage values are displayed below or next to the chart.

[0083] Mark each grid cell of the target logistics service area on the map, and use color depth or marker size to represent the peak degree of different grids according to the size of the cargo transportation demand peak. For example, the higher the demand peak of a grid, the darker the marker color or the larger the size. Combined with the time axis component, the trend of cargo transportation demand peak over time is displayed. Users can view the demand peak distribution of different time periods by dragging the time axis or selecting a specific time range, in order to analyze the peak and trough periods of demand.

[0084] Draw the scheduling path of the vehicle on the map, and distinguish different vehicle routes by different colors or line styles. At the same time, the starting point, ending point and passing cargo loading and unloading points of the vehicle can be marked, and detailed information such as the expected arrival time and transportation cargo volume of the vehicle can be displayed.

[0085] Use tables or timelines to display the scheduling of human resources. The table lists information such as employee name, job title, working time and task content; the timeline can intuitively display the work arrangement of employees at different time periods, so that users can view the dynamic allocation of human resources.

[0086] Through the visualization of warehouse layout and cargo storage location, the allocation scheme of warehouse resources is presented. The storage area and storage volume of different goods are marked on the warehouse layout map, and the in-out warehouse time arrangement and transportation task association of the goods are displayed.

[0087] Use pie charts or stacked column charts to display the constituent ratio of total cost, including transportation cost, warehouse cost, labor cost and shared transportation cost, etc. In the chart, the name and proportion of each cost type are clearly marked, and the drill-down function can be provided to let users view the detailed data of each cost type.

[0088] In this embodiment, the structured spatio-temporal data set is stored by using blockchain technology, including the following steps:

[0089] Integrate with Hyperledger Fabric consortium chain platform to build a full-process on-chain architecture for data collection, transmission and storage;

[0090] Buffer high-concurrency logistics data streams through Apache Kafka distributed message queue to establish a data preprocessing buffer layer;

[0091] Encapsulate the buffered structured spatio-temporal data set as a blockchain transaction proposal;

[0092] Send the transaction proposal to the endorsement node cluster, and obtain the signature endorsement after the chain code smart contract verifies the validity of the transaction logic; package the endorsed transactions into candidate blocks in a certain order;

[0093] The candidate block is broadcast to all Peer nodes, and after read-write set conflict verification and endorsement policy verification, it is sequentially written into the distributed ledger;

[0094] The encrypted hash value or metadata is notarized in the blockchain ledger to form a traceable logistics data notarization chain.

[0095] Specifically, a Hyperledger Fabric network is deployed in a server or cloud environment. This includes installing tools such as Docker, Docker-Compose, and binary files of Hyperledger Fabric. For example, use the docker-compose command to start network nodes, ordering services, and chaincode containers, and other components. Set up the organizational structure of the consortium chain, including multiple participants such as logistics enterprises, regulatory departments, and transportation companies as different organizational nodes. Configure encryption materials such as certificates, keys, and other encryption materials for each organizational node, and define the channels and chaincodes of the consortium chain. Develop and deploy the smart contract of the chaincode for verifying and processing logistics data transactions. The role of the chaincode is to define transaction logic, such as verifying whether the hash value of the data set meets the requirements, whether the metadata is complete, and other requirements.

[0096] Install chaincode to each Peer node and instantiate in channels. Build an architecture for data collection, transmission, and storage on-chain. Specifically, integrate logistics data collection systems such as IoT sensor networks with blockchain clients such as Hyperledger Fabric's client SDK. Data collection devices send data to blockchain clients through APIs or message queues such as Apache Kafka. Use secure communication protocols such as TLS to ensure data security between the collection end and the blockchain network. The data storage layer is in the blockchain network, where appropriate methods are chosen to store data. In Hyperledger Fabric, the hash value and metadata of the data can be stored in the blockchain ledger, while the original data can be stored off-chain and only record the link or identification of the data in the ledger. Install and configure an Apache Kafka cluster. Kafka can handle high-concurrency logistics data streams, providing buffering and order guarantees for data transmission. Configure the logistics data collection device as a Kafka producer. The producer serializes the collected logistics data and sends it to the Kafka topic. Deploy Kafka consumers between the blockchain client and the data collection end. Consumers read data from Kafka topics and perform preprocessing such as data cleaning and format conversion. Preprocessed data is temporarily stored in Kafka's buffer area, waiting for further processing. Kafka's buffer area can alleviate the processing speed difference between the data collection end and the blockchain network, avoiding data loss or backlog. Use Kafka's partitioning function to distribute data to multiple partitions, improving system load balancing and scalability. Each partition can be consumed by different Kafka consumers, allowing parallel data processing. At the same time, dynamically adjust the number of Kafka partitions and consumers according to data flow changes to adapt to high-concurrency logistics data streams.

[0097] Structurally organize preprocessed spatiotemporal data sets to meet the requirements of blockchain transaction data. For example, combine vehicle location data and timestamps into a JSON-formatted object, and combine warehouse inventory data and timestamps into another JSON object. Use cryptographic hash algorithms to calculate the hash value of the structured spatiotemporal data set. The hash value is a unique identifier of data integrity, and any modification to the original data will result in a change in the hash value.

[0098] Metadata is generated for each structured spatiotemporal dataset. Metadata includes information such as the source of the dataset, e.g., the identity of the collection device, the type of data, e.g., vehicle data or warehouse data, the time of collection, the enterprise or organization to which the data belongs, etc. Metadata helps to quickly retrieve and verify data in the blockchain ledger. The hash value of the dataset and the metadata are packaged as a blockchain transaction proposal. The transaction proposal also includes other necessary transaction information, such as the identity of the transaction initiator, the transaction timestamp, etc.

[0099] The packaged transaction proposal is sent to the endorsement node cluster through the client SDK of Hyperledger Fabric. Endorsement nodes are responsible for executing the transaction logic in the chaincode and verifying the validity of the transaction proposal. For example, the chaincode can verify whether the hash value of the dataset meets the requirements of the data structure, whether the metadata is complete and logical, etc.

[0100] After each endorsement node receives the transaction proposal, it executes the chaincode and returns the verification result. If the transaction proposal passes the verification, the endorsement node signs the transaction proposal and returns the signed endorsement. For example, the chaincode can verify whether the data collection time is within a reasonable range, whether the data source is trustworthy, etc.

[0101] In Hyperledger Fabric, a transaction needs to meet certain endorsement policies to be considered valid. For example, the endorsement policy can require at least two endorsement nodes from two organizations to sign. After the client collects enough endorsement signatures, it packages the transaction proposal and the endorsement signatures into a transaction. The packaged transaction is submitted to the ordering service. The ordering service is responsible for packaging transactions into candidate blocks in order. A candidate block is a temporary block that contains multiple transactions and is stored in the memory of the ordering service for further processing. The ordering service usually uses consensus algorithms such as Kafka consensus algorithm, Raft consensus algorithm, etc. to ensure the order consistency of transactions.

[0102] The ordering service broadcasts the generated candidate block to all Peer nodes. After each Peer node receives the candidate block, it needs to perform a series of verification operations. The Peer node checks whether the transactions in the candidate block have read-write set conflicts. The read-write set refers to the account data read and written by the transaction during execution. If there is a conflict between the read-write sets of two transactions, the later arriving transaction may be rejected. The Peer node verifies whether the transaction meets the endorsement policy. For example, it checks whether the transaction has obtained a sufficient number of endorsement node signatures and whether the signatures come from legal endorsement nodes. If the transaction passes the verification, the Peer node writes the transactions in the candidate block to the local distributed ledger. The distributed ledgers of all Peer nodes are kept consistent through consensus algorithms. Once a block is successfully added to the ledger of a Peer node, other Peer nodes will synchronize the block to ensure the consistency of the ledger state across the entire blockchain network.

[0103] The encrypted hash value or metadata in the transaction is permanently recorded in an unalterable manner in the blockchain ledger. These data are stored in blocks of the blockchain, each block is linked to the previous block through a hash value, forming an unalterable chain structure. Due to the unalterable and traceable nature of the blockchain ledger, the recordation information of the logistics data can be queried and verified at any time. Through the transaction ID or data hash value, the collection time, source, storage location, etc. of the data can be traced in the blockchain network. For example, regulatory departments can query the recordation of logistics data to verify whether the enterprise collects and stores data according to the regulations.

[0104] In this embodiment, the vehicle resource availability rate, the human resource availability rate and the warehouse resource availability rate are calculated based on the structured spatiotemporal data set, including the following steps:

[0105] The formula for calculating the vehicle resource availability rate is:

[0106]

[0107] wherein, Q 可 is the number of available vehicles, Q 总 is the total number of vehicles, Z i,max is the vehicle load of the i th available vehicle, Z i is the current load of the i th available vehicle; a1, a2 and a3 are weight coefficients, and the sum is 1; T i is the available time length of the i th available vehicle, T c总 is the total available time length of all vehicles;

[0108] The formula for calculating the human resource availability rate is:

[0109]

[0110] wherein, P 总 is the total number of people, P 在岗 is the number of people on duty; G j (t) is the real-time task queue length of station j, G max is the maximum carrying task amount, w j is the task priority weight of station j, and m is the number of stations; b1, b2 and b3 are weight coefficients, and the sum is 1; T 在岗 is the total working time length of the people on duty, T p总 is the total working time length;

[0111] The formula for calculating the warehouse resource availability rate is:

[0112]

[0113] wherein, A 总is the total warehouse capacity, A 可 is the current available warehouse capacity, H k is the total amount of goods in and out of the warehouse for the kth kind of goods in a time period, H P,k is the average occupancy value of the warehouse capacity for the kth kind of goods in a time period, T Z is the time period, n is the number of goods categories, v k is the priority weight of the kth kind of goods; γ1, γ2 and γ3 are weight coefficients, and the sum is 1; S 可用 is the number of devices in a normal available state, S 总 is the total number of devices in the warehouse facility.

[0114] Specifically, according to the multi-source logistics data in the structured spatiotemporal data set, the current state data of the vehicle is obtained in real time, including the load capacity, current load, available time and other information of the vehicle; the on-duty state of the employee, the task queue length and task priority of the workstation and other information are obtained; the warehouse capacity, the goods in and out of the warehouse data, the device state and other information are obtained. By substituting the collected and preprocessed multi-source logistics data into the calculation formula of the vehicle resource availability, the human resource availability and the warehouse resource availability, the vehicle resource availability, the human resource availability and the warehouse resource availability are calculated. Among them, when calculating the vehicle resource availability, α1, α2 and α3 are weight coefficients, which are 0.4, 0.3 and 0.3 respectively, when calculating the human resource availability, the task priority weight of the workstation j and the priority weight of the kth kind of goods can be set according to actual requirements or can be determined by using the entropy weight method, β1, β2 and β3 are weight coefficients, which are 0.35, 0.35 and 0.3 respectively, when calculating the warehouse resource availability, γ1, γ2 and γ3 are weight coefficients, which are 0.4, 0.35 and 0.25 respectively. In practical application, according to the business requirements and historical data performance of the logistics system, the weight coefficients are adjusted to optimize the accuracy of the resource availability calculation.

[0115] In this embodiment, based on the above analysis results, the real-time evaluation of the transport resource state includes the following steps:

[0116] According to the vehicle resource availability, the human resource availability and the warehouse resource availability, the resource coefficient y is calculated by using the weighted average summation formula;

[0117] The available transport capacity percentage formula is:

[0118] Among them, D is the real-time order quantity, order_max is the maximum allowable order quantity, O is the target logistics service area meteorological environment coefficient, J V is the real-time average driving speed to the target logistics service area, J is the historical average driving speed to the target logistics service area;

[0119] According to the available capacity percentage, the available capacity threshold is set, and the capacity resource state is divided, including:

[0120] When Y is greater than or equal to 85%, the capacity resource state is a normal state;

[0121] When 50% is less than Y and Y is less than 85%, the capacity resource state is a warning state;

[0122] When Y is less than 50%, the capacity resource state is an emergency state.

[0123] Specifically, the real-time order quantity is obtained from the system, the maximum allowable order quantity is the maximum processable order quantity determined according to system design or historical data, the meteorological environment of the target logistics service area includes temperature, humidity, wind speed, rainfall, etc., real-time average speed and historical average speed, etc. According to the importance setting of vehicle resource availability, human resource availability and warehouse resource availability in the logistics system, the vehicle resource importance coefficient is set to 0.5, the human resource importance coefficient is set to 0.3, and the warehouse resource importance coefficient is set to 0.2. The resource coefficient y is calculated by using the weighted average summation formula. The importance coefficient can be changed according to actual requirements. The resource coefficient obtained and the multi-source logistics data obtained are substituted into the available capacity percentage formula to calculate the available capacity percentage. The meteorological environment coefficient of the target logistics service area is calculated by using the entropy weight method to configure the weight after obtaining the meteorological environment data including temperature, humidity, wind speed, rainfall, and then performing weighted summation calculation. The latest resource availability and order data are continuously obtained from the system to ensure the real-time performance of the evaluation results. The resource coefficient and the available capacity percentage are recalculated every certain time, such as every 10 minutes. According to the actual operation situation and historical data performance, the weight is dynamically adjusted to more accurately reflect the importance and influence degree of the current resource. The calculated resource coefficient and available capacity percentage and the corresponding capacity resource state are displayed through the visualization module.

[0124] In this embodiment, the demand prediction model based on deep spatio-temporal graph neural network analyzes the peak of goods transportation demand, including the following steps:

[0125] Divide the target logistics service area into grid units, and each grid is taken as a graph node; based on the actual road network or logistics path, a weighted adjacency matrix is constructed;

[0126] Map the structured spatio-temporal data set to the corresponding grid node according to the preset time window length to form a spatio-temporal graph sequence;

[0127] A deep spatio-temporal graph neural network model containing a graph convolution layer and a time series modeling layer is constructed; multi-source logistics data at the current time is collected, the features of goods transportation volume and order quantity are extracted, and are mapped to the grid node;

[0128] inputting the real-time graph structure into the trained deep spatio-temporal graph neural network model, and outputting a predicted value of the freight transportation demand of each grid cell in a specified future time window;

[0129] For each grid cell, the average value of the freight transportation demand in the same time window in the past set number of days is calculated to obtain a historical same-period reference value.

[0130] When the predicted value of the freight transportation demand is greater than or equal to the product of the historical same-period reference value multiplied by 2, it is determined that the grid has a freight transportation demand peak.

[0131] Specifically, the target logistics service area is divided into a plurality of hexagonal or rectangular grid cells, and the size of each grid cell is determined according to actual needs, for example, 1 square kilometer. Each grid cell is taken as an independent graph node.

[0132] Based on the actual road network or logistics path, the adjacency relationship between the grid nodes is defined. If there is a direct logistics path or physical connection between two grid cells, they are considered to be adjacent. A weighted adjacency matrix is constructed, and the weight can be calculated based on distance, road capacity or other logistics-related factors. By collecting road network data, the shortest path distance between each two adjacent grids is calculated and taken as the weight value in the adjacency matrix.

[0133] The structured spatio-temporal data set is mapped to the corresponding grid node according to the preset time window length. The time window length can be determined according to actual business needs, such as 1 hour, 2 hours or half a day, etc. For each grid node, the freight transportation order quantity, freight weight and other data in each hour in the past month are collected and associated with the time stamp to form a data point in a time step in the spatio-temporal graph sequence. These data points are arranged in chronological order to form a spatio-temporal graph sequence. The sequence contains not only the freight transportation demand data of each grid node in different time windows, but also the spatial relationship between the grid nodes. Each time window corresponds to a graph structure, which contains node features such as freight transportation quantity and adjacency matrix.

[0134] A deep spatio-temporal graph neural network model is constructed, which includes a graph convolution layer for spatial information modeling and a time series modeling layer such as LSTM or Transformer for time information modeling. The graph convolution layer is used to capture the spatial dependence relationship between grid nodes. The time series modeling layer is used to capture the dynamic changes in the time series.

[0135] The model is trained using historical data to predict the freight transportation demand of each grid cell in a specified future time window. The historical spatio-temporal graph sequence is used as input, and the freight transportation demand is used as label. An appropriate loss function such as mean square error and an optimizer are selected for model training until the model converges.

[0136] Collect multi-source logistics data at the current time, including vehicle GPS data, order generation data, warehouse in-out data, etc. Extract key features such as cargo transportation volume and order volume, and map them to grid nodes to form a real-time graph structure at the current time. Input the real-time graph structure into the trained deep spatio-temporal graph neural network model, and output the cargo transportation demand prediction value of each grid unit in the future specified time window, such as the next 1 hour.

[0137] For each grid unit, calculate the average value of cargo transportation demand in the same time window in the past set number of days, such as the past 30 days, as the historical same period reference value. When the cargo transportation demand prediction value is greater than or equal to twice the historical same period reference value, it is determined that the grid unit has a cargo transportation demand peak. Aggregate the identified peak grid units to form the cargo transportation demand peak distribution information in the target logistics service area.

[0138] Through the visualization module, dynamically display the peak distribution information, and feed back the peak information to the transport resource management system to allocate resources in advance to cope with peak demand.

[0139] For example: Suppose the target logistics service area is a logistics distribution area of a city, divided into 500 hexagonal grid units. Historical data shows that the average cargo transportation demand of a grid unit in the same time window of the past 30 days, such as every morning 10:00-11:00, is 100 units.

[0140] Use the historical spatio-temporal graph sequence data of the past year to train the deep spatio-temporal graph neural network model.

[0141] Input the real-time graph structure at the current time into the model, and predict that the cargo transportation demand of the grid unit in the next 1 hour is 300 units.

[0142] The historical same period reference value is 100 units.

[0143] Judgment condition: 300 ≥ 2 × 100, meets the peak judgment condition.

[0144] Peak output: Determine that the grid unit has a demand peak, output this information to the visualization system, and notify the transport resource management system to preferentially allocate resources.

[0145] In this embodiment, a shared transport platform based on blockchain is built to manage transport resources, including the following steps:

[0146] Build a dynamic transport resource matching layer, build a decentralized transport registration pool based on Hyperledger Fabric consortium chain, and real-time structured spatio-temporal data set;

[0147] The smart contract deployed on the blockchain is used to manage the capacity transaction process. The smart contract includes encrypted offer generation, carrier qualification document hash verification, and automatic transport contract execution logic.

[0148] The achieved capacity transaction terms and smart contract execution status are stored as transaction results in the blockchain distributed ledger.

[0149] Specifically, the pre-installed software required for Hyperledger Fabric, such as Docker, DockerCompose, Go language environment, etc. Then, use the command line tool or configuration file of Hyperledger Fabric to initialize the alliance chain network, including creating the genesis block, channel configuration file, etc.

[0150] Each organization participating in the alliance chain is clearly defined, such as different logistics enterprises, carriers, etc. Each organization is assigned a corresponding certificate and key to ensure its identity authentication and permission management on the chain.

[0151] Write smart contract code to manage the registration information of the carrier's vehicle. The smart contract defines the attributes of the vehicle, such as vehicle unique identifier, location information, carrying capacity, idle period, etc., and provides corresponding interfaces for the carrier to update the vehicle status in real time.

[0152] Develop API interfaces or use message queues and other technologies for carriers to send multi-source logistics data such as vehicle location, carrying capacity, and idle period to the shared capacity platform in real time. For example, the carrier can install an Internet of Things device on the vehicle to send the vehicle's GPS location information to the platform every 5 minutes.

[0153] Implement an encrypted offer generation algorithm in the smart contract. When the shipper publishes the goods transportation demand, the carrier generates an offer based on its own cost, market conditions, and other factors, and encrypts the offer through the encryption algorithm in the smart contract to ensure the security of the offer during transmission and storage.

[0154] The carrier uploads qualification documents such as business license and transportation license to the platform, and the platform calculates the hash value of the documents and stores it in the smart contract. During the transaction process, the authenticity and integrity of the qualification documents are verified by comparing the hash values of the documents.

[0155] Define the terms and execution conditions of the transport contract. For example, when the goods arrive at the destination and are confirmed by the consignee, the smart contract automatically triggers the payment process to pay the corresponding transportation fee to the carrier.

[0156] The compiled smart contract code is deployed to the Hyperledger Fabric consortium chain network. During deployment, configuration parameters such as the channel of the smart contract, endorsement strategy, etc. need to be specified to ensure that the smart contract can be correctly executed among multiple organization nodes.

[0157] When the transaction is completed, the information of the transaction parties, the details of the transported goods, the transportation fee amount, the transportation time, and other transaction terms are recorded in the form of structured data in the blockchain transaction. These data will be broadcast to all Peer nodes in the consortium chain network and written into the blockchain ledger after endorsement and verification.

[0158] The state changes generated during the execution of the smart contract, such as the acceptance of the offer, the signing of the transportation contract, the confirmation of the delivery of goods, the completion of the transportation fee payment, etc. are recorded in the blockchain ledger. Each state change generates a corresponding event log, which facilitates real-time query and tracking of the execution of the transaction by all parties.

[0159] Develop a visual query interface or provide a query API interface for participants to conveniently query the records of the transportation transaction and the execution status of the smart contract. For example, the shipper can query the offer corresponding to the transportation demand he published, the details of the transportation contract he signed, etc.; the carrier can query the registration information of his vehicle, the status of the transportation task he undertook, etc.

[0160] The platform operator can monitor the running status of the blockchain network and the execution of the transportation transaction in real time through monitoring tools. If an abnormal transaction or smart contract execution error is found, it will be promptly investigated and handled. At the same time, the historical transaction data is audited and traced using the tamper-proof feature of the blockchain, ensuring the fairness and transparency of the transportation transaction.

[0161] In this embodiment, the logistics operation scheme is established by using an optimization algorithm in combination with the state of the transportation resources and the peak of the goods transportation demand, including the following steps:

[0162] A multi-objective optimization model is constructed, in which the objective function is to minimize the total operation cost; the decision variables include the set of vehicle scheduling paths, the dynamic allocation matrix of human resources, and the warehouse resource allocation scheme; the state of the transportation resources is the core constraint;

[0163] If the state of the transportation resources is normal, a genetic algorithm is used to solve the cost-optimal scheduling scheme;

[0164] If the state of the transportation resources is a warning state, the shared transportation platform is activated, and the blockchain smart contract automatically matches third-party transportation resources;

[0165] If the state of the transportation resources is an emergency state, dynamic path re-planning is started; according to the grid priority weight, the peak grid resources are preferentially allocated;

[0166] The demand peak grid is clustered according to spatial proximity, the future period road network state is predicted based on a deep spatio-temporal graph neural network, and the path weight is dynamically adjusted; the external transport capacity is verified through a blockchain shared transport platform and an intelligent contract is signed;

[0167] An output verifiable operation scheme is generated, a structured decision package including vehicle dispatching paths, human resource scheduling tables, and warehouse allocation schemes is generated, and the decision package hash value is stored in the blockchain.

[0168] Specifically, the current transport resource state is obtained in real time from the system transport resource state evaluation, including the availability of vehicle, human and warehouse resources, and the comprehensive evaluation of the transport resource state includes: normal, warning, and emergency. The peak demand distribution information of each grid unit in the target logistics service area is obtained from the demand prediction model, including the grid location of the peak, the expected freight transport volume, etc. The historical logistics data including vehicle dispatching records, freight transport order details, and warehouse inventory changes are obtained by calling the structured spatio-temporal data set in the storage chain through the blockchain API interface, providing data support for the optimization algorithm.

[0169] The total operation cost is minimized as the objective function of optimization. The total operation cost can include transportation cost including: vehicle fuel cost, road toll, driver salary, etc., warehouse cost including: freight storage cost, warehouse operation cost, etc., human cost including: employee salary, overtime pay, etc., and the cost of using shared transport, etc.

[0170] Among them, the decision variable determination includes: the vehicle dispatching path set includes: determining the driving path of each vehicle, including the starting point, the passing point and the ending point, and the freight loading and unloading operation at each point. The dynamic allocation matrix of human resources includes: making employee work task allocation plan, such as allocating employees to different vehicle loading and unloading operations, warehouse management tasks or transportation monitoring posts, etc. The warehouse resource allocation scheme includes: planning the freight storage and allocation scheme of each warehouse facility, including the in-out warehouse time, storage location arrangement and freight allocation between different warehouse facilities, etc. The transport resource state is taken as the core constraint condition. For example, according to the vehicle resource availability, the number of schedulable vehicles and the carrying capacity are constrained; according to the human resource availability, the number of assignable employees and the working time are constrained; according to the warehouse resource availability, the available warehouse space and the number of equipment are constrained, etc.

[0171] If the state of the transport resource is normal, a genetic algorithm is used to solve the scheduling scheme with the lowest cost; the population size of the genetic algorithm is determined, such as 100-200 individuals, the number of iterations is determined, such as 50-100 generations, the crossover probability is determined, such as 0.8-0.9, and the mutation probability is determined, such as 0.1-0.2. The decision variables are coded, such as representing the vehicle scheduling path as a gene sequence on a chromosome, each gene representing a driving section or operation of a vehicle; the human resource allocation and warehouse allocation scheme is similarly coded. Then the coding on the chromosome is converted into the actual scheduling scheme through decoding. The total operating cost is used as the fitness function, and the fitness value of each individual is calculated. The lower the fitness value, the lower the operating cost of the scheduling scheme corresponding to the individual, and the better the individual. According to the fitness value, the excellent individuals are selected to enter the next generation population, the crossover operation is performed to generate new individuals, and the mutation operation is performed on the individuals with a certain mutation probability to increase the diversity of the population. Repeat the iteration until the set number of iterations is reached, and obtain the optimal or better scheduling scheme.

[0172] If the state of the transport resource is a warning state, the shared transport platform is activated, and the blockchain smart contract automatically matches the third-party transport capacity; when the state of the transport resource is a warning, the transport matching function of the shared transport platform is automatically triggered through the blockchain smart contract. The smart contract publishes the transport task demand to the third-party transport capacity according to the current transport demand and available vehicle information. The qualification of the third-party transport capacity responding to the task is verified by using the stored hash value of the qualification file on the blockchain. The qualification file hash value provided by the third-party transport capacity is compared with the stored hash value on the blockchain to ensure the authenticity and reliability of the qualification. Once the third-party transport capacity passes the qualification verification and reaches a transport agreement with the shipper, the smart contract automatically signs the transport contract to clearly define the rights and obligations of both parties, including the transport price, delivery time, and insurance of the goods. The smart contract will automatically execute the contract terms, such as automatically paying the transport fee to the third-party transport capacity after the goods are delivered.

[0173] If the transport resource state is in an emergency state, start dynamic path re-planning; according to the grid priority weight, preferentially allocate peak grid resources; when the transport resource state is in an emergency, immediately start the dynamic path re-planning mechanism. According to the peak distribution information of the freight transport demand and the grid priority weight, preferentially allocate resources to meet the peak grid freight transport demand. The demand peak grid is clustered according to spatial proximity to form multiple transport demand hotspots. For example, the K-Means clustering algorithm or the DBSCAN density clustering algorithm can be used to cluster according to the geographical location distance between grids and the similarity of transport demand. Based on the deep spatiotemporal graph neural network model, the road network state in the future period is predicted, including road congestion, predicted travel time, etc. According to the prediction result, the path weight is dynamically adjusted to provide a more optimal path selection for vehicle scheduling. For example, if a road is predicted to have severe congestion in the next 1 hour, the path weight of the road segment is increased to make the vehicle avoid the road segment as much as possible.

[0174] Publish the urgent transport demand through the blockchain shared transport platform to attract external transport to participate. Qualify the responding external transport, and sign a smart contract with it after passing the qualification to ensure that it completes the transport task on time and with quality.

[0175] The obtained optimized scheduling scheme is arranged into a structured decision package, which includes detailed vehicle scheduling paths including: the travel route of each vehicle, the freight loading and unloading sequence, the human resource scheduling table including: the work tasks, work time and work location of employees, and the warehouse allocation scheme including: the storage location of goods, the in-out warehouse time arrangement and other information.

[0176] The decision package file is calculated for a hash value to obtain a unique hash identifier. The hash value is stored in the blockchain distributed ledger through the blockchain API to ensure the non-tamperability and traceability of the decision package. At the same time, the related metadata of the decision package such as the generation time, the data hash value according to which the decision package is generated, etc. are also stored to facilitate subsequent query and audit.

[0177] In this embodiment, the optimization management module includes the following steps:

[0178] According to the decision variables in the logistics operation scheme, the cost factors including: transportation cost, warehouse cost, labor cost, shared transport cost are extracted; the total cost is obtained by statistical summation;

[0179] The cost benefit value formula is:

[0180] Wherein, y before is the resource coefficient before optimization of the logistics operation scheme, y after is the resource coefficient after optimization of the logistics operation scheme, and cb is the total cost; T before and Tafter respectively are the average delivery time of orders before and after optimization; ω1 and ω2 are weight coefficients;

[0181] If the cost-benefit value is less than the cost-benefit threshold, the current logistics operation scheme is re-optimized; otherwise, the scheme is stored in the blockchain.

[0182] Specifically, the transportation cost related data is extracted from the decision variables of the logistics operation scheme, including the driving mileage, fuel consumption, toll fee, driver's salary, etc. According to the actual transportation path and vehicle scheduling arrangement, the transportation cost of each vehicle is calculated. For example, for each vehicle, the transportation cost can be calculated according to the following formula: transportation cost = driving mileage × (fuel consumption × oil price + toll fee unit price) + driver's salary, wherein the driving mileage can be calculated by the vehicle scheduling path and map data; the fuel consumption is estimated according to the vehicle type and historical fuel consumption data; the oil price is the current market oil price; the toll fee unit price is determined according to the actual road toll standard; the driver's salary is calculated according to the driving time and salary standard.

[0183] The storage information of goods in the warehouse resource allocation scheme is extracted, and the warehouse cost is calculated. The warehouse cost mainly includes warehouse rental fee, goods storage fee, warehouse operation cost such as equipment depreciation, water and electricity fee, etc. For example: the warehouse cost is equal to the warehouse area usage amount × unit area rental fee + the goods storage amount × unit goods storage fee + the warehouse operation fixed cost; the warehouse area usage amount is calculated according to the storage demand of goods and the warehouse layout; the unit area rental fee and the unit goods storage fee can be determined according to the warehouse service contract; the warehouse operation fixed cost includes equipment depreciation, water and electricity fee, etc. daily operation expenditure.

[0184] According to the dynamic allocation matrix of human resources, the working time and working intensity of employees in each post are counted, and the human cost is calculated. The human cost includes the basic salary, overtime pay, welfare fee, etc. of employees. For example: the human cost is equal to the sum of the basic salary, overtime pay, and welfare fee of all employees.

[0185] If shared transport such as third-party transport or external transport is used in the logistics operation scheme, relevant data is extracted from the shared transport usage record, and the shared transport cost is calculated. The shared transport cost mainly includes the transportation fee and service fee paid to the shared transport provider. For example: the shared transport cost is equal to the shared transport usage times × single transportation fee + shared transport service fee. Among them, the shared transport usage times and the single transportation fee are counted according to the actual transportation contract and the transportation task completion situation; the service fee is calculated according to the charging standard of the shared transport platform. The transportation cost, warehouse cost, human cost and shared transport cost extracted above are summarized, and each cost is added up by statistical summation to obtain the total cost value of the entire logistics operation scheme.

[0186] The resource coefficient is obtained from the transport resource state evaluation module before and after the optimization of the logistics operation scheme. The resource coefficient reflects the utilization efficiency and comprehensive state of the transport resource.

[0187] From historical order data and current order execution, the average delivery time of orders before and after optimization is calculated. This can be done by calculating the average difference between the delivery time and the order time of a certain number of orders.

[0188] According to the degree of attention of the enterprise to cost and delivery time, the weight coefficient is determined, and the sum is 1. If the enterprise pays more attention to cost control, ω1 can be set to 0.6, and ω2 can be set to 0.4. If more attention is paid to delivery time, the weight ratio can be adjusted. The cost benefit value is calculated according to the given formula.

[0189] According to the enterprise cost management and operation benefit target, a cost benefit threshold is set. This threshold can be determined by referring to historical data, industry standards or enterprise strategic goals. If the enterprise expects the cost benefit value to be at least 0.1, that is, each unit of cost can bring 0.1 of benefit improvement, the threshold is set to 0.1.

[0190] Compare the calculated cost benefit value with the set threshold. If the cost benefit value is less than the threshold, it means that the optimization effect of the current logistics operation scheme does not meet the expectation, and optimization needs to be performed again; otherwise, if the cost benefit value is greater than or equal to the threshold, it is considered that the scheme has good cost benefit, and it can be stored in the blockchain.

[0191] For the logistics operation scheme that meets the cost benefit requirement, its related key information such as vehicle scheduling path, human resource allocation plan, warehouse allocation scheme, etc. is arranged into a structured document or data format. The hash value of the document or data is calculated to obtain its unique hash identifier. Through the blockchain API interface, the hash value is stored in the blockchain distributed ledger to ensure the non-tamperability and traceability of the scheme, so as to query, audit and trace the execution of the scheme in the future.

[0192] The above examples are only used to illustrate the technical method of the present application and are not limited. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. Intelligent big data logistics operation system, characterized by: Includes the following modules: Data collection and processing module: real-time collection of multi-source logistics data and pre-processing, building a structured spatiotemporal data set; using blockchain technology to store the structured spatiotemporal data set; Data Analysis Module: Analyzes and calculates vehicle resource availability, human resource availability, and warehousing resource availability based on structured spatiotemporal datasets. Combines these analysis results to assess the status of transportation resources in real time. Analyzes peak freight transport demand using a demand forecasting model built using a deep spatiotemporal graph neural network. Operations Management Module: Manages transportation resources by building a shared transportation platform based on blockchain. It uses optimization algorithms to establish logistics operation plans based on the status of transportation resources and peak cargo transportation demand. Optimization management module: Optimize the operation management plan by analyzing the cost-effectiveness of the logistics operation plan; Visualization display module: Through the web front-end interactive dashboard, dynamically display the transportation resource status, cargo transportation demand peak, logistics operation plan, and cost-effectiveness results.

2. The intelligent big data logistics operation system according to claim 1 is characterized in that: Using blockchain technology to store structured spatiotemporal datasets includes the following steps: Integrate with the Hyperledger Fabric consortium chain platform to build a full-process chain architecture for data collection, transmission and storage; Use Apache Kafka distributed message queues to buffer high-concurrency logistics data flows and establish a data pre-processing buffer layer; Encapsulate the buffered structured spatiotemporal dataset into a blockchain transaction proposal; Send the transaction proposal to the endorsement node cluster, and obtain the signature endorsement after the chaincode smart contract verifies the validity of the transaction logic; package the endorsed transactions into candidate blocks in a determined order; The candidate blocks are broadcast to all peer nodes, and after verification of read-write set conflicts and endorsement policies, they are written into the distributed ledger in order; The encrypted hash value or metadata is stored in the blockchain ledger to form a traceable logistics data storage chain.

3. The intelligent big data logistics operation system according to claim 1 is characterized in that: Calculating vehicle resource availability, human resource availability, and storage resource availability based on structured spatiotemporal dataset analysis includes the following steps: The formula for calculating vehicle resource availability is: Among them, Q 可 is the number of available vehicles, Q 总 is the total number of vehicles, Z i,max is the vehicle load of the i-th available vehicle, Z i is the current load of the i-th available vehicle; α1, α2 and α3 are weight coefficients, and their sum is 1; T i is the available time of the i-th available car, T c总 is the total available time of all vehicles; The formula for calculating human resource availability is: Among them, P 总 is the total number of people, P 在岗 is the number of people on the job; G j (t) is the length of the real-time task queue of station j, G max is the maximum load capacity, w j is the priority weight of the task at workstation j, m is the number of workstations; β1, β2 and β3 are weight coefficients, and their sum is 1; T 在岗 is the total working hours of employees on the job, T p总 is the total working hours; The formula for calculating the availability of storage resources is: Among them, A 总 is the total storage capacity, A 可 is the current available storage capacity, H k is the total amount of goods of the kth type entering and leaving the warehouse during the time period, H P,k is the average occupancy value of the storage capacity of the kth type of goods in the time period, T Z is the time period, n is the number of goods, v k is the priority weight of the kth cargo; γ1, γ2 and γ3 are weight coefficients, and their sum is 1; S 可用 is the number of devices in normal available state, S 总 It is the total number of storage facilities and equipment.

4. The intelligent big data logistics operation system according to claim 3 is characterized in that: Based on the above analysis results, the real-time assessment of the transport resource status includes the following steps: According to the vehicle resource availability rate, human resource availability rate and storage resource availability rate, the resource coefficient y is calculated using the weighted average summation formula; The formula for calculating the available capacity percentage is: Where D is the real-time order quantity, order_max is the maximum order quantity allowed, O is the meteorological environment coefficient of the target logistics service area, J V is the real-time average speed to the target logistics service area, and J is the historical average speed to the target logistics service area; According to the percentage of available capacity, set the available capacity threshold and classify the capacity resource status into the following: When Y ≥ 85%, the capacity resource status is normal; When 50%≤Y<85%, the transport resource status is in a warning state; When Y<50%, the capacity resource status is emergency.

5. The intelligent big data logistics operation system according to claim 1, characterized in that: The demand forecasting model built based on the deep spatiotemporal graph neural network analyzes the peak demand for freight transportation, including the following steps: Divide the target logistics service area into grid cells, with each grid cell serving as a graph node. Build a weighted adjacency matrix based on the actual road network or logistics path. Map the structured spatiotemporal dataset to the corresponding grid nodes according to the preset time window length to form a spatiotemporal graph sequence; Construct a deep spatiotemporal graph neural network model including a graph convolution layer and a time series modeling layer; collect multi-source logistics data at the current moment, extract the characteristics of cargo transportation volume and order volume, and map them to the grid nodes; The real-time graph structure is input into the trained deep spatiotemporal graph neural network model, which outputs the predicted value of freight transportation demand corresponding to each grid cell within the specified future time window. For each grid cell, calculate the average value of cargo transportation demand in the same time window within the past set number of days to obtain the historical reference value for the same period; When the forecast value of freight transport demand is greater than or equal to the product of 2 times the historical reference value for the same period, it is determined that a freight transport demand peak has occurred in the grid.

6. The intelligent big data logistics operation system according to claim 1, characterized in that: By building a shared transportation platform based on blockchain, we can manage transportation resources, including the following steps: Build a dynamic matching layer for transportation resources, establish a decentralized transportation registration pool based on the Hyperledger Fabric consortium chain, and generate real-time structured spatiotemporal data sets; Utilize smart contracts deployed on the blockchain to manage the transport transaction process; these smart contracts include encrypted quote generation, hash verification of carrier qualification documents, and automated transport contract execution logic. The agreed capacity transaction terms and smart contract execution status will be recorded as transaction results in the blockchain distributed ledger.

7. The intelligent big data logistics operation system according to claim 6, characterized in that: Based on the transport resource status and peak cargo transport demand, an optimization algorithm is used to establish a logistics operation plan, which includes the following steps: Construct a multi-objective optimization model, where the objective function is to minimize the total operating cost; the decision variables include the vehicle scheduling path set, the human resource dynamic allocation matrix, and the warehouse resource allocation plan; the transportation resource status is the core constraint; If the transport resource status is normal, the genetic algorithm is used to solve the scheduling plan with the optimal cost; If the transport resource status is in a warning state, the shared transport platform will be activated, and the blockchain smart contract will automatically match third-party transport capacity; If the transport resource status is emergency, dynamic route replanning is initiated; peak grid resources are allocated first according to grid priority weights; Clustering peak demand grids by spatial proximity, predicting future road network status based on deep spatiotemporal graph neural networks, and dynamically adjusting path weights; verifying external transport capacity qualifications and signing smart contracts through a blockchain-based shared transport capacity platform; Output a verifiable operation plan, generate a structured decision package including vehicle dispatch routes, human resource schedules, and warehouse allocation plans, and store the decision package hash value in the blockchain.

8. The intelligent big data logistics operation system according to claim 1, characterized in that: The optimization management module includes the following steps: Based on the decision variables in the logistics operation plan, the cost factors are extracted, including transportation cost, warehousing cost, labor cost, and shared transportation cost; the total cost is obtained through statistical summation; The formula for calculating cost-effectiveness is: Among them, y before is the resource coefficient before the optimization of the logistics operation plan, y after is the resource coefficient after the logistics operation plan is optimized, cb is the total cost; T before and T after are the average order delivery times before and after optimization respectively; ω1 and ω2 are weight coefficients; If the cost-effectiveness value is less than the cost-effectiveness threshold, the current logistics operation plan will be re-optimized; otherwise, the plan will be recorded in the blockchain.

Images