Vaccination service reservation optimization system based on resource scheduling and path planning

Abstract

The invention relates to the technical field of resource scheduling and path planning, and particularly discloses a vaccination service reservation optimization system based on resource scheduling and path planning. The system comprises a user service interface module, a resource state sensing module, a dynamic coupling analysis module, a multi-objective optimization decision module and an execution control module. By constructing a complete technical chain of resource state perception, dynamic coupling analysis, multi-objective optimization decision and execution control, deep collaboration of resource scheduling and path planning in the vaccination service is realized, the system can accurately predict the vaccine inventory turnover risk, and the problem of vaccine waste caused by appointment breaking of a user is effectively avoided; and an optimal solution is sought under multiple constraint conditions through a multi-objective optimization algorithm, so that the inoculation service efficiency and the resource utilization rate are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of resource scheduling and path planning technology, specifically relating to a vaccine appointment optimization system based on resource scheduling and path planning. Background Technology

[0002] Medical resource allocation and optimization is a key research direction in the field of smart healthcare, with its core objective being to improve the efficiency of medical services and resource utilization through information technology. Vaccination services, as an important component of the public health system, involve complex resource coordination and allocation issues, urgently requiring efficient allocation and optimization methods.

[0003] The vaccine appointment optimization system is a key technology for improving vaccination efficiency and enhancing user experience. The system aims to achieve a dynamic balance between vaccination sites, vaccine inventory, and user demand through reasonable resource allocation and route planning, thereby ensuring the orderly conduct of vaccination services.

[0004] Existing technologies typically employ separate inventory management and appointment scheduling modules, failing to adequately consider the dynamic coupling relationships in scenarios involving multiple vaccines. The inventory system is not linked to user appointment behavior in real time, making it impossible to accurately predict the risk of vaccines nearing their expiration date due to user no-shows, resulting in low inventory turnover and resource waste.

[0005] Meanwhile, traditional route planning algorithms do not consider real-time load at vaccination sites and vaccine inventory status, making it difficult to achieve coordinated optimization of vaccination efficiency and resource utilization under multiple constraints. Furthermore, existing systems lack the ability to integrate and analyze data across the entire "appointment-inventory-route" chain, failing to effectively cope with sudden fluctuations in demand and resource conflicts, thus limiting overall service efficiency. Therefore, there is an urgent need to develop a vaccine appointment optimization system that can deeply integrate resource scheduling and route planning. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of existing vaccine appointment systems, such as the separation of resource scheduling and path planning, the inability to effectively deal with the dynamic coupling relationship of multiple vaccine mixed scenarios, and the lack of ability to integrate and analyze data across the entire chain. The present invention provides a vaccine appointment optimization system based on resource scheduling and path planning that can achieve dynamic balance between vaccination sites, vaccine inventory, and user demand.

[0007] To achieve the above objectives, the present invention provides a vaccine appointment optimization system based on resource scheduling and path planning, the system comprising:

[0008] It includes a user service interface module, a resource status awareness module, a dynamic coupling analysis module, a multi-objective optimization decision-making module, and an execution control module. The user service interface module receives vaccination appointment requests submitted by users, which include the vaccine type, desired vaccination time window, and user's geographical location information. The resource status awareness module collects and updates real-time data on vaccine inventory at each vaccination site, vaccination station operating status, medical staff scheduling information, and real-time vaccination load data.

[0009] The dynamic coupling analysis module, based on the output of the resource status perception module and the appointment request data from the user service interface module, establishes a vaccine inventory turnover risk prediction model and a vaccination site service capacity assessment model. The multi-objective optimization decision module receives the output of the dynamic coupling analysis module and generates the optimal appointment allocation scheme and vaccination route planning scheme by solving the joint optimization problem of resource scheduling and route planning. The execution control module decomposes the optimal scheme generated by the multi-objective optimization decision module into specific appointment confirmation instructions, inventory allocation instructions, and route navigation instructions, and distributes them to the corresponding terminal devices.

[0010] Furthermore, the dynamic coupling analysis module includes an inventory turnover risk prediction unit and a service capacity assessment unit. The inventory turnover risk prediction unit calculates the near-expiration risk coefficient for various vaccines by analyzing the matching degree between historical no-show data and current appointment data. This unit uses time series analysis to fit the trend of appointment data and constructs a risk warning model by combining it with vaccine expiration date data. The service capacity assessment unit calculates the service capacity margin of each vaccination site within a specified time interval based on the working status of vaccination stations, medical staff configuration, and real-time load data. This unit quantifies the service pressure distribution of vaccination sites by establishing a load balancing index.

[0011] Furthermore, the multi-objective optimization decision-making module comprises an objective function construction unit and a constraint processing unit. The objective function construction unit defines three core objectives for system optimization: minimizing average user waiting time, maximizing vaccine inventory turnover rate, and minimizing the load imbalance at vaccination sites. The constraint processing unit sets four types of hard constraints: upper limit constraint on vaccination site service capacity, vaccine inventory availability constraint, user time window satisfaction constraint, and path reachability constraint. This module employs an improved genetic algorithm to solve the multi-objective optimization problem, enhancing search efficiency by introducing an elite retention strategy and an adaptive crossover / mutation operator.

[0012] Furthermore, the solution process of the multi-objective optimization decision module includes the following steps: First, initialize the population, with each individual representing a complete reservation allocation and path planning scheme; then calculate the fitness value of each individual, with the fitness function obtained by weighted summation of the objective function; next, perform a selection operation, employing a roulette wheel selection mechanism combined with an elite retention strategy; subsequently, perform a crossover operation, crossing over in both the reservation allocation scheme and the path planning scheme dimensions; finally, perform a mutation operation, locally adjusting the inoculation point allocation or path order in the scheme with a preset probability. This process is iteratively executed until the convergence condition is met or the maximum number of iterations is reached.

[0013] Furthermore, the execution control module includes an instruction generation unit and a distribution and scheduling unit. The instruction generation unit parses the optimized solution into three types of executable instructions: an appointment confirmation instruction containing the specific location of the vaccination site, the vaccination time, and the appointment code information; an inventory allocation instruction specifying the type and quantity of vaccines to be allocated and the target vaccination site; and a route navigation instruction generating the optimal driving route from the user's current location to the designated vaccination site. The distribution and scheduling unit uses a message queue mechanism to ensure reliable transmission of instructions and prioritizes emergency instructions for timely processing.

[0014] Furthermore, the system includes a feedback learning module, which collects deviation information between actual execution data and the expected plan, including differences between actual vaccination time and appointment time, actual travel time and estimated time, and actual vaccine usage and estimated usage. The feedback learning module uses this deviation data to perform online calibration of the parameters of the dynamic coupling analysis model and the multi-objective optimization model, continuously optimizing the model's prediction accuracy through a gradient descent algorithm.

[0015] Furthermore, the user service interface module supports multiple access methods, including mobile applications, web pages, and self-service terminals. This module provides appointment inquiry, appointment modification, and appointment cancellation functions, with all operation changes synchronized in real time to the resource status awareness module and the dynamic coupling analysis module. The user service interface module also integrates electronic payment functionality, supporting online payment of vaccination fees.

[0016] Furthermore, the resource status perception module collects real-time data from vaccination sites via IoT devices, including cold chain temperature and humidity sensors, vaccination station status monitoring equipment, and personnel positioning devices. This module establishes a data quality assessment mechanism to clean and repair abnormal data, ensuring the accuracy and reliability of the input data.

[0017] Furthermore, the system adopts a distributed architecture, with modules exchanging data through well-defined interface protocols. A unified time synchronization mechanism ensures consistency in data processing across modules. The system also features comprehensive logging and audit trail functionality to meet the compliance requirements of healthcare services.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0019] 1. This invention achieves deep collaboration between resource scheduling and path planning in vaccine vaccination services by constructing a complete technical chain of resource status perception, dynamic coupling analysis, multi-objective optimization decision-making and execution control. The system can accurately predict vaccine inventory turnover risks and effectively avoid vaccine waste caused by users not showing up.

[0020] 2. By using a multi-objective optimization algorithm to find the optimal solution under multiple constraints, the efficiency of vaccination services and resource utilization have been significantly improved.

[0021] 3. The system's feedback learning mechanism continuously optimizes the decision-making model, enhancing its adaptability to sudden fluctuations in demand. The overall solution addresses the shortcomings of existing technologies where resource scheduling and path planning are disconnected, providing optimized support for the entire vaccination service chain. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall technical architecture of the vaccine appointment optimization system based on resource scheduling and path planning proposed in this invention;

[0023] Figure 2 This is a schematic diagram of the core principle framework of the dynamic coupling analysis module in this invention;

[0024] Figure 3 This is a logical flowchart of the multi-objective optimization decision-making module in this invention;

[0025] Figure 4 This is a schematic diagram of the instruction generation and distribution scheduling framework of the execution control module in this invention;

[0026] Figure 5 This is a schematic diagram illustrating the online correction and model optimization principle of the feedback learning module in this invention. Detailed Implementation

[0027] Please refer to the attached document. Figures 1 to 5 This embodiment details the technical implementation of a vaccine appointment optimization system based on resource scheduling and path planning. The system employs a distributed architecture, with modules exchanging data through well-defined interface protocols and a unified time synchronization mechanism to ensure time consistency in data processing across the entire system. The core components of the system include a user service interface module, a resource status awareness module, a dynamic coupling analysis module, a multi-objective optimization decision-making module, an execution control module, and a feedback learning module. These modules work collaboratively to form a complete closed loop from user demand awareness to instruction execution and feedback optimization.

[0028] The user service interface module is the primary portal for system interaction with external users, supporting multiple access methods including mobile applications, web pages, and self-service terminals. Deployed behind a load balancer, this module receives and parses user requests through the application gateway. When a user initiates a vaccination appointment request, the request data packet is encrypted using a secure transport layer protocol before being sent to the interface module. The data packet contains three key fields: an enumeration value for the vaccine type, the start and end timestamps of the desired vaccination time window, and the user's geographical location (latitude and longitude coordinates).

[0029] The user service interface module has a built-in request verification subunit. This subunit performs format compliance checks and logical rationality verification on the received data. For example, it verifies whether the time window is within a reasonable future range and whether the geographical coordinates are within the service coverage area. After successful verification, the request data is converted into an internally unified data format and a timestamp and a unique session identifier are added. Subsequently, the data is asynchronously pushed to the message middleware for consumption by the resource status awareness module and the dynamic coupling analysis module.

[0030] The user service interface module also provides appointment inquiry, appointment modification, and appointment cancellation functions. Any operational changes will generate corresponding event messages, which are broadcast to all relevant modules in the system in real time via a publish-subscribe model to ensure status synchronization. In addition, this module integrates an electronic payment subunit, interfaces with third-party payment platforms, and supports various online payment methods for vaccination fees. Payment status is closely linked to appointment status; successful payment is one of the necessary prerequisites for appointment confirmation.

[0031] The resource status awareness module is responsible for collecting, cleaning, and aggregating the real-time status of all inoculation points across the network. Please refer to the appendix. Figure 1 This module acquires raw data streams through a cluster of IoT devices deployed at various vaccination sites. These IoT devices include, but are not limited to, cold chain temperature and humidity sensors, vaccination station status monitoring equipment, and location beacons for healthcare personnel. The cold chain temperature and humidity sensors collect real-time temperature and humidity data from vaccine storage refrigerators once per minute, with a data accuracy of ±0.5 degrees Celsius for temperature and ±3% for humidity.

[0032] The vaccination station status monitoring equipment uses infrared and pressure sensors to determine whether the station is idle, in operation, or being cleaned, and reports status changes immediately. Medical staff location beacons, based on Bluetooth Low Energy or UWB technology, track the location of medical staff within the vaccination site in real time. Combined with scheduling system data, the availability of medical staff can be inferred. All this device data is aggregated through an IoT gateway, undergoes preliminary filtering and outlier detection by edge computing nodes, and is then transmitted to the central data platform via a dedicated network. The resource status awareness module includes a data quality assessment and repair subunit.

[0033] This sub-unit maintains a data quality rule base. For example, temperature data is marked as abnormal if it exceeds a preset threshold for three consecutive periods; a status confirmation request is triggered if the status of a vaccination station remains unchanged for more than four hours. For abnormal data, the sub-unit first attempts to repair it by interpolation based on historical data. If the repair fails, the data point is discarded and logged. The cleaned and reliable data is updated to the central status database. The core tables in the database include a vaccine inventory table, a vaccination station status table, a medical staff schedule table, and a real-time load statistics table. The vaccine inventory table records the current inventory quantity, batch number, and expiration date of various vaccines at each vaccination site. The vaccination station status table records the real-time working status of each vaccination station and its most recent status update timestamp. The medical staff schedule table records the shift information and professional skill tags of each medical staff member. The real-time load statistics table dynamically calculates the estimated load of each vaccination site in different future time segments by aggregating the number of currently ongoing vaccination services with the number of confirmed future appointments.

[0034] The dynamic coupling analysis module is one of the intelligent cores of the system; please refer to the appendix for its core principle framework. Figure 2 This module continuously monitors the appointment request stream from the user service interface module and the status update stream from the resource status awareness module. Internally, the module contains two core analysis units: an inventory turnover risk prediction unit and a service capacity assessment unit. The inventory turnover risk prediction unit focuses on predicting the risk of near-expiration and spoilage of various vaccines at different vaccination sites due to factors such as no-shows.

[0035] This unit first retrieves all current appointment records from the resource status awareness module and performs matching analysis against user no-show data from the past 365 days stored in the historical database. The analysis process includes calculating the historical no-show rate for similar vaccines within similar time periods and user groups. An internal time series analysis model is established to fit the current appointment data to predict vaccine demand at specific future points in time. Simultaneously, the unit accesses vaccine inventory data, particularly expiration date information. Based on predicted demand and current inventory, combined with the expiration date, a near-expiration risk coefficient is calculated. This coefficient calculation involves a core algorithm principle: quantifying risk by comparing the inventory depletion rate with the time decay rate. Specifically, the risk is directly proportional to the current inventory level and inversely proportional to the product of the predicted depletion rate and the remaining expiration date.

[0036] The service capacity assessment unit focuses on evaluating the future service supply capacity of each vaccination site. Inputs to this unit include the operational status of vaccination stations, staff configuration and skills, and real-time load data. Internally, the unit calculates the service capacity margin for each vaccination site within a specified time interval. Service capacity margin is defined as the maximum number of appointments that a vaccination site can serve within a specific time period minus the sum of currently booked appointments and the estimated number of services in progress. Furthermore, this unit quantifies the uniformity of pressure distribution across the entire vaccination service network by establishing a load balancing index. The load balancing index is calculated based on the ratio of the standard deviation to the average real-time load of all vaccination sites. The output of the dynamic coupling analysis module is a series of quantitative evaluation indicators and early warning signals. These results are encapsulated into a standard data structure for use by the downstream multi-objective optimization decision-making module.

[0037] The multi-objective optimization decision-making module receives the evaluation results from the dynamic coupling analysis module and solves a complex joint optimization problem involving resource scheduling and path planning. Please refer to the appendix for the logical flow framework of this module. Figure 3 The module contains an objective function construction unit and a constraint processing unit. The objective function construction unit defines the three core objectives that the system needs to optimize simultaneously.

[0038] The first objective is to minimize the average user wait time, which is the time taken from when a user arrives at the vaccination site to when the entire process of vaccination is completed.

[0039] The second objective is to maximize vaccine inventory turnover, which is the ratio of vaccine usage to average inventory levels per unit of time.

[0040] The third objective is to minimize the load imbalance at the inoculation points, i.e., the load balance index mentioned above. The constraint processing unit sets four types of hard constraints that must be met.

[0041] The first type is the upper limit constraint on the service capacity of vaccination sites, which means that the number of appointments allocated to any vaccination site cannot exceed its service capacity margin for that time period.

[0042] The second type is vaccine inventory availability constraints, which means that the type of vaccine to be allocated must have sufficient inventory at the designated vaccination site and be within its validity period.

[0043] The third type is the user time window satisfaction constraint, which means that the vaccination time allocated to the user must fall within their desired time window, or within an acceptable adjustment range confirmed by the user. The fourth type is the path accessibility constraint, which means that the path from the user's current location to the assigned vaccination point must be accessible in the actual transportation network, and the estimated travel time must be within a reasonable range.

[0044] This module employs an improved genetic algorithm to solve the aforementioned multi-objective optimization problem. The solution process begins with the initialization phase, where the algorithm generates an initial population containing hundreds of individuals. Each individual encodes a complete system solution, which comprises two main parts: a specific appointment allocation scheme and an associated path planning scheme. The appointment allocation scheme specifies the exact vaccination point and its timestamp for each user's appointment request. The path planning scheme generates a recommended route sequence for each user from their submitted starting point to the assigned vaccination point.

[0045] Next, the fitness assessment phase begins, where a fitness value is calculated for each individual in the population. The fitness function is obtained by weighted summation of the three objective functions mentioned above after normalization, and the weighting coefficients can be dynamically adjusted according to the management strategy. Following this, a selection operation is performed using a roulette wheel selection mechanism, combined with an elite preservation strategy: the top 10% of individuals with the highest fitness in each generation are directly replicated to the next generation to avoid the loss of superior genes. Then, a crossover operation is performed, with crossover occurring separately along two dimensions: a pre-assignment scheme and a path planning scheme.

[0046] For the reservation allocation scheme dimension, two parent individuals are randomly selected, and crossover points are randomly chosen on their reservation allocation lists to exchange vaccination point allocation information for some users. For the path planning scheme dimension, sequential crossover is performed on the path sequence. The probability of the crossover operation is controlled by an adaptive operator, which dynamically adjusts the crossover rate based on changes in population diversity. Finally, a mutation operation is performed to locally perturb the scheme with a preset small probability. Mutation may manifest as randomly changing the vaccination point of a user or adjusting the order of some path nodes in the path planning. The entire selection, crossover, and mutation process constitutes one iteration. The iteration process is executed repeatedly until the convergence condition is met, i.e., the improvement of the optimal fitness for several consecutive generations is less than a minimum threshold, or the preset maximum number of iterations is reached, such as 1000. Finally, the algorithm outputs the scheme represented by the individual with the highest fitness in the current population as the optimal reservation allocation scheme and vaccination path planning scheme.

[0047] The execution control module is responsible for transforming the optimal solution generated by the multi-objective optimization decision-making module into specific, executable instructions, and ensuring their reliable distribution and execution. Please refer to the appendix. Figure 4 This module includes an instruction generation unit and a distribution and scheduling unit. After receiving the optimized plan, the instruction generation unit first parses the instructions. The parsing process decomposes the plan into three types of executable instructions. The first type is the appointment confirmation instruction, which includes a unique identifier for the specific vaccination site, a precise vaccination time schedule, a unique appointment QR code or numeric code generated for the user, and necessary precautions text. The second type is the inventory allocation instruction, generated when the system decision involves the allocation of vaccines between different vaccination sites.

[0048] The instructions include the international code for the vaccine type to be allocated, the quantity to be allocated, the source vaccination site identifier, the target vaccination site identifier, and the suggested allocation time window. The third type is route navigation instructions, which generate a detailed driving route from the user's current location to the designated vaccination site based on a route planning scheme. Route information includes turn instructions, estimated travel time, traffic congestion alerts, and key landmarks. The instruction generation unit ensures that all instruction formats are standardized and conform to the data interface specifications of each execution terminal. The distribution and scheduling unit is responsible for the transmission and scheduling of instructions.

[0049] This unit employs a highly available message queue mechanism to ensure reliable instruction transmission. Different types of instructions are delivered to different message topics. Appointment confirmation instructions are typically sent to user terminals via mobile application push services or SMS gateways. Inventory allocation instructions are sent to the console or mobile terminal of the regional logistics dispatch center. Route navigation instructions are integrated into the navigation application used by the user. The distribution and dispatch unit also implements a priority management mechanism; for example, for urgent vaccine allocation instructions or appointment confirmations from high-priority users, higher message priorities are set to ensure they are processed first. The sending status, delivery status, and final execution result of all instructions are recorded and fed back to the system's log and audit module.

[0050] The feedback learning module is a key component for the system to achieve continuous self-optimization. Please refer to the appendix for its online calibration and model optimization principles. Figure 5 This module continuously collects actual system operation data and compares this data with the expected data generated by the multi-objective optimization decision-making module when it initially generated the plan, calculating deviation information. The collected deviation data mainly covers three aspects. The first aspect is time deviation, including the difference between the user's actual vaccination time and the appointment time, and the difference between the user's actual travel time and the system's estimated time.

[0051] These time-related data are obtained through the check-in system at vaccination sites and travel reports from navigation applications. The second aspect is quantity deviation, primarily the difference between the actual amount of vaccines used and the amount pre-allocated by the system, reflecting the accuracy of inventory forecasting. The third aspect is status deviation, such as the difference between the actual and estimated service time at the vaccination station, or the difference between the actual work efficiency of medical staff and the scheduling assumptions. The feedback learning module utilizes these continuous deviation data streams to perform online correction of the parameters of the upstream key models. Specifically, the module maintains a deviation dataset and periodically, for example, initiates a model parameter optimization task every 24 hours. The task employs a gradient descent algorithm to minimize the overall error between predicted and actual values.

[0052] For the inventory turnover risk prediction model in the dynamic coupling analysis module, the gradient descent algorithm adjusts the coefficients in its time series fitting model and the weight parameters in the risk coefficient calculation. For the service capacity assessment model, it adjusts the implicit parameters in its service capacity calculation, such as the work efficiency of medical staff and the turnover speed of vaccination stations. For the multi-objective optimization decision-making module, the feedback learning module analyzes the actual system performance under different objective weights and fine-tunes the weight coefficients in the fitness function to make the system behavior more consistent with actual operational goals.

[0053] All parameter adjustments are made incrementally to ensure stable system operation. The adjusted parameters are then hot-updated to the corresponding analysis and decision models, enabling online self-evolution of the models. The entire feedback learning process is meticulously recorded, including the parameters adjusted each time and the model performance metrics before and after the adjustment, for auditing and analysis by system administrators.

[0054] During deployment, the system employs a microservice architecture, encapsulating each module as an independent service. Services communicate via remote procedure calls or asynchronous messaging. All services are registered with a unified service registration and discovery center. A centralized configuration management center manages the operational parameters of all modules. The data persistence layer uses a hybrid approach of relational and time-series databases; the relational database stores business entities and relational data, while the time-series database stores a large amount of sensor readings and real-time monitoring data. The system front-end, i.e., the presentation layer of the user service interface module, adopts a responsive web design to ensure a good user experience on both mobile and desktop devices. The entire system is deployed on cloud infrastructure, possessing elastic scaling capabilities to handle the load pressure during peak appointment periods. The system also establishes comprehensive logging, monitoring and alerting, and audit trail functions to ensure all operations are traceable and meet the stringent requirements of medical data security and service compliance.

[0055] A user initiates a specific appointment request through a mobile application, and the process embodies the collaboration of the aforementioned modules. The user opens the application, selects the desired vaccine type, such as a combination of influenza and pneumococcal vaccines, enters their preferred vaccination time window (e.g., 2 PM to 4 PM), and authorizes the application to access their current geographical location. The user service interface module receives this request, performs basic validation, and places it in the pending request queue. Almost simultaneously, the resource status awareness module updates the status of all vaccination sites throughout the city, including the inventory of influenza and pneumococcal vaccines at each site, the availability of each vaccination station, and the on-duty status of medical staff.

[0056] The dynamic coupling analysis module consumes this data to calculate the inventory turnover risk coefficient and service capacity margin for the two vaccines at different vaccination sites during the period from 2 PM to 4 PM in the future. For example, it identifies that vaccination site A has sufficient influenza vaccine inventory but limited service capacity, while vaccination site B has ample service capacity but the pneumococcal vaccine is at risk of expiring. The multi-objective optimization decision-making module then runs its improved genetic algorithm based on these analysis results and the appointment requests of all other users. After hundreds of iterations, the algorithm finds an optimal solution, deciding to assign the user to vaccination site C at 2:30 PM, and plans an optimal driving route from the user's current location to vaccination site C, with an estimated travel time of 20 minutes.

[0057] The execution control module then generates three instructions: a confirmation instruction sent to the user's mobile application, including the address of vaccination site C, the 2:30 time slot, and a unique QR code; a route navigation instruction integrated into the user's map application; and an internal instruction notifying the vaccination site C system of the newly added appointment. The user arrives at vaccination site C at the appointed time, scans the QR code to confirm, and completes the vaccination. After vaccination, the feedback learning module collects the actual execution data of this appointment, including whether the user arrived on time, the actual vaccination time, the vaccine batch number, and actual usage, and compares it with the expected data for subsequent model fine-tuning and optimization. The entire process achieves a high degree of coordination between resource scheduling and route planning, ensuring optimal service efficiency and resource utilization.

[0058] When handling large-scale concurrent booking requests, the system's internal data flow and decision-making processes become more complex. Suppose that at a certain moment, the system simultaneously receives thousands of booking requests from different areas of the city. The load balancer of the user service interface module distributes these requests to multiple backend service instances for processing. Each request undergoes rigorous validation and formatting. The data collection frequency of the resource status awareness module may automatically increase during peak periods, for example, adjusting from once per minute to once every 30 seconds, to ensure the timeliness of status data. The dynamic coupling analysis module faces massive, high-dimensional data analysis tasks.

[0059] Its inventory turnover risk prediction unit needs to process risk calculations for dozens of vaccine types and hundreds of vaccination sites in parallel. The service capacity assessment unit needs to simulate the load changes at each vaccination site in real time over the next few hours. These computational tasks may be broken down into multiple sub-tasks and executed in parallel within a computing cluster. The multi-objective optimization decision module faces a massive combinatorial optimization problem. Its genetic algorithm population size may need to be dynamically expanded to thousands of individuals, and crossover and mutation operators need adaptive adjustments to avoid premature convergence. The computation process may consume significant computing resources and last for several minutes.

[0060] During this period, new appointment requests continue to pour in, requiring the system to handle this dynamic input. The execution control module needs to handle extremely high message throughput during instruction distribution. Its message queue needs high availability and persistence to prevent instruction loss. Inventory allocation instructions may trigger cross-regional vaccine logistics scheduling, requiring deep integration with external logistics systems. After peak periods, the feedback learning module will acquire massive amounts of deviation data for comprehensive model retraining, significantly improving the system's prediction and decision-making accuracy for similar future peak periods. Through deep collaboration and elastic scaling of its modules, the entire system ensures stable operation and service performance under high-concurrency scenarios.

[0061] The system's robustness is also demonstrated by its handling mechanism for anomalies. For example, when the resource status awareness module detects that the cold chain temperature at a vaccination site is continuously exceeding the standard, it immediately marks the affected vaccine inventory at that site as unavailable and generates a high-priority alarm event. This event will notify the dynamic coupling analysis module in real time, which will immediately recalculate the service capacity and risk indicators of the relevant vaccination site, and may conclude that the site is temporarily unable to provide a certain vaccine service.

[0062] In the next round of optimization calculations, the multi-objective optimization decision module will automatically avoid allocating to that vaccination site and reschedule appointment requests already assigned to that site to other available sites. The execution control module will then send appointment change notifications to affected users and provide new appointment options. The entire exception handling process is automated, minimizing the impact on user services. For example, when the user service interface module detects an abnormal surge in appointment requests from a certain geographical area, potentially indicating a local outbreak or a specific promotional activity, this information will be flagged and passed to the dynamic coupling analysis module. The dynamic coupling analysis module can then adjust its demand forecasting model accordingly, providing early warnings of potential resource shortages. The multi-objective optimization decision module can then initiate optimization calculations in advance, pre-allocate resources, or coordinate with resource suppliers to increase vaccine allocation. These mechanisms collectively ensure that the system can maintain efficient and stable service output even in the face of internal and external disturbances.

[0063] In summary, this embodiment details the specific technical implementation of a vaccine appointment optimization system based on resource scheduling and path planning. From user request access and verification, to real-time perception and cleansing of the entire network resource status, to dynamic coupling analysis and risk assessment based on multi-dimensional data, and then to generating collaborative appointment allocation and path planning schemes through an improved multi-objective optimization algorithm, the system executes the schemes through a reliable instruction generation and distribution mechanism, supplemented by continuous feedback learning to achieve self-evolution of the system model. This system, through deep integration of resource scheduling and path planning, achieves refined and intelligent management and optimization of the entire vaccination service chain.

[0064] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A vaccine appointment optimization system based on resource scheduling and path planning, characterized in that, include: The user service interface module is used to receive vaccination appointment requests submitted by users. The vaccination appointment request includes the type of vaccine to be administered, the desired vaccination time window, and the user's geographical location information. The resource status awareness module is used to collect and update vaccine inventory data, vaccination station working status, medical staff scheduling information and real-time vaccination load data at each vaccination site in real time. The dynamic coupling analysis module is used to establish a vaccine inventory turnover risk prediction model and a vaccination site service capacity assessment model based on the output of the resource status perception module and the reservation request data of the user service interface module. The multi-objective optimization decision module receives the output results from the dynamic coupling analysis module and generates the optimal appointment allocation scheme and vaccination route planning scheme by solving the joint optimization problem of resource scheduling and path planning. The execution control module is used to decompose the optimal solution generated by the multi-objective optimization decision-making module into specific reservation confirmation instructions, inventory allocation instructions, and route navigation instructions, and distribute them to the corresponding terminal devices.

2. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 1, characterized in that, The dynamic coupling analysis module includes an inventory turnover risk prediction unit and a service capability assessment unit. The inventory turnover risk prediction unit calculates the near-expiration risk coefficient of various vaccines by analyzing the matching degree between historical no-show data and current reservation data. The service capacity assessment unit calculates the service capacity margin of each vaccination site within a specified time interval based on the working status of the vaccination station, the configuration of medical staff, and real-time load data.

3. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 2, characterized in that, The inventory turnover risk prediction unit uses time series analysis to fit the trend of reservation data and combines it with vaccine expiration date data to build a risk warning model. The service capacity assessment unit quantifies the distribution of service pressure at vaccination sites by establishing a load balancing index.

4. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 1, characterized in that, The multi-objective optimization decision module includes an objective function construction unit and a constraint processing unit; The objective function construction unit defines three core objectives for system optimization: minimizing the average user waiting time, maximizing vaccine inventory turnover rate, and minimizing the load imbalance at vaccination sites. The constraint processing unit sets four types of hard constraints, including upper limit constraint on vaccination site service capacity, vaccine inventory availability constraint, user time window satisfaction constraint, and path reachability constraint.

5. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 4, characterized in that, The multi-objective optimization decision module uses an improved genetic algorithm to solve the multi-objective optimization problem; The improved genetic algorithm enhances search efficiency by introducing an elite retention strategy and an adaptive crossover mutation operator.

6. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 5, characterized in that, The solution process of the improved genetic algorithm includes the following steps: First, the population is initialized, with each individual representing a complete reservation allocation and path planning scheme. Then, the fitness value of each individual is calculated, and the fitness function is obtained by weighted summation of the objective function; Next, a selection process is performed, employing a roulette wheel selection mechanism combined with an elite retention strategy; Then, cross-operations are performed, cross-operations are performed on both the reservation allocation scheme and the route planning scheme; Finally, a mutation operation is performed to locally adjust the allocation of inoculation points or the order of paths in the scheme with a preset probability.

7. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 1, characterized in that, The execution control module includes an instruction generation unit and a distribution and scheduling unit; The instruction generation unit parses the optimization scheme into three types of executable instructions: the appointment confirmation instruction includes the specific location of the vaccination site, the vaccination time, and the appointment code information; the inventory allocation instruction specifies the type, quantity, and target vaccination site of the vaccine to be allocated; and the path navigation instruction generates the optimal driving route from the user's current location to the designated vaccination site. The distribution and scheduling unit uses a message queue mechanism to ensure reliable transmission of instructions and uses priority settings to ensure timely processing of urgent instructions.

8. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 1, characterized in that, The system also includes a feedback learning module; The feedback learning module collects deviation information between actual execution data and expected plans, including the difference between actual vaccination time and appointment time, the difference between actual route travel time and estimated time, and the difference between actual vaccine usage and estimated usage. The feedback learning module uses these deviation data to perform online calibration of the parameters of the dynamic coupling analysis model and the multi-objective optimization model.

9. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 8, characterized in that, The feedback learning module continuously optimizes the model's prediction accuracy using the gradient descent algorithm. The online correction process includes periodically initiating model parameter optimization tasks and using the gradient descent algorithm to minimize the overall error between the predicted and actual values.

10. The vaccine appointment optimization system based on resource scheduling and path planning according to claim 1, characterized in that, The user service interface module supports multiple terminal access methods, including mobile application, web page and self-service terminal; The user service interface module provides appointment query, appointment modification and appointment cancellation functions. All operation changes are synchronized to the resource status awareness module and dynamic coupling analysis module in real time. The user service interface module also integrates electronic payment functionality, supporting online payment of vaccination fees.

Images