Multi-source data fusion vaccine supply chain collaborative optimization method and system

Abstract

The application provides a vaccine supply chain collaborative optimization method and system based on multi-source data fusion, and relates to the technical field of vaccine management. The application obtains the immunization archives and the vaccination plan of the target inoculation object, extracts the inoculation track and performs time matching to form a demand queue. The dose gap and available inventory are calculated in combination with the demand queue, and the off-site inoculation details are desensitized, and the travel time between inoculation units is calculated to generate a collaborative resource pool. The resource occupation sequence and inoculation execution timing are generated in the collaborative resource pool. The adverse reaction symptoms are combined to extend the observation time and split the inoculation control nodes. When there is an off-site inoculation task, the available inventory is mapped and distributed to form an inoculation task sequence. Finally, according to the actual execution information, the un-validated doses and temperature abnormal nodes are identified to generate a task compensation queue, which can realize the closed-loop collaboration of vaccine demand, inventory, travel, observation and abnormal compensation.

Description

Technical Field

[0001] This application relates to the field of vaccine management technology, and in particular to a method and system for collaborative optimization of the vaccine supply chain through multi-source data fusion. Background Technology

[0002] The vaccine supply chain collaborative optimization method is mainly used to unify the management of vaccine procurement, inventory, cold chain distribution, vaccination implementation and abnormal handling, so that vaccine resources can be allocated according to actual vaccination needs. With the increasing requirements for cross-regional vaccination, appointment-based vaccination and cold chain traceability, this type of method has good application prospects in regional immunization programs, primary vaccination services and public health emergency response.

[0003] In the prior art, the patent with publication number CN118798765B discloses the following: a solution for predicting vaccine demand, generating replenishment lists, planning delivery routes, and monitoring temperature, humidity and vehicle location during transportation by collecting real-time and historical data from vaccination sites. This method can manage vaccine replenishment and cold chain logistics, and focuses on solving the problems of inventory replenishment and transportation environment control.

[0004] However, in actual vaccination scenarios, vaccination demand, cross-unit undertaking, inventory holding, cold chain anomalies, and execution feedback are often scattered across different business processes. Existing methods struggle to establish continuous coordination between the vaccination and supply ends. When vaccination tasks are not completed or cold chain conditions deteriorate, subsequent compensation arrangements are prone to lag, affecting the timely allocation of vaccine resources and the continuous execution of the vaccination plan. Therefore, existing technologies suffer from insufficient coordination between vaccination execution and anomaly compensation in the vaccine supply chain. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for collaborative optimization of the vaccine supply chain based on multi-source data fusion, so as to solve the problem of insufficient coordination between vaccine delivery execution and anomaly compensation in the existing technology.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a multi-source data fusion-based collaborative optimization method for the vaccine supply chain, the method comprising:

[0007] Obtain the immunization records and vaccination plans of the target vaccination subjects; extract the vaccination trajectory corresponding to the target vaccination subjects based on the historical vaccination information in the immunization records; and match the vaccination trajectory with the vaccination plan in time to generate a demand queue for each vaccination unit.

[0008] Based on the demand queue, the vaccine dose shortage and available inventory of each vaccination unit are calculated. The off-site vaccination details in the immunization record are anonymized to obtain anonymized records. At the same time, the travel time between each vaccination unit is calculated within a preset reference framework. A collaborative resource pool is generated based on the anonymized records and the travel time.

[0009] Within the collaborative resource pool, a resource occupancy sequence is generated according to the order of vaccination dates, and the doses continuously received by the same target vaccination subject in different vaccination units are organized into a vaccination execution sequence according to time.

[0010] Based on the adverse reaction symptoms of the target subjects in the immunization records, the observation time after vaccination is extended. The vaccination execution sequence is divided into multiple vaccination control nodes with the vaccination window and the extended observation time as the boundary.

[0011] When there are off-site vaccination tasks in the collaborative resource pool, the available inventory is mapped and allocated according to the resource occupancy sequence and the occupancy identifier is updated. The target occupancy identifier and the corresponding vaccination control node are concatenated according to the vaccination timestamp to form a vaccination task sequence.

[0012] Obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

[0013] Secondly, this application provides a multi-source data fusion-based vaccine supply chain collaborative optimization system, which includes:

[0014] The extraction module is used to obtain the immunization records and vaccination plans of the target vaccination subjects, extract the vaccination trajectory corresponding to the target vaccination subjects based on the historical vaccination information in the immunization records, and match the vaccination trajectory with the vaccination plan in time to generate a demand queue for each vaccination unit.

[0015] The calculation module is used to calculate the vaccine dose shortage and available inventory of each vaccination unit based on the demand queue, perform data anonymization processing on the off-site vaccination details in the immunization file to obtain anonymized records, and calculate the travel time between each vaccination unit within a preset reference framework, and generate a collaborative resource pool based on the anonymized records and the travel time.

[0016] The organization module is used to generate a resource occupancy sequence according to the order of vaccination days within the collaborative resource pool, and to organize the doses that the same target vaccination object receives consecutively in different vaccination units into a vaccination execution sequence in chronological order.

[0017] The splitting module is used to extend the observation time after vaccination by combining the adverse reaction symptoms of the target vaccination subjects in the immunization record. The vaccination execution sequence is split into multiple vaccination control nodes with the vaccination window and the extended observation time as the boundary.

[0018] The splicing module is used to map and allocate the available inventory according to the resource occupancy sequence and update the occupancy identifier when there are off-site vaccination tasks in the collaborative resource pool, and splice the target occupancy identifier and the corresponding vaccination control node according to the vaccination timestamp to form a vaccination task sequence.

[0019] The generation module is used to obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

[0020] The multi-source data fusion-based vaccine supply chain collaborative optimization method provided in this application has the following beneficial effects: By acquiring immunization records and vaccination plans and generating demand queues for each vaccination unit, this application can transform individual vaccination progress into demand arrangements that vaccination units can handle; by calculating vaccine dose gaps and available inventory and combining anonymized records and travel time to generate a collaborative resource pool, it can determine the resource relationships that can be handled while protecting vaccination privacy; by generating resource occupancy sequences and vaccination execution sequences, it can arrange cross-unit vaccination tasks continuously according to the vaccination date sequence; by combining adverse reaction symptoms to extend the observation time and splitting vaccination control nodes, it can make verification, vaccination, and observation arrangements more in line with individual risks; by mapping and allocating available inventory and generating vaccination task sequences, it can implement inventory occupancy relationships into executable tasks; by identifying unverified doses and temperature abnormality nodes and generating task compensation queues, it can provide a basis for subsequent supplementary vaccination, verification, and abnormal handling. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a multi-source data fusion-based collaborative optimization method for the vaccine supply chain disclosed in this application.

[0023] Figure 2 This is a regional vaccination collaborative application layout diagram disclosed in this application;

[0024] Figure 3 This application discloses a flowchart of the vaccination observation sequence breakdown.

[0025] Figure 4 This application discloses a compensation queue ledger structure diagram;

[0026] Figure 5 This is a schematic diagram of the structure of a multi-source data fusion collaborative optimization system for the vaccine supply chain disclosed in this application.

[0027] Figure label:

[0028] 21-Regional boundary; 22-Road network topology; 23-Vaccination unit; 24-Cold chain storage point; 25-Cold chain delivery vehicle; 26-Vaccination recipient location; 41-Task compensation ledger; 42-Assignment weight record; 43-Collaborative resource pool record; 44-Immunization record backfilling record. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] Figure 1 This is a flowchart illustrating a multi-source data fusion-based collaborative optimization method for the vaccine supply chain disclosed in this application. The method includes:

[0031] S101. Obtain the immunization records and vaccination plans of the target vaccination subjects. Based on the historical vaccination information in the immunization records, extract the vaccination trajectory corresponding to the target vaccination subjects and match the vaccination trajectory with the vaccination plan in time to generate the demand queue of each vaccination unit.

[0032] The immunization record includes data such as the vaccination records completed by the target recipients, vaccine names, dose numbers, and vaccination completion status; the vaccination plan includes data such as the vaccination windows opened by the vaccination unit, the types of vaccines available, and the planned number of doses; the vaccination trajectory is used to represent the vaccination continuity between the target recipients' completed doses and the doses to be completed; and the demand queue is used to record the target vaccination doses that each vaccination unit needs to receive within its corresponding vaccination window.

[0033] In this embodiment of the application, S101 specifically includes:

[0034] S1011. Analyze the historical vaccination time and the standard interval between adjacent doses in the historical vaccination information.

[0035] Historical vaccination information includes the historical vaccination time, the name of the vaccine administered, the number of the completed doses, and the vaccination completion status. The standard interval is used to indicate the vaccination time requirement between adjacent doses of the same vaccine, such as at least 1 month, or 1 to 2 months, which can be derived from immunization program rules, vaccination rule tables, or pre-configured business rules of the vaccination unit.

[0036] In this embodiment, historical vaccination information is read from the immunization record by field recognition, and the corresponding standard interval time is queried according to the vaccine name and dose number. When reading, it is first confirmed whether the historical vaccination information is in a completed state, and then the corresponding historical vaccination time is extracted to avoid using cancellation, no-show or incomplete records as the basis for subsequent dose calculation.

[0037] S1012. Based on the historical vaccination time and the standard interval time, calculate the vaccination time range of the target doses that have not yet been administered, and form a vaccination trajectory with time constraints.

[0038] Among them, the target vaccination dose refers to the dose that the target vaccination target has not yet completed according to the vaccination sequence and needs to be arranged; the vaccination time range is used to indicate the time interval in which the target vaccination dose can be arranged.

[0039] In this embodiment of the application, the vaccination time range of the target dose is first determined by starting from the historical vaccination time and pushing forward at least one month, or one to two months, etc., standard interval time. Then, the completed doses, the target doses, and the vaccination time range are connected in chronological order to form a vaccination trajectory. This vaccination trajectory is used to represent the continuous vaccination relationship of the target subject in terms of vaccine doses.

[0040] S1013. Calculate the overlap interval between the vaccination time range and the target vaccination window arranged in the vaccination plan, and extract the target vaccination doses within the overlap interval.

[0041] The target vaccination window refers to the time period in the vaccination plan that is available for vaccination units to actually schedule vaccinations; the overlapping period is used to indicate the available time period that simultaneously meets the requirements of the vaccination time range and the target vaccination window. This step can prevent the target doses from being scheduled in vaccination windows that are not actually available in terms of time.

[0042] S1014. Based on the geographical coordinates of each vaccination unit and the spatial relationship between the target vaccination object, the selected target vaccination doses are associated and allocated to the corresponding vaccination units to generate a demand queue.

[0043] Among them, geographic coordinates are used to represent the actual service location of the vaccination unit; spatial location relationship is used to represent the distance relationship, service range relationship or appointment selection relationship between the target vaccination object and the vaccination unit. If the target vaccination object has already selected a vaccination unit on the appointment platform, the appointment selection relationship can be used first; if no vaccination unit has been selected, the spatial location relationship can be determined based on the registered address, commonly used vaccination units or service radius.

[0044] Specifically, the geographical coordinates of each vaccination unit are first read, and then the spatial relationship between the target vaccination target and each vaccination unit is determined by combining the registered address, appointment selection location or commonly used vaccination unit of the target vaccination target. Then, the target vaccination doses are allocated to vaccination units that are available in terms of time and space, and the demand queues of each vaccination unit are generated separately according to the vaccination unit.

[0045] The above steps enable the historical vaccination information in the immunization record to be continuously linked with the vaccination window in the vaccination plan, thereby providing a clear demand basis for subsequent inventory matching, cross-institutional collaboration, and vaccination task generation.

[0046] S102. Based on the demand queue, calculate the vaccine dose shortage and available inventory of each vaccination unit, perform data anonymization processing on the off-site vaccination details in the immunization record to obtain anonymized records, and calculate the travel time between each vaccination unit within a preset reference framework. Generate a collaborative resource pool based on the anonymized records and the travel time.

[0047] Among them, vaccine dose shortage is used to indicate the number of doses that a vaccination unit needs to receive within a certain vaccination window but whose existing available inventory cannot cover; available inventory is used to indicate the vaccine inventory that has not expired, has not been frozen, is in normal cold chain condition, and can be handed over before the vaccination window; off-site vaccination details include records of changes in vaccination location for the target vaccination subject between different vaccination units;

[0048] The anonymized records are used to retain the statistical characteristics of off-site vaccination behavior while hiding personal identity associations; the preset reference framework is used to place the location of vaccination unit, road network nodes, road sections, cold chain equipment, handover points and ambient temperature collection time periods in the same calculation basis, and use vaccination unit number, road section number and collection time as association fields; the collaborative resource pool is used to record which vaccination units can share resources when demand, inventory and cold chain timeliness are all met.

[0049] like Figure 2As shown, the regional boundary 21 represents the geographical area covered by the vaccination collaborative service, and the road network topology 22 represents the road connection relationship between vaccination units 23, cold chain storage points 24 and vaccination recipient locations 26 within the region. Multiple vaccination units 23 are distributed in different locations in the road network topology 22, the cold chain storage point 24 is set on the starting branch road of the road network topology 22, and the cold chain delivery vehicle 25 is located on the delivery path between the cold chain storage point 24 and the vaccination unit 23. The vaccination recipient location 26 is connected to the road segment of the road network topology 22 through the branch road, which is used to represent the spatial relationship between the target vaccination recipient and the vaccination unit that can accept the vaccination.

[0050] In this embodiment of the application, S102 specifically includes:

[0051] S1021. Extract the vaccination frequency and vaccination distance from the vaccination details in the different locations, dynamically adjust the privacy budget based on the vaccination distance in the different locations, and inject random perturbation values ​​that satisfy a preset probability distribution into the vaccination frequency to generate desensitized records.

[0052] Among them, vaccination frequency is used to represent the number of times the target vaccination subject or target area appears in cross-vaccination units, where the actual vaccination unit is different from the original vaccination unit; cross-location vaccination distance is used to represent the distance between the original vaccination unit and the actual vaccination unit, which can be divided into several distance levels. Each distance level corresponds to a privacy budget. The higher the distance level, the lower the corresponding privacy budget, and the stronger the rewriting of the vaccination frequency; the privacy budget is a value used to control the amplitude of random perturbation. When using Laplace perturbation, the smaller the privacy budget, the greater the fluctuation of the random perturbation value, and the stronger the desensitization strength; the larger the privacy budget, the smaller the fluctuation of the random perturbation value, and the higher the fidelity of the statistical results; the random perturbation value is used to rewrite the vaccination frequency so that the processed data can still reflect the business trend, but cannot directly restore the real behavior of a single target vaccination subject; the preset probability distribution can refer to the Laplace distribution.

[0053] The vaccination frequency in the desensitization records is shown in formula (1):

[0054] (1)

[0055] in, The original vaccination frequency; The frequency of vaccination in the desensitization record; For random perturbation values, obey , Indicates the Laplace distribution, This represents the maximum impact of a single out-of-town vaccination record on vaccination frequency. This is a privacy budget determined based on the distance of vaccinations in different locations; the above parameters are used for vaccination frequency statistics and do not indicate the identity of the target vaccination recipient. If the vaccination frequency in the anonymized record is not an integer, it will be processed according to the existing rounding rules.

[0056] As a specific example, the original vaccination frequency of target vaccination subjects in a certain region in the off-site vaccination details. The maximum impact of a single off-site vaccination record on vaccination frequency is 8. The privacy budget is set at 1, based on the distance between vaccination locations. It is 0.5, according to the Laplace distribution. If a random disturbance value is drawn this time, then... If the value is -1.2, then the vaccination frequency in the desensitization records is... This preserves the statistical significance of the frequency of vaccinations in different locations while avoiding direct exposure of actual vaccination behavior. All exemplary values ​​in the above embodiments are for illustrative purposes only and are not intended to limit the scope of this application.

[0057] S1022. Within a preset reference frame that serves as a spatial measurement benchmark, analyze the actual trajectory distance of each inoculation unit in the road network topology, and determine the travel time based on the actual trajectory distance.

[0058] Among them, the road network topology is used to represent the connection relationship between roads and junctions between vaccination units; the actual trajectory distance is used to represent the actual road distance traveled by cold chain transport vehicles from one vaccination unit to another.

[0059] In this embodiment, each inoculation unit and the cold chain handover point are first treated as nodes in the road network topology, and road sections are treated as connecting edges between nodes. The shortest path search technology is used to analyze the actual trajectory distance between each inoculation unit. This process does not simply use straight-line distance, but finds the passable actual trajectory path according to the road network topology.

[0060] The transportation time is calculated segment by segment based on the actual trajectory distance, and the transportation time of each segment is added together to obtain the travel time, which is used to ensure that vaccine batches can be received within the cold chain time limit. In this embodiment, the dynamic speed limit threshold can be calculated according to formula (2):

[0061] (2)

[0062] in, For the first Dynamic speed limit thresholds for each road section; For the first Speed ​​limits for each road section; For cold chain delivery vehicles in the first The rated stable speed of each road section.

[0063] The passage time can be calculated according to formula (3):

[0064] (3)

[0065] in, For the time taken to pass through; For the first The actual trajectory distance of each road segment; For the first Dynamic speed limit thresholds for each road section.

[0066] S1023. Based on the desensitized records, target vaccination subjects with a tendency to be vaccinated in other locations are screened out, the target vaccine batches required by the screened target vaccination subjects are determined, the cold chain timeliness of the target vaccine batches is extracted, and the vaccination units with a passage time shorter than the cold chain timeliness are matched and associated with resources to build a collaborative resource pool.

[0067] Among them, "individual vaccination tendency" is used to indicate the possibility that the target vaccination target has the business of completing vaccination across vaccination units; "target vaccine batch" is used to indicate the specific vaccine batch that can meet the required number of doses for the target vaccination target; "cold chain timeliness" is used to indicate the remaining time that the target vaccine batch can still meet the temperature requirements from leaving the warehouse, transportation, and handover to when it is ready for vaccination; "resource matching association" is used to record the transferable and acceptable relationship between the supply vaccination unit and the demand vaccination unit.

[0068] As a specific example, the cold chain delivery vehicle needs to pass through two road sections from the cold chain storage point to the vaccination unit: in the first road section, It is 60km / h. If it is 50km / h, then =50km / h; in the second road section It is 40km / h. If it is 45km / h, then =40km / h. If the actual track distances of the two road sections are 12km and 8km respectively, then the travel time is... 12 / 50 + 8 / 40 = 0.44 h, or 26.4 min. If the cold chain time for the target vaccine batch is 60 min, then this path can enter the collaborative resource pool.

[0069] This application, through step 102, can protect the privacy of vaccination details in different locations while unifying demand queues, available inventory, road network access, and cold chain timeliness into the same collaborative judgment process, so that the resource matching and association in the collaborative resource pool has an executable business basis and cold chain constraint basis.

[0070] S103. Within the collaborative resource pool, a resource occupancy sequence is generated according to the order of vaccination dates, and the doses continuously received by the same target vaccination subject in different vaccination units are organized into a vaccination execution sequence according to time.

[0071] Among them, the resource occupancy sequence is used to represent the queuing order and time occupancy relationship of vaccination units on the vaccination day; the vaccination execution sequence is used to represent the actual execution order when the same target vaccination object completes multiple doses across vaccination units.

[0072] In this embodiment of the application, S103 specifically includes:

[0073] S1031. Obtain the queuing sequence of each vaccination unit on the vaccination day, and extract the appointment time slot corresponding to the queuing sequence.

[0074] S1032. Generate a resource occupancy sequence based on the queuing sequence and the reservation time period.

[0075] The queuing sequence indicates the order in which individuals are vaccinated at a vaccination unit on that day; the appointment time slot indicates the available vaccination time slot for each queuing position.

[0076] S1033. Retrieve multiple vaccination details that the same target vaccination subject needs to complete within a preset time period, and identify the vaccination unit to which each vaccination detail belongs.

[0077] Among them, the vaccination details are used to indicate the specific doses that the same target vaccination subject needs to complete within a preset time period; the vaccination unit is used to indicate the actual vaccination site corresponding to each vaccination detail.

[0078] In this embodiment, the target recipient is used as an index to retrieve multiple vaccination details that have not yet been completed within a preset time period from the immunization record and vaccination plan, and then the vaccination units corresponding to these vaccination details are marked respectively. Through this process, the vaccination arrangements of the same target recipient in different vaccination units can be linked together, avoiding the situation of only seeing a single dose and ignoring the continuous completion relationship.

[0079] S1034. Extract the spatial transfer time distributed between different vaccination units from the preset road network topology, insert the spatial transfer time between the planned vaccination times of the vaccination details to perform timing verification, arrange the vaccination details that pass the timing verification in the order of vaccination time, and generate the vaccination execution timing sequence.

[0080] Among them, the road network topology is used to represent the road connection relationship between vaccination units; the spatial transfer time is used to represent the time required for the same target vaccination object to be transferred from one vaccination unit to another; and the timing check is used to determine whether there is enough transfer time between two consecutive vaccination details.

[0081] In this embodiment, the spatial transfer time between different vaccination units is first estimated based on the road network topology. Then, this time is inserted between the planned vaccination times of adjacent vaccination details for verification. If the interval between two planned vaccination times can cover the transfer time, it is determined that the group of vaccination details can be continuously accepted. If it cannot cover the transfer time, the group of vaccination details is not included in the same execution chain. Then, the verified vaccination details are arranged in sequence according to the vaccination time to generate the vaccination execution sequence.

[0082] This application enables the queuing resource occupancy relationship within the vaccination day and the continuous vaccination sequence across vaccination units to be coordinated and consistent in the same time chain through step 103, thereby providing a reliable sequential basis for subsequent vaccination control node splitting and vaccination task sequence generation.

[0083] S104. Based on the adverse reaction symptoms of the target vaccination subjects in the immunization record, extend the vaccination observation time, and divide the vaccination execution sequence into multiple vaccination control nodes with the vaccination window and the extended vaccination observation time as the boundary.

[0084] Among them, adverse reaction symptoms refer to the physical reactions recorded in the target recipients after previous vaccinations; vaccination control nodes are executable time periods formed by breaking down the original vaccination execution sequence.

[0085] like Figure 3 As shown in the embodiment of this application, S104 specifically includes:

[0086] S301. Extract adverse reaction symptoms from the historical records of the target vaccination subject and analyze the component similarity between the historical vaccines and the vaccine to be administered.

[0087] Among them, historical records include the reaction records of the target recipients after previous vaccinations; component similarity is used to determine the degree of similarity between the historical vaccines and the vaccine to be administered in terms of antigen components and excipient components.

[0088] In this embodiment, the historical vaccination records of the target recipient are first read from the immunization record, and then the registered adverse reaction symptoms are extracted, such as local discomfort, fever, rash or other abnormal reactions after vaccination. Then, the component information of the historical vaccines and the vaccines to be administered is read, and the main components of the two are compared to obtain the component similarity. The closer the components are, the higher the attention is paid to the previous reactions during subsequent vaccination, which provides a basis for adjusting the observation time.

[0089] S302. Quantify the risk severity of the adverse reaction symptoms according to the preset assessment mapping table, and use the component similarity to weight and correct the risk severity to obtain the risk gain coefficient.

[0090] Among them, the assessment mapping table is used to map adverse reaction symptoms to rule tables of different risk levels; the risk severity is used to indicate the degree of impact of the symptom on subsequent vaccination and observation; and the risk gain coefficient is used to indicate the correction result when the observation time needs to be extended.

[0091] In this embodiment of the application, the vaccination observation rules and historical adverse reaction records are pre-configured. To facilitate the explanation of how the assessment mapping table classifies adverse reaction symptoms, the following description is provided in conjunction with Table 1:

[0092] Table 1. Assessment Mapping Table for Adverse Reaction Symptoms

[0093]

[0094] The symptom categories and assessment mapping table output in Table 1 illustrate how the severity of the risk is obtained, enabling adverse reaction symptoms to be converted into a risk expression that can be used in subsequent calculations of the risk gain coefficient. The above example is only one example of this application; in practical applications, it can also be set according to vaccination observation rules and historical adverse reaction records, and this application does not limit this.

[0095] S302 specifically includes:

[0096] A1. Retrieve the risk weight value corresponding to the adverse reaction symptoms in the preset assessment mapping table, and confirm the risk weight value as the quantified risk severity.

[0097] In this embodiment, the risk weight value corresponding to the symptom category is found from the assessment mapping table, and then the risk weight value is used as the severity of the risk. After this processing, different symptoms can be converted into a unified risk expression, which is convenient for further modification.

[0098] A2. Extract the antigen and excipient components of the historically administered vaccines and the vaccines to be administered, calculate the component differences between the historically administered vaccines and the vaccines to be administered in the multidimensional feature space, and determine the component similarity based on the component differences.

[0099] In this embodiment, the component information of historically administered vaccines and vaccines to be administered is organized into comparable component items, and then the differences between the two in terms of major components are compared. The smaller the component differences, the higher the component similarity. Through this process, it is possible to determine whether more caution is needed for subsequent vaccinations, taking into account the past reactions of the target recipient.

[0100] A3. Calculate the weighted correction value by multiplying the component similarity with the risk severity, and input the weighted correction value into a preset nonlinear function for mapping operation to output a baseline risk value.

[0101] In this embodiment, the severity of risk is weighted and corrected using component similarity, so that the closer the historical vaccination is to the vaccine to be administered, the more prominent the risk expression is; then the weighted correction value is input into a preset nonlinear function for mapping to obtain a baseline risk value. Component similarity, severity of risk, weighted correction value and baseline risk value are all dimensionless rank values.

[0102] A4. Match the preset age correction coefficient to the physiological age of the target vaccination subject, add the baseline risk value to the age correction coefficient, and obtain the risk gain coefficient.

[0103] The age correction factor is a coefficient used to adjust the observation requirements for different age groups.

[0104] The reason for this approach in this application is that different age groups have varying sensitivities to post-vaccination observation. The age correction factor can take these differences into account, as can be found in Table 2:

[0105] Table 2 Age Correction Factor Mapping Table

[0106]

[0107] The risk gain coefficient can be calculated according to formulas (4)-(6):

[0108] (4)

[0109] (5)

[0110] (6)

[0111] in, To assess the severity of the risk; For component similarity; This is a weighted adjustment value; For piecewise risk mapping function, It can be determined according to preset segmentation mapping rules, for example Output when <1 =1.0, 1≤ Output when <2 =1.2, Output when ≥2 =1.5, the specific range can be configured according to the vaccination observation rules; The benchmark risk value; This is an age correction factor; This is the risk gain coefficient; the above , , , , and All values ​​are dimensionless.

[0112] As a specific example, if a target recipient has experienced a general systemic reaction, the assessment mapping table will determine the severity of the risk. The value was determined to be 2; the similarity of components between historically administered vaccines and the vaccine to be administered. If it is 0.8, then the weighted adjustment value is... =2 × 0.8 = 1.6. If the piecewise risk mapping function specifies 1 ≤ Output the baseline risk value when <2 =1.2, and the age correction factor for the target vaccination population. If it is 0.2, then the risk gain coefficient is... =1.2 + 0.2 = 1.4. Therefore, the observation period after vaccination is extended, and the extended observation period is incorporated into the vaccination control process. The above embodiments are only used to illustrate the processing relationship between risk severity, component similarity, segmented risk mapping function, and age correction coefficient; the listed risk levels, similarities, and correction coefficients do not constitute limitations.

[0113] This application, through A1-A4, can transform adverse reaction symptoms into a uniform risk severity level, and make corrections based on the similarity of components between historical vaccines and the vaccine to be administered. Furthermore, it introduces age correction according to the physiological age of the target recipient, so that the risk gain coefficient simultaneously reflects the symptom type, vaccine component association, and individual age differences, providing a clear and actionable basis for subsequent extension of the observation period after vaccination.

[0114] S303. Multiply the vaccination observation time with the risk gain coefficient to calculate the extended vaccination observation time.

[0115] Among them, the vaccination observation time is used to represent the standard observation duration; the extended vaccination observation time is used to represent the actual observation duration after adjusting for individual risk.

[0116] In this embodiment, the extended observation time after vaccination is obtained by multiplying the standard observation time by the risk gain coefficient. If the risk gain coefficient is high, it indicates that the target subject needs a longer observation time; if the risk gain coefficient is low, the observation time is closer to the standard value.

[0117] S304. Sequentially mark the appointment verification window, vaccination window, and extended vaccination observation time on the timeline to extract time sequence segmentation points.

[0118] The verification window is used for identity confirmation and information verification; the vaccination window is used to complete the actual vaccination; and the time sequence segmentation point is used to mark the boundary between different business stages.

[0119] In this embodiment, the scheduled verification window, vaccination window, and extended observation period are sequentially placed on the same timeline, and time segmentation points are extracted at the boundaries of each stage. In this way, identity verification, vaccination procedures, and observation are no longer mixed together, but rather form clear time boundaries, providing a basis for subsequently splitting the vaccination execution sequence.

[0120] S305. The vaccination execution sequence is segmented and encapsulated according to the time sequence segmentation point to generate multiple vaccination control nodes, including identity verification node, vaccination observation node and inventory replenishment node.

[0121] Among them, the identity verification node is used to confirm the identity of the vaccination recipient; the vaccination observation node is used to complete the observation period; and the inventory replenishment node is used to process inventory replenishment or inventory switching after vaccination.

[0122] In this embodiment, the vaccination execution sequence is divided according to the time sequence segmentation point, and the originally continuous execution sequence is broken into multiple executable nodes. These nodes are then encapsulated as vaccination control nodes. Each vaccination control node corresponds to a clear business action, and the vaccination unit, observation arrangement and inventory replenishment process can be scheduled according to the node, thereby ensuring clear connection between the vaccination execution process.

[0123] As an example, if a target recipient experiences a mild local reaction after a previous vaccination, such as a mild fever, and needs to complete the next dose at another vaccination unit: First, extract the recipient's historical adverse reaction symptoms and compare the similarity of the components of the previous vaccine with the vaccine to be administered; then, determine the risk level corresponding to the symptom according to the assessment mapping table, and obtain a risk gain coefficient by combining component similarity and age correction coefficient; then, use this coefficient to extend the standard observation time, and mark the three stages of verification, vaccination, and observation on the timeline; finally, divide the vaccination execution sequence into identity verification node, vaccination observation node, and inventory replenishment node.

[0124] This application incorporates the target recipients' past reactions, similarity of vaccine components, and age differences into the observation and adjustment process through S104, enabling the vaccination sequence to be broken down into executable nodes according to actual risks, thereby improving the continuity and controllability of subsequent vaccination arrangements.

[0125] S105. When there are off-site vaccination tasks in the collaborative resource pool, the available inventory is mapped and allocated according to the resource occupancy sequence and the occupancy identifier is updated. The target occupancy identifier and the corresponding vaccination control node are concatenated into a vaccination task sequence according to the vaccination timestamp.

[0126] Among them, available inventory is used to represent the vaccine inventory that can be occupied by the off-site vaccination task; target occupancy identifier is used to represent the occupancy result after the inventory is updated by mapping; vaccination timestamp is used to mark the order in which each node occurs; vaccination task sequence is used to represent the complete task arrangement that the terminal can directly execute.

[0127] In this embodiment of the application, S105 specifically includes:

[0128] S1051. In response to the remote vaccination task issued in the collaborative resource pool, the resource occupancy sequence is parsed to obtain the target demand, and an inventory locking operation is performed in the available inventory of the corresponding vaccination unit to deduct the target demand and obtain the current inventory balance.

[0129] Among them, the target demand quantity is used to indicate the actual number of doses required for the off-site vaccination task; the inventory lock is used to reserve the corresponding number of vaccines in advance to avoid them being used repeatedly by other tasks; the current inventory balance is used to indicate the remaining available inventory after the lock is completed.

[0130] In this embodiment, the off-site vaccination task issued by the collaborative resource pool is first read, and then the target demand corresponding to the task is extracted from the resource occupancy sequence. Next, an inventory locking operation is performed on the available inventory of the corresponding vaccination unit to reserve the vaccine corresponding to the target demand and deduct it from the available inventory, thus obtaining the current inventory balance. This process avoids the same batch of inventory being occupied by multiple tasks simultaneously and provides a clear inventory basis for subsequent traceability code mapping.

[0131] S1052. Read the electronic traceability code of the vaccine assigned to the off-site vaccination task, encrypt and map the current inventory balance and the electronic traceability code based on the preset data encoding rules, output the mapping feature code, and write the mapping feature code into the corresponding field of the initial occupancy identifier to obtain the target occupancy identifier.

[0132] Among them, the electronic traceability code is used to identify the smallest packaging unit of the vaccine; the data coding rules are used to specify the field order, field length, separator mark, random salt value and summary length; the mapping feature code is used to represent the correspondence between the current inventory balance and the electronic traceability code; and the initial occupancy identifier is used to represent the original occupancy status of the inventory before it has been mapped and updated.

[0133] In this embodiment, the electronic traceability code of the vaccine assigned to the off-site vaccination task is first read. Then, the current inventory balance and the electronic traceability code are organized into an coded field string according to a preset data encoding rule. The data encoding rule specifies the order of task number, current inventory balance, electronic traceability code, random salt value, and digest length. Then, hash digest processing can be used to encrypt and map this field string to obtain a mapping feature code. This encryption mapping is not for recovering plaintext, but to fix the correspondence between inventory and traceability code, which is convenient for terminal verification and background tracking. Finally, the mapping feature code is written into the corresponding field of the initial occupancy identifier, and the original field value of the initial occupancy identifier is replaced with the mapping feature code to obtain the target occupancy identifier, so that the occupancy status and vaccine traceability information are consistent.

[0134] S1053. Using the vaccination timestamp as a time reference, the associated identity verification node, vaccination observation node, and inventory replenishment node are connected in sequence according to time to form a process execution link.

[0135] The process execution link is used to connect identity verification, vaccination observation and inventory replenishment in sequence.

[0136] In this embodiment, the vaccination timestamp is used as a unified benchmark. The identity verification node, vaccination observation node and inventory replenishment node are extracted from the corresponding vaccination control node and then connected in sequence according to time to form a process execution link. This process enables the verification, vaccination observation and inventory replenishment to no longer be processed separately, but to be executed sequentially according to the same link.

[0137] S1054. Encapsulate the target occupancy identifier into a scheduling control instruction, and configure the scheduling control instruction in the process execution link to generate an inoculation task sequence that can be executed by the terminal.

[0138] Among them, the scheduling control instructions are used to package the target occupancy identifier and execution order into control content that can be recognized by the terminal.

[0139] In this embodiment, the target occupancy identifier is first encapsulated into a scheduling control instruction, and then the scheduling control instruction is written into the process execution chain. In this way, after receiving the vaccination task sequence, the terminal can perform identity verification, vaccination observation, and inventory replenishment step by step according to the order given in the chain, instead of just seeing a static inventory record.

[0140] This application, through step 105, enables the target recipient to complete subsequent doses at another vaccination unit, unifying remote vaccination tasks, inventory occupancy relationships, and terminal execution links into the same scheduling process. This allows inventory locking, traceability code mapping, and node serialization to be completed continuously, thereby ensuring that the vaccination task sequence has a clear execution order and traceable occupancy basis.

[0141] S106. Obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

[0142] The actual execution information includes the scanning record generated by the vaccination terminal on the vaccination day, the vaccination completion status, the vaccination time and the corresponding electronic traceability code; unverified doses are used to indicate doses that have been scheduled for vaccination but have not been verified in the actual execution information; temperature anomaly nodes are used to indicate temperature control sections where the temperature continuously deviates from the specified range during cold chain storage and transportation; and task compensation queues are used to record the task order that needs to be rearranged for vaccination, inventory replenishment or abnormal handling.

[0143] like Figure 4 As shown, the task compensation ledger 41 represents the entire ledger carrier area enclosed by the outer solid line. The task compensation ledger 41 includes columns for appointment traceability code, actual execution, abnormal temperature, unredeemed doses, compensation status, and file backfilling. These columns together constitute the main table area of ​​the task compensation queue, used to record the traceability code verification results, actual execution results, abnormal temperature conditions, unredeemed dose conditions, compensation processing status, and file backfilling status generated after the vaccination task sequence is executed. The lower part of the task compensation ledger 41 also includes allocation weight records 42, collaborative resource pool records 43, and immunization file backfilling records 44. Allocation weight records 42 are used for resource adjustments in subsequent allocation cycles, collaborative resource pool records 43 correspond to the aforementioned resource matching relationships, and immunization file backfilling records 44 correspond to the backfilling results of vaccination completion status and adverse reaction data.

[0144] In this embodiment of the application, S106 specifically includes:

[0145] S1061. After the designated vaccination date, collect the electronic traceability code that has been scanned from the terminal device as the actual execution information.

[0146] The terminal equipment includes vaccination registration equipment, mobile scanning equipment, or cold chain handover scanning equipment used by vaccination units for scanning and verifying vaccines; the electronic traceability code is used to identify the smallest packaging unit of the vaccine and to match the vaccination task with the actual vaccine used.

[0147] In this embodiment, after the vaccination day ends, the electronic traceability code that has been scanned is read from the terminal device, and the scanning time, vaccination unit number and vaccination completion status are read simultaneously; the collected data are collected according to the task number in the vaccination task sequence, and the scanning results that can be matched with specific vaccination tasks are organized into actual execution information; if the same electronic traceability code is scanned repeatedly, the first valid scanning record is retained, and subsequent repeated records are saved as verification information.

[0148] S1062. Compare the appointment traceability code contained in the vaccination task sequence with the electronic traceability code in the actual execution information, and mark the dose corresponding to the unmatched electronic traceability code as an unredeemed dose.

[0149] The appointment traceability code is used to represent the vaccine traceability code pre-assigned to the target recipient in the vaccination task sequence; unsuccessful matching means that the appointment traceability code cannot find the corresponding electronic traceability code in the actual execution information, or the found electronic traceability code is inconsistent with the task recipient, vaccination unit, and vaccination time.

[0150] In this embodiment, the appointment traceability code is used as the verification benchmark. The corresponding electronic traceability code is searched for item by item in the actual execution information. If the electronic traceability code exists and the vaccination recipient, vaccination unit, and vaccination time all conform to the vaccination task sequence, the dose is recorded as cancelled. If no corresponding electronic traceability code is found, or the scan record is inconsistent with the vaccination task sequence, the dose corresponding to the electronic traceability code is marked as an uncancelled dose. Through this process, the collected actual execution information is converted into incomplete records that can be used for compensation arrangements.

[0151] S1063. Real-time acquisition of ambient temperature data of each temperature control node during cold chain storage and transportation. When the duration of the ambient temperature data deviating from the preset temperature range exceeds the preset duration threshold, the temperature control segment where the ambient temperature data deviates is marked as a temperature abnormal node.

[0152] Among them, temperature control nodes include outbound temporary storage points, transport vehicles, refrigerated boxes, in-transit handover points, and refrigeration equipment at vaccination units; ambient temperature data is used to represent the temperature records collected by temperature control nodes during cold chain storage and transportation; preset temperature ranges and preset duration thresholds can be set according to vaccine storage and transportation management requirements, vaccine instructions, and cold chain management rules of vaccination units.

[0153] In this embodiment, the ambient temperature data of each temperature control node is read in chronological order of acquisition time, and the ambient temperature data is continuously checked against the preset temperature range: when a single deviation recovers quickly, it is only recorded as a temperature fluctuation; when the duration of the ambient temperature data deviating from the preset temperature range exceeds a preset duration threshold, it is determined that the temperature control segment no longer belongs to ordinary fluctuation, and the corresponding temperature control segment is marked as a temperature anomaly node. After this processing, the temperature anomaly node can be included in the subsequent compensation judgment along with the unrecovered doses.

[0154] S1064. Extract the associated business attributes of the unredeemed doses and the temperature anomaly node, calculate the compensation priority of the unredeemed doses and the temperature anomaly node using a preset weighted scoring model, and write them into the compensation task table in descending order of priority to generate a task compensation queue.

[0155] Among them, the associated business attributes include the vaccine batch to which the dose belongs, the planned vaccination time of the target recipient, the vaccination unit, the inventory status, the location of the temperature anomaly node and the duration of the anomaly; the compensation priority is used to indicate the order in which different compensation tasks are processed; the compensation task table is used to store the compensation recipient, the reason for compensation, the processing order and the corresponding vaccination unit.

[0156] In this embodiment, the preset weighted scoring model can be configured using historical compensation records, cold chain exception handling rules, and vaccination business rules. For example, the expression for the preset weighted scoring model is formula (7):

[0157] (7)

[0158] in, Priority for compensation; The score is based on the timing of the vaccination schedule; The score represents the dose sequence. The score represents the inventory occupancy status; The score represents the range of influence of temperature control. The score represents the capacity of the vaccination unit to handle the workload; , , , and These are the corresponding weights, and they are dimensionless proportionality coefficients. .

[0159] As a specific implementation, if a certain unrecovered dose is simultaneously associated with a temperature anomaly node, the vaccination schedule time, dose sequence, inventory status, temperature control impact range, and vaccination unit capacity are converted into dimensionless rank scores, with values ​​of 4, 3, 5, 4, and 2 respectively; the corresponding weights are 0.25, 0.20, 0.25, 0.20, and 0.10 respectively. The compensation priority is then determined. =0.25×4+0.20×3+0.25×5+0.20×4+0.10×2=3.85, and write them into the compensation task table according to the compensation priority from high to low, forming a task compensation queue. The above embodiment is only used to illustrate the method of determining the compensation priority of each business item score and corresponding weight. The listed scores and weights are not limited. In actual application, they can be set according to the vaccination task status, cold chain abnormality handling rules and the acceptance capacity of the vaccination unit. This application does not limit them in this regard.

[0160] This application, through step 106, can uniformly incorporate the results of vaccination scanning and verification and the results of abnormal cold chain temperature into the task compensation queue, so that incomplete doses, abnormal batches and subsequent compensation arrangements can form a continuous closed loop.

[0161] After generating the task compensation queue, the method further includes:

[0162] B1. Based on the number of records in the task compensation queue, adjust the allocation weight of the corresponding vaccination unit in the next allocation cycle, and iteratively update the subsequent vaccination plan based on the allocation weight.

[0163] The allocation cycle is used to represent the time period between one resource allocation and the next resource allocation.

[0164] In this embodiment, the number of records in the task compensation queue is first counted, and the compensation records are summarized separately for each vaccination unit. If a vaccination unit has a large number of unredeemed doses or a large number of tasks that need to be rescheduled due to temperature abnormalities, the allocation weight of that vaccination unit in the next allocation cycle is increased, so that subsequent resources are given priority to that vaccination unit. If a vaccination unit has few compensation records, the original allocation weight is maintained or only slightly adjusted. Then, the updated allocation weight is written into the subsequent vaccination plan so that the next round of scheduling can continue the compensation results of the previous round.

[0165] B2. When the number of records is zero or the difference between the assigned weights in two adjacent iterations is less than a preset convergence threshold, the iterative update is stopped.

[0166] The preset convergence threshold is used to represent the minimum weight change that can be actually executed by the inoculated unit when the weight allocation continues to be adjusted. For example, it can be set to 0.10, which means that the iteration update stops when the weight allocation difference between two adjacent iterations is less than 0.10.

[0167] In this embodiment, the task compensation queue is rechecked after each update: if the number of records in the queue is zero, it means that the items requiring compensation in the current round have been processed, and further adjustments are stopped; if there are still records in the queue, but the change in allocation weight between the current round and the previous round is less than a preset convergence threshold, it means that further adjustments will not have a significant impact on subsequent arrangements, and iterative updates are also stopped. This avoids repeated fluctuations in the plan and also avoids the additional processing costs caused by repeated adjustments.

[0168] B3. The vaccination completion status and adverse reaction data monitored after vaccination in the actual execution information are synchronously filled back into the immunization record of the target vaccination subject.

[0169] Among them, the adverse reaction data monitored after vaccination are used to record abnormal reactions that occur during the post-vaccination observation, follow-up or reporting process.

[0170] In this embodiment, the vaccination completion status from the actual execution information is written back to the target recipient's immunization record, and adverse reaction data monitored after vaccination are simultaneously added to the corresponding fields. This way, when the immunization record is accessed again later, the number of doses completed by the target recipient, the post-vaccination reaction status, and whether special observation is required can be directly read, thus providing a complete basis for the next vaccination schedule.

[0171] This application, through B1-B3, enables a vaccination unit to reverse the allocation of resources after unredeemed doses and reassignment of tasks in the previous round, and to synchronously record the actual vaccination results in the immunization record, thereby ensuring the continuity and consistency between the vaccination plan, execution results and subsequent arrangements.

[0172] This application, through S101-S106, can integrate multi-source information such as immunization records, vaccination plans, inventory, road network access, and cold chain temperature into the same collaborative process. Under the premise of protecting the privacy of vaccination information in different locations, it can complete demand identification, inventory locking, task scheduling, observation segmentation, anomaly compensation, and record backfilling, forming a traceable, replenishable, and closed-loop vaccine supply chain collaborative optimization mechanism. This improves the efficiency of supply and demand matching and the continuity of vaccination execution, and reduces the risks of duplicate inventory, temperature control deviation, and unverified inventory.

[0173] Figure 5 This is a schematic diagram of the structure of a multi-source data fusion vaccine supply chain collaborative optimization system disclosed in this application, with reference to... Figure 5 The system may include:

[0174] Extraction module 51 is used to obtain the immunization records and vaccination plans of the target vaccination subjects, extract the vaccination trajectory corresponding to the target vaccination subjects based on the historical vaccination information in the immunization records, and match the vaccination trajectory with the vaccination plan in time to generate the demand queue of each vaccination unit;

[0175] The calculation module 52 is used to calculate the vaccine dose shortage and available inventory of each vaccination unit based on the demand queue, perform data anonymization processing on the off-site vaccination details in the immunization file to obtain anonymized records, and calculate the travel time between each vaccination unit within a preset reference framework, and generate a collaborative resource pool based on the anonymized records and the travel time.

[0176] The organization module 53 is used to generate a resource occupancy sequence in the collaborative resource pool according to the order of vaccination days, and to organize the doses that the same target vaccination object receives consecutively in different vaccination units into a vaccination execution sequence in chronological order.

[0177] The splitting module 54 is used to extend the vaccination observation time based on the adverse reaction symptoms of the target vaccination subjects in the immunization record, and to split the vaccination execution sequence into multiple vaccination control nodes with the vaccination window and the extended vaccination observation time as the boundary.

[0178] The splicing module 55 is used to map and allocate the available inventory according to the resource occupancy sequence and update the occupancy identifier when there is a remote vaccination task in the collaborative resource pool, and splice the target occupancy identifier and the corresponding vaccination control node according to the vaccination timestamp to form a vaccination task sequence.

[0179] The generation module 56 is used to obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

[0180] The multi-source data fusion vaccine supply chain collaborative optimization system of this application embodiment is used to implement the aforementioned multi-source data fusion vaccine supply chain collaborative optimization method. Therefore, the specific implementation of the multi-source data fusion vaccine supply chain collaborative optimization system can be found in the embodiment section of the multi-source data fusion vaccine supply chain collaborative optimization method above. The specific implementation can be referred to the description of the corresponding embodiment, and will not be repeated here.

[0181] The foregoing has provided a detailed description of a multi-source data fusion-based vaccine supply chain collaborative optimization method and system provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for collaborative optimization of the vaccine supply chain through multi-source data fusion, characterized in that, include: Obtain the immunization records and vaccination plans of the target vaccination subjects; extract the vaccination trajectory corresponding to the target vaccination subjects based on the historical vaccination information in the immunization records; and match the vaccination trajectory with the vaccination plan in time to generate a demand queue for each vaccination unit. Based on the demand queue, the vaccine dose shortage and available inventory of each vaccination unit are calculated. The off-site vaccination details in the immunization record are anonymized to obtain anonymized records. At the same time, the travel time between each vaccination unit is calculated within a preset reference framework. A collaborative resource pool is generated based on the anonymized records and the travel time. Within the collaborative resource pool, a resource occupancy sequence is generated according to the order of vaccination dates, and the doses continuously received by the same target vaccination subject in different vaccination units are organized into a vaccination execution sequence according to time. Based on the adverse reaction symptoms of the target subjects in the immunization records, the observation time after vaccination is extended. The vaccination execution sequence is divided into multiple vaccination control nodes with the vaccination window and the extended observation time as the boundary. When there are off-site vaccination tasks in the collaborative resource pool, the available inventory is mapped and allocated according to the resource occupancy sequence and the occupancy identifier is updated. The target occupancy identifier and the corresponding vaccination control node are concatenated according to the vaccination timestamp to form a vaccination task sequence. Obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

2. The method according to claim 1, characterized in that, The process involves desensitizing the off-site vaccination details in the immunization records to obtain desensitized records, calculating the travel time between vaccination units within a preset reference framework, and generating a collaborative resource pool based on the desensitized records and the travel time, including: Extract the vaccination frequency and distance from the vaccination details in the different locations, dynamically adjust the privacy budget based on the vaccination distance, and inject random perturbation values ​​that satisfy a preset probability distribution into the vaccination frequency to generate desensitized records; Within a pre-defined reference framework serving as a spatial measurement benchmark, the actual trajectory distance of each inoculation unit in the road network topology is analyzed, and the travel time is determined based on the actual trajectory distance. Based on the desensitized records, target recipients with a tendency to receive vaccinations in other locations are screened out. The target vaccine batches required by the screened target recipients are determined, the cold chain timeliness of the target vaccine batches is extracted, and vaccination units with a passage time shorter than the cold chain timeliness are matched and associated with resources to build a collaborative resource pool.

3. The method according to claim 1, characterized in that, The method of extending the observation period after vaccination by incorporating adverse reaction symptoms of the target subjects in the immunization record, and using the vaccination window and the extended observation period as boundaries, divides the vaccination execution sequence into multiple vaccination control nodes, including: Extract adverse reaction symptoms from the historical records of the target vaccination subjects and analyze the component similarity between the historical vaccines and the vaccines to be administered; The risk severity of the adverse reaction symptoms is quantified according to a preset assessment mapping table, and the risk severity is weighted and corrected using the component similarity to obtain a risk gain coefficient. The extended vaccination observation time is calculated by multiplying the vaccination observation time by the risk gain coefficient. The verification window, vaccination window, and extended vaccination observation time are sequentially marked on the timeline to extract the time sequence segmentation points; The vaccination execution sequence is segmented and encapsulated according to the time sequence segmentation points to generate multiple vaccination control nodes, including identity verification nodes, vaccination observation nodes, and inventory replenishment nodes.

4. The method according to claim 3, characterized in that, The step of quantifying the risk severity of the adverse reaction symptoms according to a preset assessment mapping table, and using the component similarity to weight and correct the risk severity to obtain a risk gain coefficient, includes: Retrieve the risk weight value corresponding to the adverse reaction symptoms from the preset assessment mapping table, and confirm the risk weight value as the quantified risk severity; Extract the antigen and excipient components of the historically administered vaccines and the vaccines to be administered, calculate the component differences between the historically administered vaccines and the vaccines to be administered in a multidimensional feature space, and determine the component similarity based on the component differences; The risk severity is calculated by multiplying the component similarity to obtain a weighted correction value, and the weighted correction value is input into a preset nonlinear function for mapping operation to output a baseline risk value; The risk gain coefficient is obtained by matching the target vaccination subject's physiological age with a preset age correction coefficient and adding the baseline risk value to the age correction coefficient.

5. The method according to claim 3, characterized in that, When there are off-site vaccination tasks in the collaborative resource pool, the available inventory is mapped and allocated according to the resource occupancy sequence, and the occupancy identifier is updated. The target occupancy identifier and the corresponding vaccination control node are concatenated according to the vaccination timestamp to form a vaccination task sequence, including: In response to the remote vaccination task issued in the collaborative resource pool, the resource occupancy sequence is parsed to obtain the target demand, and an inventory locking operation is performed in the available inventory of the corresponding vaccination unit to deduct the target demand and obtain the current inventory balance. Read the electronic traceability code of the vaccine assigned to the off-site vaccination task, encrypt and map the current inventory balance and the electronic traceability code based on the preset data encoding rules, output the mapping feature code, and write the mapping feature code into the corresponding field of the initial occupancy identifier to obtain the target occupancy identifier; Using the vaccination timestamp as a time reference, the associated identity verification node, vaccination observation node, and inventory replenishment node are linked together in chronological order to form a process execution link; The target occupancy identifier is encapsulated into a scheduling control instruction, and the scheduling control instruction is configured in the process execution link to generate an inoculation task sequence that can be executed by the terminal.

6. The method according to claim 1, characterized in that, The method of organizing the consecutive doses received by the same target vaccine recipient across different vaccination units into a vaccination execution sequence according to time order includes: Retrieve multiple vaccination details that the same target vaccination subject needs to complete within a preset time period, and identify the vaccination unit to which each vaccination detail belongs; The spatial transfer time distributed between different vaccination units is extracted from the preset road network topology. The spatial transfer time is inserted between the planned vaccination times of the vaccination details to perform timing verification. The vaccination details that pass the timing verification are arranged in the order of vaccination time to generate the vaccination execution timing.

7. The method according to claim 1, characterized in that, The process of extracting the vaccination trajectory corresponding to the target vaccination subject based on the historical vaccination information in the immunization record, and matching the vaccination trajectory with the vaccination plan in time to generate a demand queue for each vaccination unit includes: The historical vaccination time and the standard interval between adjacent doses in the historical vaccination information are analyzed. Based on the historical vaccination time and the standard interval time, the vaccination time range of the target doses that have not yet been administered is calculated to form a vaccination trajectory with time constraints. Calculate the overlap between the vaccination time range and the target vaccination window arranged in the vaccination plan, and extract the target number of vaccination doses within the overlap interval; Based on the geographical coordinates of each vaccination unit and the spatial location of the target vaccination recipient, the selected target vaccination doses are associated and allocated to the corresponding vaccination units to generate a demand queue.

8. The method according to claim 1, characterized in that, The step of obtaining the actual execution information of the vaccination task sequence, identifying unredeemed doses and abnormal temperature nodes, and generating a task compensation queue includes: After the designated vaccination date, the electronic traceability code that was actually scanned is collected from the terminal device as actual execution information; By comparing the appointment traceability code contained in the vaccination task sequence with the electronic traceability code in the actual execution information, the doses corresponding to the unmatched electronic traceability codes are marked as unredeemed doses. Real-time acquisition of ambient temperature data at each temperature control node during cold chain storage and transportation; when the duration of the ambient temperature data deviating from the preset temperature range exceeds a preset duration threshold, the temperature control segment where the ambient temperature data deviates is marked as a temperature abnormal node. Extract the associated business attributes of the unredeemed doses and the temperature anomaly nodes, calculate the compensation priority of the unredeemed doses and the temperature anomaly nodes using a preset weighted scoring model, and write them into the compensation task table in descending order of priority to generate a task compensation queue.

9. The method according to claim 1, characterized in that, After generating the task compensation queue, the method further includes: Based on the number of records in the task compensation queue, the allocation weight of the corresponding vaccination unit in the next allocation cycle is adjusted, and the subsequent vaccination plan is iteratively updated based on the allocation weight. The iterative update stops when the number of records is zero or the difference between the assigned weights in two adjacent iterations is less than a preset convergence threshold. The vaccination completion status and adverse reaction data monitored after vaccination in the actual execution information will be synchronously filled back into the immunization record of the target vaccination subject.

10. A multi-source data fusion-based vaccine supply chain collaborative optimization system, characterized in that, include: The extraction module is used to obtain the immunization records and vaccination plans of the target vaccination subjects, extract the vaccination trajectory corresponding to the target vaccination subjects based on the historical vaccination information in the immunization records, and match the vaccination trajectory with the vaccination plan in time to generate a demand queue for each vaccination unit. The calculation module is used to calculate the vaccine dose shortage and available inventory of each vaccination unit based on the demand queue, perform data anonymization processing on the off-site vaccination details in the immunization file to obtain anonymized records, and calculate the travel time between each vaccination unit within a preset reference framework, and generate a collaborative resource pool based on the anonymized records and the travel time. The organization module is used to generate a resource occupancy sequence according to the order of vaccination days within the collaborative resource pool, and to organize the doses that the same target vaccination object receives consecutively in different vaccination units into a vaccination execution sequence in chronological order. The splitting module is used to extend the observation time after vaccination by combining the adverse reaction symptoms of the target vaccination subjects in the immunization record. The vaccination execution sequence is split into multiple vaccination control nodes with the vaccination window and the extended observation time as the boundary. The splicing module is used to map and allocate the available inventory according to the resource occupancy sequence and update the occupancy identifier when there are off-site vaccination tasks in the collaborative resource pool, and splice the target occupancy identifier and the corresponding vaccination control node according to the vaccination timestamp to form a vaccination task sequence. The generation module is used to obtain the actual execution information of the vaccination task sequence, identify unrecovered doses and abnormal temperature nodes, and generate a task compensation queue.

Images