A dynamic intelligent scheduling system for cross-departmental physical examination services

By constructing a multi-dimensional resource capability profile and project dependency graph, and combining an improved particle swarm optimization algorithm and feedback optimization module, the problem of low resource utilization in cross-departmental physical examinations was solved, and an efficient scheduling scheme was adaptively adjusted and optimized.

CN123120237APending Publication Date: 2026-11-17INNER MONGOLIA SKYLARK HEALTH TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611541743.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing physical examination scheduling scheme fails to effectively handle the differences in capabilities and medical constraints of resources across departments, resulting in low resource utilization, inability to respond quickly to sudden abnormalities, and lack of feedback optimization mechanisms, making it difficult to improve scheduling efficiency.

Method used

We construct a multi-dimensional resource capability profile and project dependency graph, use an improved particle swarm optimization algorithm to generate a scheduling scheme, and adjust it in real time through execution monitoring and feedback optimization modules to ensure that the scheduling scheme meets medical constraints and demand matching.

Benefits of technology

It achieves precise matching of resources and demand, reduces waiting time for physical examinations, improves scheduling efficiency, can quickly respond to sudden anomalies, and continuously optimizes scheduling strategies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120237A_ABST
    Figure CN123120237A_ABST
Patent Text Reader

Abstract

This invention relates to the field of medical resource scheduling technology and discloses a dynamic intelligent scheduling system for cross-departmental physical examination services. The system includes a resource capability profile construction module, a project dependency graph construction module, a dynamic scheduling decision module, an execution monitoring module, a feedback optimization module, and a priority adaptation module. This invention achieves precise matching of resources and needs by constructing a multi-dimensional resource capability profile and project constraint graph, avoiding resource mismatch. It employs a particle swarm optimization algorithm customized for physical examination constraints to generate a scheduling scheme, comprehensively considering multiple objectives such as examinee waiting time, departmental load balancing, and resource utilization, thereby improving overall scheduling efficiency. Based on attendance records, it enables adaptive adjustment for abnormal events, quickly responding to sudden anomalies without the need for additional positioning infrastructure deployment, reducing the impact of anomalies on the physical examination process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of medical resource scheduling technology, specifically a dynamic intelligent scheduling system for cross-departmental physical examination services. Background Technology

[0002] The physical examination service involves the collaboration of multiple departments and includes multiple items such as blood drawing, ultrasound, electrocardiogram, and imaging examination. There are complex constraints between different items, such as fasting constraints, time-series dependencies, and cross-departmental collaboration. At the same time, there are differences in the number of equipment, personnel qualifications, and real-time load of each department.

[0003] Most existing physical examination scheduling schemes adopt a simple queuing and diversion model, allocating examinees solely based on the current number of people in the queue at each department. This fails to consider differences in departmental resource capabilities and the constraints of different procedures, making it impossible to achieve a precise match between resources and demand. When abnormal events occur, such as equipment failure, temporary staff absences, or sudden increases in the number of people in the queue, existing schemes cannot quickly adjust their scheduling strategies, leading to increased waiting times for examinees and decreased resource utilization. Furthermore, existing scheduling schemes lack feedback and optimization mechanisms, making it impossible to continuously optimize scheduling strategies based on historical results, thus hindering the improvement of overall scheduling efficiency.

[0004] While existing general scheduling algorithms (such as genetic algorithms, ant colony algorithms, and standard particle swarm optimization algorithms) can be used for path optimization, they are not adapted to the specific constraints of physical examination scenarios. They cannot handle hard constraints such as fasting constraints and medical timing constraints. Direct application of these algorithms will result in scheduling schemes that violate medical standards.

[0005] Therefore, a dynamic intelligent scheduling system for cross-departmental physical examination services is designed. Summary of the Invention

[0006] In order to overcome the shortcomings of the prior art, this invention provides a dynamic intelligent scheduling system for cross-departmental physical examination services, which effectively solves the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a dynamic resource intelligent scheduling system for cross-departmental physical examination services, comprising: The resource capability profile building module is used to collect static attribute data and dynamic operation data of each department to build a multi-dimensional resource capability profile that includes equipment status, personnel qualifications, real-time load, historical efficiency, and spatial location. The project dependency graph construction module is used to parse the medical constraints, temporal constraints, and cross-departmental collaborative constraints of physical examination items, and construct a project dependency graph for physical examination items. The dynamic scheduling decision module is used to receive the project list of the examinee and generate an initial scheduling scheme by combining the resource capability profile and project dependency graph with a multi-objective optimization algorithm. The execution monitoring module is used to collect resource status data and project progress data in real time during the execution of the scheduling plan, monitor abnormal events, and trigger dynamic adjustments to the scheduling plan. The feedback optimization module is used to collect efficiency data and resource utilization data after the scheduling scheme is executed, and to iteratively optimize the algorithm parameters of the dynamic scheduling decision module.

[0008] Preferably, the resource capability profiling module includes: The static attribute acquisition unit is used to collect information such as the model and quantity of equipment in the department, the qualification level of medical staff, and the spatial location of the department. The dynamic data monitoring unit is used to collect equipment operating status, current number of people in queue, average execution time of a single project, and on-duty status of medical staff in real time according to a hierarchical collection strategy. Among them, equipment status data is collected on a minute-level cycle, while queue and personnel status data are collected on a second-level cycle. The profile generation unit is used to integrate static attribute data with dynamic data to generate a capability score matrix for each department's resources. The capability score matrix includes four dimensions: equipment availability, personnel matching degree, load saturation, and location convenience. The location convenience is calculated based on the distance between the department location corresponding to the examinee's most recent check-in record and the target department.

[0009] Preferably, the project dependency graph construction module includes: The constraint parsing unit is used to identify fasting constraints, pre- and post-dependent constraints, and cross-departmental collaborative constraints for physical examination items. Fasting constraints mark the attributes of items that need to be completed on an empty stomach, pre- and post-dependent constraints mark the execution order requirements between items, and cross-departmental collaborative constraints mark related items that need to be completed in sequence within a preset time window. The graph generation unit is used to transform constraint relationships into a directed acyclic graph. Each node corresponds to a physical examination item, and each edge corresponds to the constraint relationship between items. The edge weight is the constraint priority. Cross-departmental collaborative constraints are expressed by setting directed edges with time window labels between related item nodes.

[0010] Preferably, the dynamic scheduling decision module includes: The constraint verification unit is used to verify whether the initial scheduling scheme conforms to the constraint relationship in the project dependency graph and to eliminate scheduling paths that do not conform to the constraints. The multi-objective optimization unit takes the shortest total waiting time for examinees, the highest departmental load balance, and the maximum resource utilization as its objective functions. It uses an improved particle swarm optimization algorithm to solve for the optimal scheduling path and generate an initial scheduling scheme. The improvements to the particle swarm optimization algorithm include: introducing a constraint violation penalty function into the fitness function, encoding high-priority constraints in the project dependency graph as penalty terms, and adding penalty values ​​to the fitness value when a scheduling path violates a high-priority constraint. The degree of violation of the fasting constraint is quantified by the length of time elapsed after eating, with a larger penalty value for longer time. An adaptive inertia weight strategy is adopted, with the inertia weight decreasing linearly with the number of iterations. A Gaussian mutation operator is introduced, applying a Gaussian perturbation to the current optimal particle position when the global optimal solution is not updated after a preset number of consecutive iterations.

[0011] Preferably, the execution monitoring module includes: The status acquisition unit is used to collect real-time data on changes in the number of people queuing in each department, equipment malfunction status, temporary absence status of medical staff, and check-in records of examinees. Anomaly identification unit is used to identify four types of abnormal events: equipment failure, staff leaving their posts, sudden increase in the number of people in the queue, and medical examinees staying in the queue. Among them, medical examinees staying in the queue is determined by the fact that a medical examinee has not signed in to the next department after a preset time. Sudden increase in the number of people in the queue is determined by the fact that the number of people in the queue exceeds a preset proportion of the department's maximum capacity for receiving patients. The maximum capacity for receiving patients is the maximum number of examinations that can be completed by all equipment and personnel in the department per unit time. The dynamic adjustment unit triggers local rescheduling when an abnormal event is detected, adjusting only the execution order and corresponding departments of the affected projects, while keeping the scheduling scheme of the unaffected projects unchanged.

[0012] Preferably, the dynamic adjustment unit automatically reschedules affected items to nearby departments with similar equipment when equipment malfunctions, wherein the nearby department is the department with the closest physical location and the same examination function; when staff are absent from their posts, the unit reschedules items to departments with qualified medical staff on duty; when there is a sudden increase in the number of people in the queue, the unit adjusts the corresponding items for subsequent examinees to similar departments with lower workloads; and when examinees are delayed, the unit re-plans the order of subsequent items for that examinee.

[0013] Preferably, the feedback optimization module includes: The data acquisition unit is used to collect the total time spent by examinees, the average load rate of each department, the equipment idle rate, and the number of abnormal adjustments for each scheduling plan. The effectiveness evaluation unit compares the collected data with historical scheduling data and calculates the efficiency score of the current scheduling plan using a normalization method. The efficiency score is calculated as follows: Efficiency score = 0.4 × max(0,1 - total time spent by examinees / historical average total time spent) + 0.3 × max(0,1 - standard deviation of load rate of each department / historical average standard deviation) + 0.3 × max(0,1 - equipment idle rate / historical average equipment idle rate), where the max function ensures that the score of each item is not lower than 0, and the historical average is the moving average of the past 30 natural days. The parameter optimization unit is used to adjust the weight coefficients of each objective function in the multi-objective optimization unit according to the efficiency score. When the efficiency scores of multiple consecutive scheduling schemes are lower than the threshold, global optimization of algorithm parameters is triggered.

[0014] Preferably, it also includes a priority adaptation module, which is used to assign different scheduling priorities based on the examinee's age, health status, and appointment type. The scheduling scheme for high-priority examinees prioritizes the execution order of fasting and emergency items.

[0015] Preferably, the scheduling scheme generated by the dynamic scheduling decision module is pushed to the examinee through the terminal. The terminal displays the current project progress, the location of the next project department, and the estimated waiting time in real time. At the same time, it supports the examinee to adjust the execution order of non-constrained projects. The adjusted scheme can only be executed after it has been verified by the constraint verification unit.

[0016] A method for intelligent scheduling of dynamic resources for cross-departmental physical examination services based on the above system includes the following steps: Step 1: The resource capability profile building module collects static attributes and dynamic operational data of each department in real time according to the hierarchical collection strategy, and generates a multi-dimensional resource capability profile. Step 2: The project dependency graph construction module parses the constraints of the physical examination items and generates a project dependency graph; Step 3: The priority matching module assigns scheduling priorities based on the examinee's age, health status, and appointment type; Step 4: The dynamic scheduling decision module receives the list of items for the examinee, and combines the resource capability profile, item dependency graph and scheduling priority to generate an initial scheduling scheme using an improved particle swarm algorithm; Step 5: Push the scheduling plan to the examinee's terminal to display the project progress and estimated waiting time in real time; Step 6: The monitoring module monitors the execution status of the scheduling plan in real time based on the check-in records, and triggers local rescheduling when an abnormal event occurs; Step 7: The feedback optimization module collects scheduling execution data, iteratively optimizes scheduling algorithm parameters, and improves the efficiency of subsequent scheduling schemes.

[0017] Compared with the prior art, the beneficial effects of the present invention are: This invention achieves precise matching of resources and needs by constructing a multi-dimensional resource capability profile and project constraint graph, avoiding resource mismatch. It employs a particle swarm optimization algorithm tailored to physical examination constraints to generate a scheduling scheme, comprehensively considering multiple objectives such as examinee waiting time, departmental load balancing, and resource utilization, thereby improving overall scheduling efficiency. Based on attendance records, it enables adaptive adjustment for abnormal events, quickly responding to sudden anomalies without the need for additional location infrastructure deployment, reducing the impact of anomalies on the physical examination process. Through a feedback optimization mechanism, the scheduling strategy is continuously iterated, and scheduling efficiency can be continuously improved over usage time, making it suitable for cross-departmental resource scheduling scenarios in various physical examination institutions. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0019] In the attached diagram: Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the scheduling process of the present invention; Figure 3 This is a logical block diagram of the project dependency graph of the present invention; Figure 4 This is a block diagram of the dynamic scheduling logic of the present invention; Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] Depend on Figures 1-4 The present invention relates to a dynamic intelligent scheduling system for cross-departmental physical examination services, comprising a resource capability profile construction module, a project dependency graph construction module, a dynamic scheduling decision module, an execution monitoring module, and a feedback optimization module.

[0022] The resource capability profile building module collects static attribute data and dynamic operational data from each department to construct a multi-dimensional resource capability profile that includes equipment status, personnel qualifications, real-time load, historical efficiency, and spatial location, providing resource capability data support for scheduling decisions. The dynamic data adopts a hierarchical collection strategy, with equipment status data collected on a minute-by-minute basis and queuing and personnel status data collected on a second-by-second basis, balancing data timeliness and system load. When calculating location convenience, the profile generation unit determines the examinee's current location based on the department location corresponding to their most recent check-in record, without the need for additional positioning infrastructure deployment.

[0023] The project dependency graph construction module analyzes the medical constraints, temporal constraints, and cross-departmental collaborative constraints of physical examination items, constructs a project dependency graph, clarifies the execution order and collaborative requirements between items, and avoids scheduling schemes from violating medical constraints. Among them, cross-departmental collaborative constraints are for related items that need to be completed in sequence within a preset time window. For example, an enhanced CT examination requires kidney function blood sampling, contrast agent injection, and then scanning. Such related items are represented in a directed acyclic graph by setting directed edges with time window labels between related item nodes, ensuring that the scheduling scheme arranges each related item in sequence within the time window.

[0024] The dynamic scheduling decision module receives the examination patient's item list, combines resource capability profiles and item dependency graphs, and generates an initial scheduling scheme using a multi-objective optimization algorithm. This multi-objective optimization algorithm can be implemented using various swarm intelligence optimization algorithms such as particle swarm optimization, genetic algorithm, and ant colony optimization; in this embodiment, an improved particle swarm optimization algorithm is preferred. This algorithm incorporates three customized improvements for the physical examination scenario: First, a constraint violation penalty function is introduced into the fitness function, encoding high-priority constraints in the item dependency graph as penalty items. The degree of violation of the fasting constraint is quantified by the length of time elapsed after eating; the longer the time, the larger the penalty value, ensuring that the generated scheduling path does not violate the fasting constraint and medical timing constraints. Second, an adaptive inertia weight strategy is adopted, with the inertia weight decreasing linearly with the number of iterations, emphasizing global search in the early stages and local fine-grained search in the later stages. Third, a Gaussian mutation operator is introduced; when the global optimal solution is not updated after a preset number of iterations, a Gaussian perturbation is applied to the current optimal particle position to avoid the algorithm getting trapped in local optima. The goal is to achieve precise matching of resources and needs, minimizing the total waiting time for examinees, maximizing departmental load balance, and maximizing resource utilization.

[0025] The monitoring module collects resource status and project progress data in real time during the execution of the scheduling plan, monitoring four types of abnormal events: equipment failure, staff absence, sudden increases in queues, and patients remaining in their medical examinations. Patient retention is determined by attendance records—if a patient completes attendance in one department but fails to complete attendance in the next department within a preset timeframe, they are considered to be in a state of retention. This method does not require additional indoor positioning infrastructure. A sudden increase in queues is determined by the number of people in the queue exceeding a preset percentage of the department's maximum capacity, which is the maximum number of examinations that can be completed by all equipment and personnel in the department per unit time. When an anomaly occurs, a local rescheduling is triggered, adjusting only the affected items and departments while maintaining the stability of the unaffected parts of the scheduling.

[0026] The feedback optimization module collects efficiency and resource utilization data after the scheduling scheme is executed, calculates the efficiency score using a normalization method, and ensures that the score of each sub-item is not lower than 0 through a max function to avoid negative values. The historical average value in the efficiency score calculation is the moving average of the past 30 calendar days to ensure the timeliness of the historical benchmark. Based on the efficiency score, the algorithm parameters of the dynamic scheduling decision module are iteratively optimized to continuously improve the efficiency of the scheduling scheme.

[0027] Example 1

[0028] A cross-departmental physical examination business dynamic resource intelligent scheduling system includes a resource capability profile construction module, a project dependency graph construction module, a dynamic scheduling decision module, an execution monitoring module, a feedback optimization module, and a priority adaptation module.

[0029] The resource capability profiling module includes a static attribute collection unit, a dynamic data monitoring unit, and a profiling generation unit. The static attribute collection unit connects to the hospital information system to collect information on equipment models, quantities, medical staff qualification levels, and departmental locations for each department. Medical staff qualification levels include the scope of projects they can perform, years of experience, and proficiency scores. The dynamic data monitoring unit connects to departmental execution terminals and IoT devices to collect real-time data on equipment operating status, current queue size, average execution time for a single project, and medical staff on-duty status, according to a tiered collection strategy. Equipment operating status data is collected on a minute-by-minute basis, while queue size and on-duty status data are collected on a second-by-second basis, balancing data timeliness and system computational load. The generation unit integrates static attribute data with dynamic data to generate a capability score matrix for each department's resources. The capability score matrix includes four dimensions: equipment availability, personnel matching, load saturation, and location convenience. Each dimension has a score range of 0-100. Equipment availability is calculated based on the percentage of normal equipment uptime. Personnel matching is calculated based on the degree of matching between the qualifications of on-duty personnel and project requirements. Load saturation is calculated based on the ratio of the current number of people in the queue to the department's maximum capacity. Location convenience is calculated based on the distance between the department location corresponding to the examinee's most recent check-in record and the target department. The examinee's current location does not need to be obtained through additional positioning equipment; it is directly determined by the location of the department where the examinee last checked in, as recorded by the check-in terminal.

[0030] The project dependency graph construction module includes a constraint parsing unit and a graph generation unit. The constraint parsing unit reads the constraint information of each item from the physical examination item rule base, identifying fasting constraints, pre- and post-dependent constraints, and cross-departmental collaborative constraints. Fasting constraints mark the attributes that require fasting to complete, including fasting duration requirements and the time range during which the item cannot be executed after eating. Pre- and post-dependent constraints mark the execution order requirements between items; for example, abdominal ultrasound must be performed after blood draw to avoid food affecting the examination results. Cross-departmental collaborative constraints mark related items that need to be completed sequentially within a preset time window; for example, enhanced CT scans require kidney function blood draws before contrast agent injection and then scanning, and exercise stress tests require simultaneous completion of ECG monitoring data within a preset time window after exercise begins. The data collection and graph generation unit transforms the constraint relationships into a directed acyclic graph (DAG). Each node corresponds to a physical examination item, and each edge corresponds to the constraint relationship between items. The edge weight represents the constraint priority, which is divided into three levels: high, medium, and low. High-priority constraints are medical constraints that must be met, medium-priority constraints are time-series constraints that improve efficiency, and low-priority constraints are collaborative constraints that optimize user experience. Among them, cross-departmental collaborative constraints are expressed in the DAG by setting directed edges with time window labels between related item nodes. The direction of the directed edge indicates the execution order between items, and the time window label on the edge indicates that all related items must be completed within this time range, ensuring that the scheduling algorithm arranges related items to be executed sequentially within the same time window during the solution process.

[0031] The dynamic scheduling decision module includes a constraint verification unit and a multi-objective optimization unit. The constraint verification unit receives the examinee's item list, traverses all constraint edges in the item dependency graph, verifies whether the initial scheduling path conforms to all high-priority constraints, and eliminates scheduling paths that do not conform to the constraints. The multi-objective optimization unit uses the shortest total waiting time for examinees, the highest departmental load balance, and the highest resource utilization as objective functions, and employs an improved particle swarm optimization algorithm to solve for the optimal scheduling path. The multi-objective optimization algorithm can be implemented using various swarm intelligence optimization algorithms such as particle swarm optimization, genetic algorithm, and ant colony optimization; in this embodiment, the improved particle swarm optimization algorithm is preferred. The improvements to the particle swarm optimization algorithm in this embodiment include the following three aspects: First, a constraint violation penalty function is introduced into the fitness function. High-priority constraints in the project dependency graph are encoded as penalty items. When a scheduling path violates a high-priority constraint (such as scheduling a project that needs to be completed on an empty stomach in the post-meal time period), the fitness value is superimposed with a penalty value, and the penalty value is proportional to the degree of constraint violation. The degree of violation of the fasting constraint is quantified by the length of time elapsed after eating. For example, if an empty stomach project is scheduled 1 hour after eating, the penalty value is the base penalty coefficient multiplied by 1; if it is scheduled 3 hours after eating, the penalty value is the base penalty coefficient multiplied by 3. The longer the time, the larger the penalty value, ensuring that feasible solutions that satisfy all hard constraints are generated first during the algorithm search process.

[0032] Second, an adaptive inertia weight strategy is adopted; the inertia weight decreases linearly with the number of iterations, with an initial value of 0.9 and a final value of 0.4, so that the algorithm focuses on global search and explores a larger range of solution space in the early stage of iteration, and focuses on local fine search and converges to the vicinity of the optimal solution in the later stage of iteration.

[0033] Third, a Gaussian mutation operator is introduced. When the global optimal solution is not updated after a preset number of iterations, it is determined that the algorithm may be trapped in a local optimum. At this time, a Gaussian perturbation is applied to the current optimal particle position. The perturbation amplitude follows a Gaussian distribution with a mean of 0 and a standard deviation that is the reciprocal of the current iteration number, which helps the algorithm escape from the local optimum.

[0034] In this embodiment, the algorithm parameters are set as follows: 50 particles, 200 iterations, 2.0 learning factor, and 30 consecutive times without updating the global optimal solution. Finally, an initial scheduling scheme is generated, which includes the execution order of each project, the corresponding department, the expected start time, and the expected waiting time.

[0035] The monitoring module includes a status acquisition unit, an anomaly identification unit, and a dynamic adjustment unit. The status acquisition unit collects real-time data on changes in the number of people queuing in each department, equipment malfunction status, temporary absence of medical staff, and check-in records of examinees, updating data every 5 seconds. The anomaly identification unit identifies four types of abnormal events based on the collected data: equipment malfunctions are identified through equipment operating status signals; staff absences are identified through check-in status on medical staff terminals; a sudden increase in the number of people queuing is identified when the number of people queuing exceeds 80% of the department's maximum capacity (the maximum number of examinations that can be completed by all equipment and personnel in the department per unit time); and examinee delays are determined when an examinee fails to check in at a department for more than a preset time (15 minutes in this embodiment) after checking in at a previous department. The dynamic adjustment unit triggers local rescheduling when an abnormal event is detected: for equipment malfunctions, the affected items are automatically rescheduled to other departments. For nearby departments with similar equipment, which are physically the closest departments with the same examination functions, the selection of nearby departments prioritizes the department with the highest equipment availability in the capability scoring matrix. For abnormal staff absences, projects are reassigned to departments with qualified medical staff on duty, prioritizing departments with the highest staff matching score. For abnormal surges in queue numbers, subsequent examinees' corresponding projects are adjusted to similar departments with lower loads, prioritizing departments with the lowest load saturation score. For abnormal examinee delays, the order of subsequent projects for that examinee is replanned, prioritizing projects that are currently closest and have the shortest waiting time. Local rescheduling only adjusts the execution order and corresponding departments of affected projects, keeping the scheduling scheme for unaffected projects unchanged to avoid process chaos caused by global adjustments.

[0036] The feedback optimization module includes a data acquisition unit, an effectiveness evaluation unit, and a parameter optimization unit. The data acquisition unit collects the total time spent by examinees, the average load rate of each department, the equipment idle rate, and the number of abnormal adjustments for each scheduling plan; the data is stored daily. The effectiveness evaluation unit compares the collected data with historical scheduling plan data and uses a normalization method to calculate the efficiency score of the current scheduling plan. The efficiency score calculation formula is as follows: Efficiency score = 0.4 × max(0,1 - total time spent by examinees / historical average total time spent) + 0.3 × max(0,1 - standard deviation of workload rate of each department / historical average standard deviation) + 0.3 × max(0,1 - equipment idle rate / historical average equipment idle rate) The max function ensures that the score of each sub-item is not lower than 0, avoiding negative scores when the current indicator is worse than the historical average. The historical average total time, historical average standard deviation, and historical average equipment idle rate are all taken as the moving average of the past 30 natural days to ensure the timeliness of the historical benchmark and avoid scoring distortion due to outdated historical data. The parameter optimization unit adjusts the weight coefficients of each objective function in the multi-objective optimization unit according to the efficiency score. When the efficiency score is higher than 80, the current weight coefficients remain unchanged. When the efficiency score is between 60 and 80, the weight coefficients are fine-tuned to increase the weight of objective functions with lower efficiency scores. When the efficiency scores of 7 consecutive scheduling schemes are lower than 60, global optimization of algorithm parameters is triggered, and the number of particles, number of iterations, inertia weight, and learning factor of the particle swarm algorithm are readjusted to improve the algorithm's optimization ability.

[0037] The priority matching module assigns different scheduling priorities based on the examinee's age, health status, and appointment type. The priorities are divided into three levels: high, medium, and low. Elderly examinees (age ≥ 70 years), examinees with mobility difficulties, and examinees with emergency needs are given high priority. The scheduling plan prioritizes the execution order of fasting and emergency items to shorten their overall waiting time. Examinees with regular appointments are given medium priority and are executed according to the regular scheduling rules. Examinees in groups are given low priority and are scheduled without affecting examinees with high and medium priority.

[0038] The scheduling plan generated by the dynamic scheduling decision module is pushed to the examinee through the examinee's terminal. The examinee's terminal displays the current project progress, the location of the next project department, and the estimated waiting time in real time. It also supports examinees to adjust the execution order of non-constrained projects. The adjusted plan must be verified by the constraint verification unit and can only be executed after it meets the constraint requirements.

[0039] The method for dynamic resource intelligent scheduling of cross-departmental physical examination services in this embodiment includes the following steps: Step 1: The resource capability profile building module collects static attributes and dynamic operational data of each department in real time according to the hierarchical collection strategy, and generates a multi-dimensional resource capability profile. The profile data is continuously updated according to the hierarchical cycle.

[0040] Step 2: The project dependency graph construction module parses the constraints of the physical examination projects and generates a project dependency graph. The graph data is updated synchronously when the project rules are updated.

[0041] Step 3: The priority adaptation module assigns scheduling priorities based on the examinee's age, health status, and appointment type.

[0042] Step 4: The dynamic scheduling decision module receives the list of items for the examinee, and combines the resource capability profile, item dependency graph and scheduling priority to generate an initial scheduling scheme using an improved particle swarm algorithm.

[0043] Step 5: Push the scheduling plan to the examinee's terminal to display the project progress and estimated waiting time in real time.

[0044] Step 6: The execution monitoring module monitors the execution status of the scheduling plan in real time based on the check-in and attendance records, and triggers local rescheduling when an abnormal event occurs.

[0045] Step 7: The feedback optimization module collects scheduling execution data, calculates efficiency scores using a normalization method, iteratively optimizes scheduling algorithm parameters, and continuously improves the efficiency of subsequent scheduling schemes.

[0046] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0047] 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 dynamic resource intelligent scheduling system for cross-departmental physical examination services, characterized in that, include: The resource capability profile building module is used to collect static attribute data and dynamic operation data of each department to build a multi-dimensional resource capability profile that includes equipment status, personnel qualifications, real-time load, historical efficiency, and spatial location. The project dependency graph construction module is used to parse the medical constraints, temporal constraints, and cross-departmental collaborative constraints of physical examination items, and construct a project dependency graph for physical examination items. The dynamic scheduling decision module is used to receive the project list of the examinee and generate an initial scheduling scheme by combining the resource capability profile and project dependency graph with a multi-objective optimization algorithm. The execution monitoring module is used to collect resource status data and project progress data in real time during the execution of the scheduling plan, monitor abnormal events, and trigger dynamic adjustments to the scheduling plan. The feedback optimization module is used to collect efficiency data and resource utilization data after the scheduling scheme is executed, and to iteratively optimize the algorithm parameters of the dynamic scheduling decision module.

2. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The resource capability profiling module includes: The static attribute acquisition unit is used to collect information such as the model and quantity of equipment in the department, the qualification level of medical staff, and the spatial location of the department. The dynamic data monitoring unit is used to collect equipment operating status, current number of people in queue, average execution time of a single project, and on-duty status of medical staff in real time according to a hierarchical collection strategy. Among them, equipment status data is collected on a minute-level cycle, while queue and personnel status data are collected on a second-level cycle. The profile generation unit is used to integrate static attribute data with dynamic data to generate a capability score matrix for each department's resources. The capability score matrix includes four dimensions: equipment availability, personnel matching degree, load saturation, and location convenience. The location convenience is calculated based on the distance between the department location corresponding to the examinee's most recent check-in record and the target department.

3. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The project dependency graph construction module includes: The constraint parsing unit is used to identify fasting constraints, pre- and post-dependent constraints, and cross-departmental collaborative constraints for physical examination items. Fasting constraints mark the attributes of items that need to be completed on an empty stomach, pre- and post-dependent constraints mark the execution order requirements between items, and cross-departmental collaborative constraints mark related items that need to be completed in sequence within a preset time window. The graph generation unit is used to transform constraint relationships into a directed acyclic graph. Each node corresponds to a physical examination item, and each edge corresponds to the constraint relationship between items. The edge weight is the constraint priority. Cross-departmental collaborative constraints are expressed by setting directed edges with time window labels between related item nodes.

4. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The dynamic scheduling decision module includes: The constraint verification unit is used to verify whether the initial scheduling scheme conforms to the constraint relationships in the project dependency graph and to eliminate scheduling paths that do not conform to the constraints. The multi-objective optimization unit takes the shortest total waiting time for examinees, the highest departmental load balance, and the maximum resource utilization as its objective functions. It uses an improved particle swarm optimization algorithm to solve for the optimal scheduling path and generate an initial scheduling scheme. The improvements to the particle swarm optimization algorithm include: introducing a constraint violation penalty function into the fitness function, encoding high-priority constraints in the project dependency graph as penalty terms, and adding penalty values ​​to the fitness value when a scheduling path violates a high-priority constraint. The degree of violation of the fasting constraint is quantified by the length of time elapsed after eating, with a larger penalty value for longer time. An adaptive inertia weight strategy is adopted, with the inertia weight decreasing linearly with the number of iterations. A Gaussian mutation operator is introduced, applying a Gaussian perturbation to the current optimal particle position when the global optimal solution is not updated after a preset number of consecutive iterations.

5. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The execution monitoring module includes: The status acquisition unit is used to collect real-time data on changes in the number of people queuing in each department, equipment malfunction status, temporary absence status of medical staff, and check-in records of examinees. Anomaly identification unit is used to identify four types of abnormal events: equipment failure, staff leaving their posts, sudden increase in the number of people in the queue, and medical examinees staying in the queue. Among them, medical examinees staying in the queue is determined by the fact that a medical examinee has not signed in to the next department after a preset time. Sudden increase in the number of people in the queue is determined by the fact that the number of people in the queue exceeds a preset proportion of the department's maximum capacity for receiving patients. The maximum capacity for receiving patients is the maximum number of examinations that can be completed by all equipment and personnel in the department per unit time. The dynamic adjustment unit triggers local rescheduling when an abnormal event is detected, adjusting only the execution order and corresponding departments of the affected projects, while keeping the scheduling scheme of the unaffected projects unchanged.

6. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 5, characterized in that, In response to equipment malfunctions, the dynamic adjustment unit automatically reassigns affected items to nearby departments with similar equipment. These nearby departments are the closest in physical location to the department with the same examination function. In response to staff absences, the unit reassigns items to departments with qualified medical staff on duty. In response to a sudden increase in the number of people in the queue, the unit adjusts the corresponding items for subsequent examinees to similar departments with lower workloads. In response to examinees experiencing prolonged periods of congestion, the unit re-plans the order of subsequent items for that examinee.

7. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The feedback optimization module includes: The data acquisition unit is used to collect the total time spent by examinees, the average load rate of each department, the equipment idle rate, and the number of abnormal adjustments for each scheduling plan. The effectiveness evaluation unit compares the collected data with historical scheduling data and calculates the efficiency score of the current scheduling plan using a normalization method. The efficiency score is calculated as follows: Efficiency score = 0.4 × max(0,1 - total time spent by examinees / historical average total time spent) + 0.3 × max(0,1 - standard deviation of load rate of each department / historical average standard deviation) + 0.3 × max(0,1 - equipment idle rate / historical average equipment idle rate), where the max function ensures that the score of each item is not lower than 0, and the historical average is the moving average of the past 30 natural days. The parameter optimization unit is used to adjust the weight coefficients of each objective function in the multi-objective optimization unit according to the efficiency score. When the efficiency scores of multiple consecutive scheduling schemes are lower than the threshold, global optimization of algorithm parameters is triggered.

8. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, It also includes a priority adaptation module, which assigns different scheduling priorities based on the examinee's age, health status, and appointment type. The scheduling scheme for high-priority examinees prioritizes the execution order of fasting and emergency procedures.

9. The intelligent scheduling system for dynamic resources of cross-departmental physical examination services according to claim 1, characterized in that, The scheduling plan generated by the dynamic scheduling decision module is pushed to the examinee through the terminal. The terminal displays the current project progress, the location of the next project department, and the estimated waiting time in real time. It also supports the examinee to adjust the execution order of non-constraint projects. The adjusted plan can only be executed after it has been verified by the constraint verification unit.

10. A method for intelligent scheduling of dynamic resources for cross-departmental physical examination services based on the system described in any one of claims 1-9, characterized in that, Includes the following steps: Step 1: The resource capability profile building module collects static attributes and dynamic operational data of each department in real time according to the hierarchical collection strategy, and generates a multi-dimensional resource capability profile. Step 2: The project dependency graph construction module parses the constraints of the physical examination items and generates a project dependency graph; Step 3: The priority matching module assigns scheduling priorities based on the examinee's age, health status, and appointment type; Step 4: The dynamic scheduling decision module receives the list of items for the examinee, and combines the resource capability profile, item dependency graph and scheduling priority to generate an initial scheduling scheme using an improved particle swarm algorithm; Step 5: Push the scheduling plan to the examinee's terminal to display the project progress and estimated waiting time in real time; Step 6: The monitoring module monitors the execution status of the scheduling plan in real time based on the check-in records, and triggers local rescheduling when an abnormal event occurs; Step 7: The feedback optimization module collects scheduling execution data, iteratively optimizes scheduling algorithm parameters, and improves the efficiency of subsequent scheduling schemes.