Medical resource intelligent dynamic scheduling method and system based on double disease features

Abstract

The invention provides a medical resource intelligent dynamic scheduling method and system based on double disease features, and the method comprises the steps: carrying out the matching in a preset resource demand tensor according to the information of a patient, and obtaining a standard template value; and according to the inspection index set, the historical index mean value, the historical index standard deviation and the disease type superposition indication variable, constructing an individual correction item, and combining the standard template value, the individual correction item and a preset resource adaptation matrix to obtain a resource use prediction matrix. And constructing a scheduling objective function according to the resource scheduling matrix, the resource use prediction matrix, the individual correction term, the disease superposition indication variable, the resource conflict matrix and the dynamic resource load index. And maximizing the scheduling objective function to obtain an optimized resource scheduling matrix. The scheduling task is screened out from the optimized resource scheduling matrix, an execution command structure is constructed according to the scheduling task, the task scheduling time and the resource metadata, resource scheduling is conducted according to the execution command structure, and intelligent dynamic scheduling of the medical resources is achieved.

Description

Technical Field

[0001] This invention belongs to the field of medical resources, and in particular relates to a method and system for intelligent dynamic scheduling of medical resources based on dual disease characteristics. Background Technology

[0002] In modern large-scale general hospitals and specialized hospitals, with the increasing specialization and informatization of medical services, the efficiency of medical resource allocation has become a crucial factor affecting overall operational levels and emergency response capabilities. This is particularly pronounced in scenarios involving the co-treatment of infectious diseases and liver diseases, where resource allocation conflicts are especially acute. Infectious diseases are characterized by their sudden onset and cluster-like nature; for example, the concentrated admission of patients with dengue fever, COVID-19, and HIV / AIDS within a short period leads to a rapid increase in demand for isolation wards, protective equipment, and medical personnel. Liver diseases, such as chronic hepatitis B, cirrhosis, and liver cancer, are characterized by long disease courses and reliance on specialized equipment (such as artificial livers), requiring continuous occupancy of beds and specialized medical care, exhibiting a stability and cyclicality completely different from infectious diseases. This structural characteristic of "acute outbreak" and "chronic dependence" creates multiple conflicts in resource allocation for hospitals in scenarios involving the co-treatment of these two diseases. On the one hand, the centralized treatment during infectious disease outbreaks often leads to a surge in demand for hospital beds, manpower, and resources, creating an "emergency shortage." On the other hand, liver disease patients experience high equipment occupancy rates and long recovery periods during long-term treatment; improper resource allocation can result in severe treatment delays and resource underutilization. Therefore, how to achieve intelligent and dynamic scheduling of medical resources and efficient allocation of medical resources has become an urgent technical problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to design a method and system for intelligent dynamic scheduling of medical resources based on dual disease characteristics, which can realize intelligent dynamic scheduling of medical resources and efficient allocation of medical resources.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for intelligent dynamic scheduling of medical resources based on dual-disease characteristics, the method comprising: Obtain patient information, and match it in a preset resource demand tensor to obtain standard template values; wherein, the patient information includes a set of test indicators; Based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superimposed indicator variable, an individual correction term is constructed. The standard template value, the individual correction term, and the preset resource adaptation matrix are combined to obtain a resource usage prediction matrix. The scheduling objective function is constructed based on the preset resource scheduling matrix, the resource usage prediction matrix, the individual correction term, the disease superposition indicator variable, the preset resource conflict matrix, and the preset dynamic resource load index; wherein, maximizing the scheduling objective function yields the optimized resource scheduling matrix; Scheduling tasks are selected from the optimized resource scheduling matrix, the task scheduling time of each scheduling task is obtained according to the preset resource preparation time, and the corresponding resource metadata is obtained according to the scheduling task. An execution command structure is constructed based on the scheduling task, the task scheduling time, and the resource metadata, and resource scheduling is performed based on the execution command structure.

[0005] Furthermore, before obtaining patient information and matching it against a preset resource requirement tensor to obtain a standard template value, the method further includes: Constructing the resource requirement tensor and the resource adaptation matrix specifically includes: Obtain information on disease category, resource type, disease stage, and time window; The resource demand tensor is constructed based on the disease category, the resource type, the disease stage, and the time window. A resource adaptation matrix is ​​constructed based on whether the corresponding resource category can be used for the disease category.

[0006] Furthermore, the resource adaptation matrix includes multiple adaptation matrix elements, and the step of constructing the resource adaptation matrix based on whether the corresponding resource category can be used for the disease category includes: Determine whether each of the disease categories can use each of the resource categories. If the determination result is yes, the corresponding adaptation matrix element is a first preset value; if the determination result is no, the corresponding adaptation matrix element is a second preset value.

[0007] Furthermore, after performing resource scheduling based on the execution command structure, the method further includes: If the resource scheduling fails, the corresponding scheduling task is marked as a failed task and the failed task is added to the failed task set. In the next scheduling cycle, an execution command structure is constructed for the failed task, and resource scheduling is performed based on the execution command structure.

[0008] Further, maximizing the scheduling objective function to obtain the optimized resource scheduling matrix includes: Obtain the resource inventory quantity; The resource data required for all the optimized resource scheduling matrices is less than or equal to the resource inventory quantity, and the optimized resource scheduling matrix satisfies the resource adaptation matrix.

[0009] Further, the set of test indicators includes the test indicator value corresponding to each test indicator, and the step of constructing an individual correction term based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease-specific superimposed indicator variable includes: The difference between the test index value corresponding to each test index and the historical index mean is divided by the historical index standard deviation, and then multiplied by the first index value to obtain the first data. The preset resource sensitivity weight is multiplied by the disease-specific overlay indicator variable to obtain the second data. The first data and the second data are added together to obtain the third data. The individual correction term is obtained by summing the third data corresponding to all test indicators.

[0010] Furthermore, the optimized resource scheduling matrix includes multiple scheduling matrix elements, where each element is a first preset value or a second preset value. The step of selecting scheduling tasks from the optimized resource scheduling matrix includes: The optimized resource scheduling matrix selects entries whose elements are the first preset values ​​as the scheduling tasks.

[0011] In a second aspect, the present invention provides an intelligent dynamic scheduling system for medical resources based on dual-disease characteristics, the system comprising: An acquisition unit is used to acquire patient information and match it in a preset resource requirement tensor to obtain a standard template value; wherein, the patient information includes a set of test indicators. The construction unit is used to construct individual correction terms based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superposition indicator variable, and to combine the standard template value, the individual correction terms, and the preset resource adaptation matrix to obtain a resource use prediction matrix. An optimization unit is used to construct the scheduling objective function based on a preset resource scheduling matrix, the resource usage prediction matrix, the individual correction term, the disease superposition indicator variable, a preset resource conflict matrix, and a preset dynamic resource load index; wherein, maximizing the scheduling objective function yields the optimized resource scheduling matrix; The filtering unit is used to filter out scheduling tasks from the optimized resource scheduling matrix, obtain the task scheduling time of each scheduling task according to the preset resource preparation time, and obtain the corresponding resource metadata according to the scheduling task. The scheduling unit is used to construct an execution command structure based on the scheduling task, the task scheduling time, and the resource metadata, and to perform resource scheduling based on the execution command structure.

[0012] In a third aspect of the invention, an electronic device is provided, the electronic device including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the method described in the first aspect above.

[0013] In a fourth aspect of the invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0014] The beneficial technical effects of the present invention are at least as follows: To address the aforementioned issues, this invention provides a method and system for intelligent dynamic scheduling of medical resources based on dual-disease characteristics. Its core lies in establishing four organically linked steps: structured disease-specific resource representation, individualized resource usage prediction, dynamic scheduling optimization, and execution configuration generation. This achieves intelligent management of the entire process from resource demand modeling to scheduling implementation. First, a resource demand tensor is constructed to express the interaction between disease, resources, disease stage, and time window in a four-dimensional tensor form. This systematically characterizes the resource consumption patterns of infectious diseases and liver diseases at different stages and times, and introduces a resource adaptation matrix to reflect infection control constraints and resource exclusivity. Second, based on the disease model combined with individual patient test indicators and disease superposition characteristics, an individualized resource usage prediction matrix is ​​established to generate a sequence of patient resource demands within multiple future time windows, thus achieving a dynamic mapping from group statistics to individual behavior. Third, a multi-objective resource scheduling mechanism is designed. Considering the intensity of predicted demand, individual bias, disease superposition priority, and resource conflict suppression, a multi-constraint optimization model is generated and solved using integer programming to obtain the resource scheduling matrix, ensuring the balance and emergency response capability of different diseases in resource competition. Finally, the scheduling results are transformed into executable configuration instructions, generating a standardized execution command structure that includes patients, resources, time, and preparation actions. This structure is then linked with the hospital information system, equipment system, SPD (Supply, Development, and Production) material system, and manpower scheduling system through interfaces to form an intelligent scheduling closed loop. This invention, through a combination of multi-dimensional modeling, time series prediction, and scheduling task implementation, achieves for the first time coordinated management of infectious diseases and liver diseases within the same resource system. It effectively solves the problems of resource conflicts, scheduling delays, and execution gaps in the context of dual-disease co-treatment, achieving the dual goals of maximizing resource utilization and minimizing emergency response time. Attached Figure Description

[0015] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0016] Figure 1 This is a flowchart of the intelligent dynamic scheduling method for medical resources based on dual disease characteristics provided in the embodiments of this application.

[0017] Figure 2 This is a schematic diagram of the structure of the intelligent dynamic scheduling system for medical resources based on dual disease characteristics provided in this application embodiment. Detailed Implementation

[0018] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0019] Please refer to Figure 1 , Figure 1 This is a flowchart of the intelligent dynamic scheduling method for medical resources based on dual disease characteristics provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0020] Step S101: Obtain patient information and match it in a preset resource demand tensor to obtain standard template values; wherein, the patient information includes a set of test indicators. Step S102: Construct individual correction terms based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superposition indicator variables. Combine the standard template value, the individual correction terms, and the preset resource adaptation matrix to obtain the resource use prediction matrix. Step S103: Construct a scheduling objective function based on a preset resource scheduling matrix, resource usage prediction matrix, individual correction terms, disease-specific overlay indicator variables, a preset resource conflict matrix, and a preset dynamic resource load index; wherein, maximize the scheduling objective function to obtain the optimized resource scheduling matrix; Step S104: Select scheduling tasks from the optimized resource scheduling matrix, obtain the task scheduling time of each scheduling task according to the preset resource preparation time, and obtain the corresponding resource metadata according to the scheduling task. Step S105: Construct an execution command structure based on the scheduling task, task scheduling time, and resource metadata, and perform resource scheduling based on the execution command structure.

[0021] In some embodiments, prior to step S101, a reusable structured model is constructed to express the dynamic demand intensity of different types of resources for different disease categories (infectious diseases and liver diseases) at various treatment stages. This model must possess multidimensional structure, be accessible from existing hospital information system data, and be directly referenced by downstream prediction and scheduling steps. This structure is represented as a four-dimensional resource demand tensor. Its four dimensions represent disease category, resource type, disease stage, and time window, respectively.

[0022] Disease categories ( The classification is based on the admission diagnosis codes extracted from the Hospital Information System (HIS), typically ICD-10 classification codes. Multiple specific codes are aggregated into higher-level disease categories through mapping rules. For example, B17.1 (acute hepatitis C) and B18.1 (chronic hepatitis B) are both classified under "Liver Diseases"; U07.1 (COVID-19) and A90 (dengue fever) are classified under "Infectious Diseases". This mapping table, developed by the hospital's quality control department, has been used in other business processes and is directly reusable.

[0023] Stage of illness ( The identification of cases originates from structured entries and treatment process records in the Electronic Medical Record (EMR) system. Three standard stages are defined: initial consultation / routine stage, moderate / hospitalization stage, and severe / intensive care stage. Specific criteria include whether an ICU bed has been allocated, whether high-frequency testing (such as daily liver function tests) has been initiated, whether artificial liver support equipment has been used, and whether surgery has been completed. For example, a patient with cirrhosis admitted to the ICU post-surgery is automatically marked as being in the severe stage, while dengue fever with high fever and a sudden drop in platelets automatically qualifies as being in the moderate stage. These discrimination rules are based on disease type, and the logical rule engine performs structured parsing of EMR entries.

[0024] Resource types ( The dimensions of the resource are derived from three management systems: hospital supplies, personnel, and equipment, including but not limited to general beds, negative pressure wards, ICU beds, artificial liver equipment, protective clothing, and professional medical staff. Each resource type is assigned a unique resource category during modeling and a resource unit is defined (e.g., 1 artificial liver device, 1 negative pressure bed, 1 attending physician with infectious disease qualifications).

[0025] Time window ( Based on the hospital's shift scheduling cycle, a window value is set to roll over every 8 hours. This time granularity is consistent with the actual scheduling system, avoiding a disconnect in time resolution between modeling and execution.

[0026] With the above four dimensions, each element in the resource demand tensor... Defined as the unit intensity requirement of resources for this disease category at this stage and time point, it is constructed using the following structured expression: ; in, Indicates disease category The resource benchmark factor is derived by normalizing the average daily resource consumption for the disease in historical statistics. For example, the daily consumption of dengue fever beds is 0.75 beds, and the daily ICU occupancy for liver cancer is 1.8 beds. Indicates disease category During the stage of the disease The severity coefficient is derived from the comparison of resource usage intensity at different stages in the treatment pathway database and hospitalization records; Indicates resource type During the stage of the disease The degree of dependence is constructed by extracting usage frequency from equipment systems and scheduling records, such as the probability of calling artificial liver equipment in the critical care stage; Indicates a point in time The resource tension correction factor is set based on the growth rate of outpatient visits for each disease during that time period and the risk level released by the CDC. For example, the correction factor for dengue fever during the peak season from May to October is higher than that during the stable period. This is to reflect the exclusivity of various resources in the spatial dimension (e.g., liver disease patients cannot enter negative pressure wards).

[0027] Furthermore, construct a resource adaptation matrix. The resource adaptation matrix includes multiple adaptation matrix elements. The first preset value is 1, and the second preset value is 0. It determines whether each disease category can use each resource category. If the determination result is yes, the corresponding adaptation matrix element is 1; if the determination result is no, the corresponding adaptation matrix element is 0. As shown in the following formula: ; in, Indicates disease category Are resource types allowed? This matrix is ​​set by hospital infection control rules and physical isolation conditions; for example, artificial livers are only suitable for patients with liver diseases, and negative pressure wards are only suitable for patients with respiratory infectious diseases. In the scheduling algorithm, this matrix serves as a hard constraint to prevent illegal resource allocation.

[0028] Finally, structural model The above methods constitute a disease-driven, resource-sensitive, and time-adjustable four-dimensional expression system. Each model construction or update cycle is set to be executed once a day, and it is directly linked to the data interface of the hospital's disease database, material system, and scheduling system for updates.

[0029] In step S101 of some embodiments, patient information is jointly provided by the Hospital Information System (HIS) and the Laboratory Information System (LIS), specifically including patient diagnostic tags. Stage of the patient's condition Patient's admission time and test index set .in, It is automatically matched through ICD-10 diagnostic codes, such as B18.1 being classified as "liver disease" and A90 being classified as "infectious disease"; The hospital's disease grading module will automatically label the patient's condition, based on criteria including whether they have been admitted to the ICU, whether they have undergone a liver biopsy, and whether they have completed an antiviral treatment course. From the patient's admission registration time. Provided by the LIS system, this data covers liver function (ALT, AST, TBIL) and immune inflammation (CRP, WBC) indicators. These data are typically updated hourly, and the hospital's central platform provides standardized API access. To adapt to the time sensitivity of resource prediction, the latest results within ±4 hours are used as the standard, and are automatically extracted and mapped to these standard values.

[0030] First, call the resource demand tensor. Through the patient , and current time Locate its standard template value in the resource demand tensor. ,in This indicates a rolling forecast period, for example, every 8 hours as a window, with a continuous forecast of 72 hours. Based on the patient's diagnosis label. Stages of the disease With admission time Three pieces of information in the resource demand tensor The corresponding element value is precisely located in the four-dimensional index. For example, if a patient has been diagnosed with liver cancer, then... Mapped to the "liver disease" category; if the patient is currently in a critical postoperative stage, then... Corresponding to the "critical period"; if the admission time The future forecast time window is =8 hours, then the system will read directly The numerical value serves as the standard resource usage intensity for that patient in that future time period. This allows for precise determination of the standard template value based on a triple index of disease category, disease stage, and time window. The standard template value represents the average resource usage level for that disease category at that disease stage under a standard treatment pathway. For example, a patient who has undergone liver cancer surgery... In the critical stage, The average usage intensity of the corresponding artificial liver device in the next 24 hours is likely to be 0.8, indicating that there is an expected 80% usage of this resource.

[0031] In step S102 of some embodiments, the standard template values ​​are individually adjusted. This is achieved by analyzing the patient's set of test indicators. Introduce an individual correction term Used to indicate the patient's demand for resources The potential for additional use. The construction is based on a weighted combination of standardized bias and the cumulative effect of diseases. Specifically, the difference between the test indicator value and the historical mean of each indicator is divided by the historical standard deviation, and then multiplied by the first indicator value to obtain the first data. A preset resource sensitivity weight is multiplied by the disease-related cumulative indicator variable to obtain the second data. The first and second data are added together to obtain the third data. Finally, the third data corresponding to all test indicators are added together to obtain the individual correction term. As shown in the following formula: ; in, Indicates the first The test index value of each test index. This indicates the first [number] in the population with this disease. The historical average of each test indicator. The first in this disease population The historical standard deviation of each test indicator and The data comes from statistical information on hospitalized patients with the same disease over the past 12 months, collected from the hospital's big data platform. Indicators of test The degree of abnormality on resources The sensitivity coefficient to the increase in resource demand. For example, if a hepatitis B patient has an ALT of 280, a historical mean of 70, and a historical standard deviation of 40, then the standardized deviation of this item is 5.25. The contribution value for this item is 0.525, indicating that the patient's relative demand for artificial liver resources is higher than average. Furthermore, to address the unique resource utilization needs of patients with overlapping conditions such as "hepatitis B + dengue fever" or "liver cancer + COVID-19," a disease overlap indicator variable is introduced. If the patient belongs to both categories of disease, then assume =1, otherwise 0. For resource-sensitive weights, such as setting artificial liver equipment to 0.3 and protective clothing to 0.1.

[0032] Next, the standard template values, individual correction terms, and resource adaptation matrix are combined to form a resource usage prediction matrix. : ; in, Indicates the patient Standard template values, For individual correction items. The resource adaptation matrix represents the patient's disease category. Is this resource type allowed? For example, a liver disease patient might set their requirement for protective equipment to 0. This expression constitutes a prediction mechanism that integrates structure-driven, individual-driven, and rule-constrained elements, enabling the generation of resource demand sequences with temporal continuity and legitimacy in highly complex resource usage scenarios.

[0033] Resource usage prediction matrix It is a three-dimensional matrix with dimensions "patient × resource type × time window," representing the occupancy intensity of each patient for each type of resource over several future time windows. This matrix serves as the core input structure for the scheduling algorithm. The matrix is ​​automatically refreshed every 8 hours within the system and supports filtering by disease type, ward, and resource type, providing visualization support for management. As an implicit individual characteristic indicator, it can also be used to assist in priority ranking and severity estimation in scheduling strategies.

[0034] Through the above steps S101 and S102, resource usage time-series prediction is performed for each patient to be scheduled, and a resource usage prediction matrix is ​​output. , indicating the patient In the future Resources within a time period The projected usage intensity. This process maps resource requirements at the disease level to the individual patient level, enabling differentiated scheduling and dynamic allocation. It is particularly suitable for addressing resource shortages and frequent priority conflicts in the context of treating both infectious diseases and liver diseases.

[0035] In step S103 of some embodiments, the aim is to use the resource usage prediction matrix. With individual correction items This is effectively translated into resource scheduling behavior. By designing a multi-objective, multi-constraint scheduling optimization mechanism, a three-dimensional resource scheduling matrix is ​​generated. The matrix clearly identifies patients. Is it within a time period? Allocated resources It can directly interface with hospital bed management, medical staff scheduling, and equipment allocation systems to generate executable resource allocation instructions. Its role is not only to allocate resources, but also to achieve intelligent priority control based on differences in disease structure and individual conditions in situations where infectious diseases and liver diseases are treated simultaneously and resource conflicts frequently occur.

[0036] Resource inventory quantity The system is automatically updated via an interface from the hospital's human resources and equipment management system. Bed availability is exported from the empty bed list in the scheduling system. Equipment resources, such as the number of artificial liver units and PCR testing capacity, are uploaded daily by the equipment usage registration system. Material resources, such as the quantity of protective suits, are synchronized by the SPD intelligent material platform. All resource inventory information is stored and used at a granular level, rolling every 8 hours. Resource conflict matrix. Used to represent resources With resources Are there any usage conflicts within the same time period? For example, PCR laboratories and artificial liver equipment require different levels of air purification and cannot be scheduled simultaneously; or ICU beds and beds for certain types of infectious diseases may have incompatible disinfection intervals. This matrix is ​​jointly defined by the hospital infection control team and the logistics department, with a value of 0 or 1. Dynamic resource load index Used to represent resources in the current scheduling round In time The usage density is defined as: ; Indicates the time period Should resources be allocated? Assigned to patients A value of 1 indicates that the allocation has been completed, and a value of 0 indicates that the allocation has not been completed. The optimization process is dynamically updated, reflecting whether the resource scheduling density is too high, which is an important basis for preventing equipment overload and space congestion.

[0037] The scheduling objective function is constructed using a multi-objective trade-off approach, taking into account the following three aspects: First, priority is given to patients with high predicted resource needs, reflecting a prediction outcome-oriented approach; second, priority is given to patients with large individual deviations from the template, reflecting a tilt mechanism towards severe cases; third, priority is given to scheduling patients with overlapping diseases (such as hepatitis B combined with dengue fever) to prevent them from falling into resource blind spots; and fourth, conflicting resources are suppressed to prevent physical conflicts or overlapping personnel allocations in local scheduling. The scheduling objective function is constructed based on a preset resource scheduling matrix, resource usage prediction matrix, individual correction terms, disease overlap indicator variables, a preset resource conflict matrix, and preset dynamic resource load indicators, as shown in the following formula: ; in, This is the resource scheduling matrix; For resource usage prediction matrix; The individual correction term represents the patient's individual deviation from the resource, and is constructed by linear combination of indicators such as ALT and WBC. This is a disease-specific indicator variable derived from the patient's diagnostic system. If a patient is simultaneously diagnosed with liver disease and an infectious disease, then... =1, otherwise 0; This is a dynamic resource load metric, representing the current load intensity of resources, which is determined by the load in each iteration. Calculation of the cumulative value; This is a predefined resource conflict matrix used to determine whether a resource scheduling should be penalized. , , For the weighting coefficients that can be manually or learned, hospitals can set default values ​​based on resource constraints, such as (0.5, 0.3, 0.6).

[0038] To ensure that the scheduling results meet the actual operational conditions, the following two basic constraints are applied: ; ; in, Represents the resource scheduling matrix. Indicates the quantity of resource inventory. This represents a resource matching matrix. The first constraint ensures that the allocation of any resource within any time period does not exceed the resource inventory. The second constraint prohibits allocating resources to patients whose diseases do not match; for example, protective clothing can only be used for patients with infectious diseases, and artificial livers can only be used for patients with liver diseases. It should be noted that... The value (0 or 1) is relative to Impose hard restrictions. Because It can only take 0 or 1, if =0, this inequality forces Only 0 is allowed, indicating that the resource is prohibited from being used. Patients whose diseases do not match ;like =1 is required to allow resources to be allocated to the patient in scheduling optimization.

[0039] The scheduling objective function is solved using integer linear programming, employing optimizers such as Gurobi or CPLEX deployed in the server background, and executed every 8 hours on a rolling basis, each time calculating the scheduling window for the next 48 hours. The optimization results... The output structure is a three-dimensional scheduling matrix, which is written to the database and can be synchronously accessed by systems such as ward scheduling, ICU bed allocation, and material distribution. It also automatically generates a list of unmatched schedules (e.g., ...). Above the threshold but If the value is 0, the administrator is prompted to review the resource scheduling or request an emergency resource expansion.

[0040] In some embodiments, steps S104 to S105 aim to optimize the resource scheduling matrix. This process parses the data into executable configuration tasks within the hospital, effectively translating intelligent scheduling results into hospital resource deployment instructions. It involves not only the structured decomposition of scheduling data but also the classification and adaptation of execution paths for different resource types, including the specific execution logic for resources such as beds, equipment, supplies, and medical personnel. This step connects to multiple subsystems, including the Hospital Information System (HIS), equipment system, supplies management system (SPD), and staffing scheduling system, and is a crucial step in implementing intelligent scheduling within the hospital's multi-system environment.

[0041] The input data for scheduling is the optimized resource scheduling matrix. Each element with a value of 1 indicates that it should be within the time period. Resources Assigned to patients This matrix is ​​calculated by the scheduling engine and stored in the scheduling database. Its structure is well-defined and facilitates subsequent retrieval. Simultaneously, resource inventory data... Sourced from the hospital's equipment status database, SPD (Supply, Development, Production) platform, and staff scheduling system, this indicates the available quantity of each type of resource within each time period. Resource metadata. Includes the physical location of each type of resource Quantity of resources used Resource preparation time and preparation operation instruction set This information, including patient basic information, is maintained uniformly by the hospital's asset management system or resource master data system. Tags including patient ID, disease type, current department, and priority are used to verify the legitimacy of matching resource type with patient.

[0042] The system first selects entries with an element of 1 from the optimized resource scheduling matrix as scheduling tasks, and then constructs a scheduling task list. This indicates which patients will use which resources at what time. Each scheduling task... Each task requires preparatory actions before execution, such as bed cleaning, equipment debugging, and material distribution. Therefore, the trigger time for the preparation process should be pre-calculated for each scheduling task. ; in, It is a resource Preparation time, in minutes, is derived from [source missing]. The table defines the resource's time frame, such as 90 minutes for an artificial liver, 30 minutes for an ICU bed, and 0 for nursing staff. These time points will trigger the system's task execution engine to preload execution instructions, ensuring timely availability of resources. Task scheduling time. The task scheduling time is the time when the resources are officially put into use, plus the resource preparation time, which equals the time when the preparation process is triggered. The preparation process is triggered when the resource preprocessing process begins, such as bed cleaning, equipment preheating, or material picking. Preparation process and task scheduling time Once determined, a standardized execution command structure is constructed based on the scheduling task, task scheduling time, and resource metadata. Its format is defined as follows: ; in, This refers to standardized execution instructions generated by the scheduling system, used to invoke corresponding resources and trigger necessary preparatory operations within a specified time period; It is a resource The physical location is used to specify a particular bed number, equipment number, or nurse number; This indicates the quantity of resources used, such as 2 sets of protective clothing per shift, 1 nurse, 1 piece of equipment, and 1 bed. This is a set of pre-defined operation instructions, such as "material sorting," "ward reservation," and "equipment status verification," which are preset according to resource types, with different items for different resources. Execution instructions are written to the middleware in a structured format (JSON or API form), and various subsystems identify and process them based on the resource type. For example, the artificial liver equipment platform identifies the equipment number and registers its occupancy, the SPD system automatically generates consumable picking tasks, and the bed system marks the bed as "reserved, awaiting cleaning."

[0043] For example, the scheduling task is ( =Li Ming, =Artificial liver, =2024-11-05 08:00), the system will call The system obtains the location "5th Floor, Zone E of the Liver Disease Center", the preparation process "Equipment Preheating", "Catheter Preparation", and "Personnel Arrangement", calculates the preparation start time as "2024-11-05 06:30", writes it into the system scheduling bus, and notifies the relevant subsystems to execute resource scheduling according to the process.

[0044] If a subsystem reports a failure during task execution, such as due to scheduling conflicts, equipment malfunction, or shortage of consumables, the system will mark the task status as "failed," record the failure type, failed resources, and failure time, and add it to the failed task set. Simultaneously, in the next scheduling cycle, the failed task is re-added to the allocation pool, and an execution command structure is constructed based on the failed task. Resource scheduling is then re-performed according to the execution command structure. Failure information will be fed back to the visual management terminal, allowing schedulers to adjust weight parameters and manually intervene in resource status, thereby improving the stability of subsequent scheduling.

[0045] Steps S101 to S105 of this embodiment involve acquiring patient information and matching it against a preset resource demand tensor to obtain a standard template value. The patient information includes a set of test indicators. An individual correction term is constructed based on the test indicator set, preset historical indicator mean, preset historical indicator standard deviation, and preset disease-specific overlay indicator variables. The standard template value, individual correction term, and preset resource adaptation matrix are combined to obtain a resource usage prediction matrix. A scheduling objective function is constructed based on the preset resource scheduling matrix, resource usage prediction matrix, individual correction term, disease-specific overlay indicator variables, preset resource conflict matrix, and preset dynamic resource load indicators. Maximizing the scheduling objective function yields an optimized resource scheduling matrix. Scheduling tasks are selected from the optimized resource scheduling matrix. The task scheduling time for each task is obtained based on a preset resource preparation time, and corresponding resource metadata is acquired based on the scheduling task. An execution command structure is constructed based on the scheduling task, task scheduling time, and resource metadata. Resource scheduling is performed based on the execution command structure, achieving intelligent dynamic scheduling of medical resources and efficient allocation of medical resources.

[0046] Please see Figure 2 This application also provides an intelligent dynamic scheduling system for medical resources based on dual-disease characteristics, which can realize the above-mentioned intelligent dynamic scheduling method for medical resources based on dual-disease characteristics. The system includes: The acquisition unit 201 is used to acquire patient information and match it in a preset resource requirement tensor to obtain standard template values; wherein, the patient information includes a set of test indicators. Construction unit 202 is used to construct individual correction terms based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superposition indicator variable. The standard template value, the individual correction terms, and the preset resource adaptation matrix are combined to obtain the resource use prediction matrix. The optimization unit 203 is used to construct a scheduling objective function based on a preset resource scheduling matrix, a resource usage prediction matrix, individual correction terms, disease-related superposition indicator variables, a preset resource conflict matrix, and a preset dynamic resource load index; wherein, maximizing the scheduling objective function yields the optimized resource scheduling matrix; The filtering unit 204 is used to filter out scheduling tasks from the optimized resource scheduling matrix, obtain the task scheduling time of each scheduling task according to the preset resource preparation time, and obtain the corresponding resource metadata according to the scheduling task. The scheduling unit 205 is used to construct an execution command structure based on the scheduling task, task scheduling time, and resource metadata, and to perform resource scheduling based on the execution command structure.

[0047] The specific implementation of the intelligent dynamic scheduling system for medical resources based on dual disease characteristics is basically the same as the specific implementation of the intelligent dynamic scheduling method for medical resources based on dual disease characteristics described above, and will not be repeated here.

[0048] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for intelligent dynamic scheduling of medical resources based on dual-disease characteristics, characterized in that, The method includes: Obtain patient information, and match it in a preset resource demand tensor to obtain standard template values; wherein, the patient information includes a set of test indicators; Based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superimposed indicator variable, an individual correction term is constructed. The standard template value, the individual correction term, and the preset resource adaptation matrix are combined to obtain a resource usage prediction matrix. The scheduling objective function is constructed based on the preset resource scheduling matrix, the resource usage prediction matrix, the individual correction term, the disease superposition indicator variable, the preset resource conflict matrix, and the preset dynamic resource load index; wherein, maximizing the scheduling objective function yields the optimized resource scheduling matrix; Scheduling tasks are selected from the optimized resource scheduling matrix, the task scheduling time of each scheduling task is obtained according to the preset resource preparation time, and the corresponding resource metadata is obtained according to the scheduling task. An execution command structure is constructed based on the scheduling task, the task scheduling time, and the resource metadata, and resource scheduling is performed based on the execution command structure.

2. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 1, characterized in that, Before obtaining patient information and matching it against a preset resource requirement tensor to obtain a standard template value, the method further includes: Constructing the resource requirement tensor and the resource adaptation matrix specifically includes: Obtain information on disease type, resource type, disease stage, and time window; The resource demand tensor is constructed based on the disease category, the resource type, the disease stage, and the time window. A resource adaptation matrix is ​​constructed based on whether the corresponding resource category can be used for the disease category.

3. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 2, characterized in that, The resource adaptation matrix includes multiple adaptation matrix elements. Constructing the resource adaptation matrix based on whether the corresponding resource category can be used for the disease category includes: Determine whether each of the disease categories can use each of the resource categories. If the determination result is yes, the corresponding adaptation matrix element is a first preset value. If the determination result is no, the corresponding adaptation matrix element is a second preset value.

4. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 1, characterized in that, After performing resource scheduling based on the execution command structure, the method further includes: If the resource scheduling fails, the corresponding scheduling task is marked as a failed task and the failed task is added to the failed task set. In the next scheduling cycle, an execution command structure is constructed for the failed task, and resource scheduling is performed based on the execution command structure.

5. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 1, characterized in that, Maximizing the scheduling objective function to obtain the optimized resource scheduling matrix includes: Obtain the resource inventory quantity; The resource data required for all the optimized resource scheduling matrices is less than or equal to the resource inventory quantity, and the optimized resource scheduling matrix satisfies the resource adaptation matrix.

6. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 1, characterized in that, The set of test indicators includes the test indicator value corresponding to each test indicator. The step of constructing an individual correction term based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease-specific superimposed indicator variable includes: The difference between the test index value corresponding to each test index and the historical index mean is divided by the historical index standard deviation, and then multiplied by the first index value to obtain the first data. The preset resource sensitivity weight is multiplied by the disease-specific overlay indicator variable to obtain the second data. The first data and the second data are added together to obtain the third data. The individual correction term is obtained by summing the third data corresponding to all test indicators.

7. The intelligent dynamic scheduling method for medical resources based on dual disease characteristics according to claim 1, characterized in that, The optimized resource scheduling matrix includes multiple scheduling matrix elements, where each element is a first preset value or a second preset value. The step of selecting scheduling tasks from the optimized resource scheduling matrix includes: The optimized resource scheduling matrix selects entries whose elements are the first preset values ​​as the scheduling tasks.

8. A medical resource intelligent dynamic scheduling system based on dual-disease characteristics, characterized in that, The system includes: An acquisition unit is used to acquire patient information and match it in a preset resource requirement tensor to obtain a standard template value; wherein, the patient information includes a set of test indicators. The construction unit is used to construct individual correction terms based on the set of test indicators, the preset historical indicator mean, the preset historical indicator standard deviation, and the preset disease superposition indicator variable, and to combine the standard template value, the individual correction terms, and the preset resource adaptation matrix to obtain a resource use prediction matrix. An optimization unit is used to construct the scheduling objective function based on a preset resource scheduling matrix, the resource usage prediction matrix, the individual correction term, the disease superposition indicator variable, a preset resource conflict matrix, and a preset dynamic resource load index; wherein, maximizing the scheduling objective function yields the optimized resource scheduling matrix; The filtering unit is used to filter out scheduling tasks from the optimized resource scheduling matrix, obtain the task scheduling time of each scheduling task according to the preset resource preparation time, and obtain the corresponding resource metadata according to the scheduling task. The scheduling unit is used to construct an execution command structure based on the scheduling task, the task scheduling time, and the resource metadata, and to perform resource scheduling based on the execution command structure.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the intelligent dynamic scheduling method for medical resources based on dual disease characteristics as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent dynamic scheduling method for medical resources based on dual disease characteristics as described in any one of claims 1 to 7.

Images