Remote medical service task scheduling method and system for patients with brain diseases

Abstract

The invention discloses a brain disease patient telemedicine service task scheduling method and system, and relates to the technical field of task scheduling. A remote medical service task scheduling system for patients with brain diseases comprises a brain disease condition assessment module, a brain disease medical service scheduling module and a brain disease medical service conflict module. According to the invention, by monitoring the physiological data of the brain disease patient in real time and utilizing the advanced medical service evaluation model, the accurate evaluation of the physiological status of the patient and the rapid matching of the remote medical service are realized; medical resources can be efficiently screened and dispatched through a stacked graph convolutional layer and an improved NSGAII algorithm, and it is ensured that a patient obtains timely medical services; the medical resource configuration is optimized, the medical service experience of the patient is improved, and more accurate and efficient remote medical service is provided for the brain disease patient.

Description

Technical Field

[0001] The present invention relates to the technical field of task scheduling, and in particular to a method and system for scheduling remote medical service tasks for patients with brain diseases. Background Art

[0002] Generally speaking, telemedicine services prioritize tasks based on their urgency and importance, with the most urgent tasks being prioritized. For patients with brain diseases, those in the acute phase, those with severe symptoms, or those in life-threatening situations are usually given priority. However, in actual operations, although some hospitals or medical teams are able to centrally manage a certain number of patients, traditional resource allocation methods often lack intelligent optimization, which can lead to delays in low-priority tasks or uneven resource allocation. To improve efficiency, intelligent scheduling methods are needed to automatically allocate appropriate medical resources, such as doctors, monitoring equipment, and treatment time, to different patients. At the same time, they should also have predictive capabilities to anticipate potential acute events, allowing for early intervention and optimized resource scheduling. Summary of the Invention

[0003] The present invention aims to provide a method and system for scheduling telemedicine service tasks for patients with brain diseases, so as to reasonably schedule and allocate telemedicine services.

[0004] A method for scheduling remote medical service tasks for patients with brain diseases, comprising the following steps: Acquire real-time physiological data of brain disease patients; analyze the real-time physiological data and the medical service evaluation model for brain disease patients to obtain physiological condition indicators of brain disease patients, brain disease telemedicine service signals and preliminary screening conditions; The medical service evaluation model for brain disease patients includes a real-time physiological parameter analysis layer, a historical indicator analysis layer, a medical service evaluation layer, and a medical service fast matching layer. It is constructed in conjunction with a stacked graph convolution layer to perform real-time physiological condition assessment on brain disease patients. After receiving the brain disease telemedicine service signal, obtain all the medical resources to be screened that meet the preliminary screening conditions H i , i=1, 2, ..., I, I is the total number of all medical resources to be screened; medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Based on the medical resources to be screened H i , screening the physiological condition indicators of brain disease patients and the brain disease telemedicine service scheduling model to obtain a telemedicine service scheduling result; providing medical services to the brain disease patients corresponding to the brain disease telemedicine service signals sent based on the telemedicine service scheduling result; The telemedicine service scheduling model for brain diseases includes a medical resource analysis layer, a telemedicine service decision layer, and a telemedicine service scheduling layer. In the telemedicine service decision layer, the improved NSGAII algorithm is used to perform multi-objective screening and task scheduling. When a telemedicine service task scheduling conflict occurs, the conflicting telemedicine service scheduling task is obtained; the conflicting telemedicine service scheduling task includes conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; the number of conflicting brain disease patient medical tasks is two or more; task scheduling is performed based on the conflicting telemedicine service scheduling task and the telemedicine service task conflict model, and medical resources are reallocated for the conflicting telemedicine service scheduling task; The telemedicine service task conflict model includes a medical service demand assessment layer, a medical service conflict redistribution layer and a medical service task conflict scheduling layer, and is constructed based on the BP neural network model.

[0005] As a preferred technical solution of the present invention, the medical service evaluation model for patients with brain diseases includes a physiological parameter real-time analysis layer, a historical indicator analysis layer, a medical service evaluation layer, and a medical service rapid matching layer; The physiological parameter real-time analysis layer is used to pre-process the real-time physiological data to obtain pre-processed real-time physiological data; perform feature extraction on the pre-processed real-time physiological data to obtain features of the pre-processed real-time physiological data; The historical indicator analysis layer is used to obtain historical medical service data corresponding to patients with brain diseases; extract features from the historical medical service data to obtain the features of the historical medical service data; The medical service evaluation layer is used to perform feature matching based on pre-processed real-time physiological data features and historical medical service data features to obtain physiological status indicators of brain disease patients; The medical service rapid matching layer is used to conduct quantitative evaluation based on the physiological condition indicators of brain disease patients to obtain brain disease telemedicine service signals; and to match preliminary screening conditions based on the physiological condition indicators of brain disease patients.

[0006] As a preferred technical solution of the present invention, the specific steps of training the medical service evaluation layer include: Construct 2N graph convolution layers G k As the basic structure of the medical service evaluation layer, k=1, 2, …, n, …, 2N; The graph convolution layer G 2n-1 and graph convolution layer G 2n Parallelize to get the merged graph convolution layer G m , m=1,2,…,N, n=1,2,…,N; where the graph convolution layer G 2n-1 Represented as 2N graph convolution layers G k The 2n-1th graph convolution layer in the graph convolution layer G2n Represented as 2N graph convolution layers G k The 2nth graph convolutional layer in ; Merge all graph convolution layers G m Stacking is performed to obtain the initial medical service evaluation layer; Collecting several sets of continuous physiological condition indicator training samples; each set of physiological condition indicator training samples includes an initial physiological parameter map and a predicted physiological parameter map; using the predicted physiological parameter map as a target value in the physiological condition indicator training samples; combining several sets of continuous physiological condition indicator training samples to obtain a physiological condition indicator training set; Inputting the physiological condition indicator training set into the initial medical service evaluation layer for training with the target value as the target to obtain a pre-trained medical service evaluation layer; performing model evaluation on the pre-trained medical service evaluation layer to obtain a model evaluation result of the pre-trained medical service evaluation layer; if the model evaluation result of the pre-trained medical service evaluation layer is passed, the pre-trained medical service evaluation layer is used as the medical service evaluation layer in the medical service evaluation model for brain disease patients; otherwise, continuing model training using the physiological condition indicator training set; In the medical service evaluation layer, the pre-processed real-time physiological data features and historical medical service data features are used to construct a graph structure to obtain a heterogeneous graph of physiological parameters of brain disease patients; the node features of the heterogeneous graph of physiological parameters of brain disease patients are learned to obtain the relationship weights of physiological parameters of brain disease patients; aggregation is performed based on the relationship weights of physiological parameters of brain disease patients to obtain the physiological state embedding vector of brain disease patients; predictive analysis is performed based on the physiological state embedding vector of brain disease patients to obtain physiological condition indicators of brain disease patients.

[0007] As a preferred technical solution of the present invention, the telemedicine service scheduling model for brain diseases includes a medical resource analysis layer, a telemedicine service decision layer and a telemedicine service scheduling layer; The medical resource analysis layer is used to extract features based on the physiological condition indicators of brain disease patients and obtain the characteristics of patients receiving telemedicine services; The telemedicine service decision layer is used to determine the characteristics of telemedicine patients and all medical resources to be screened. i Perform multi-objective optimization screening to obtain the telemedicine service scheduling results; multi-objective optimization screening is combined with the improved NSGAII algorithm to construct; The telemedicine service scheduling layer is used to output telemedicine service scheduling results.

[0008] As a preferred technical solution of the present invention, the specific steps of performing multi-objective optimization screening in the telemedicine service decision-making layer include: Construct Q medical service decision-making individuals C q ,q=1,2,…,Q; each medical service decision-making individual Cq There is a set of solutions for telemedicine service scheduling; Q medical service decision individuals C q Combine and obtain the medical service decision initialization population; set the maximum number of iterations, initial mutation probability and SAA initial parameters; Calculate the individual C of medical service decision q The fitness of q , specific steps: Using the function F(C q )=W1*min(max(Tq))+W2*min(L q )+W3*min(R q ) as the medical service decision optimization function F(C q );F(C q ) is taken as the fitness S q ; Among them, T q Indicates the estimated completion time, L q Represents the expected dispatch distance, R q Indicates the number of resource scheduling times; W1, W2 and W3 are preset parameter weight values; Perform population iteration on the medical service decision initialization population and retain the optimal Pareto medical service decision solution set; During population iteration, the elite retention strategy is used to select the initial population of medical service decision-making, and the initial population of medical service decision-making is cross-recombined according to the initial mutation probability to obtain an updated initial population of medical service decision-making. The SAA initial parameters are used to perform a local search on the optimal Pareto medical service decision solution set to obtain a new optimal Pareto medical service decision solution set. When the maximum number of iterations is reached, the current optimal Pareto medical service decision solution set and the medical service decision individual with the maximum fitness in the updated medical service decision initialization population are output, which is the optimal medical service decision individual; based on the optimal medical service decision individual, the telemedicine service scheduling result is obtained.

[0009] As a preferred technical solution of the present invention, the telemedicine service task conflict model includes a medical service demand assessment layer, a medical service conflict redistribution layer and a medical service task conflict scheduling layer; The medical service demand assessment layer is used to prioritize conflicting medical tasks for patients with brain diseases in conflicting telemedicine service scheduling tasks, and obtain the priority of conflicting medical tasks for patients with brain diseases; The medical service conflict reallocation layer is used to input the conflicting brain disease patients' medical tasks with lower priority among the conflicting brain disease patients' medical tasks into the brain disease telemedicine service scheduling model to reschedule the tasks and obtain updated telemedicine service scheduling resources; The medical service task conflict scheduling layer is used to reschedule tasks based on updated telemedicine service scheduling resources and conflicting telemedicine service scheduling resources.

[0010] As a preferred technical solution of the present invention, the specific steps of training the medical service demand assessment layer include: Collecting several sets of training samples of medical task priorities for patients with brain diseases; each set of training samples of medical task priorities for patients with brain diseases contains an evaluated quantitative level of medical tasks and corresponding medical task characteristics of patients with brain diseases; combining the several sets of training samples of medical task priorities for patients with brain diseases to obtain a training set of medical task priorities for patients with brain diseases; using the evaluated quantitative level of medical tasks as a target value; The training set of medical task priorities for patients with brain diseases is input into the telemedicine service task conflict model to train the medical service demand assessment layer with the target value as the target, and obtain the initial medical service demand assessment layer; the initial medical service demand assessment layer is evaluated; if the initial medical service demand assessment layer passes the model evaluation, the initial medical service demand assessment layer is used as the medical service demand assessment layer in the telemedicine service task conflict model; otherwise, the model training is continued using the training set of medical task priorities for patients with brain diseases.

[0011] A remote medical service task scheduling system for patients with brain diseases, comprising: The brain disease condition assessment module includes a data acquisition unit and a data analysis unit; the data acquisition unit is used to obtain real-time physiological data of brain disease patients; the data analysis unit is used to analyze the real-time physiological data and the medical service assessment model for brain disease patients to obtain physiological condition indicators of brain disease patients, brain disease telemedicine service signals and preliminary screening conditions; The brain disease medical service scheduling module includes a task scheduling unit; the task scheduling unit is used to obtain all the medical resources to be screened that meet the preliminary screening conditions after receiving the brain disease remote medical service signal. i , i=1, 2, ..., I, I is the total number of all medical resources to be screened; medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Based on the medical resources to be screened H i , screening the physiological condition indicators of brain disease patients and the brain disease telemedicine service scheduling model to obtain a telemedicine service scheduling result; providing medical services to the brain disease patients corresponding to the brain disease telemedicine service signals sent based on the telemedicine service scheduling result; The brain disease medical service conflict module includes a conflict task acquisition unit and a task evaluation unit; the conflict task acquisition unit is used to obtain conflicting telemedicine service scheduling tasks when a telemedicine service task scheduling conflict occurs; the conflicting telemedicine service scheduling tasks include conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; the number of conflicting brain disease patient medical tasks is two or more; the task evaluation unit is used to perform task scheduling based on the conflicting telemedicine service scheduling tasks and the telemedicine service task conflict model, and reallocate medical resources for the conflicting telemedicine service scheduling tasks.

[0012] The present invention has the following advantages: 1. This invention monitors the physiological data of brain disease patients in real time and utilizes advanced medical service evaluation models to achieve accurate assessment of patients' physiological conditions and rapid matching of telemedicine services. Through stacked graph convolutional layers and an improved NSGAII algorithm, it can efficiently screen and dispatch medical resources to ensure that patients receive timely medical services. It intelligently handles service task conflicts and reallocates resources through a BP neural network model to improve the efficiency and quality of medical services. This not only optimizes the allocation of medical resources, but also enhances the patient's medical service experience, providing more accurate and efficient telemedicine services for patients with brain diseases.

[0013] 2. The present invention combines the characteristics of real-time physiological data with the characteristics of historical medical service data to construct a heterogeneous graph of physiological parameters of patients with brain diseases. This can comprehensively consider data from different sources. This heterogeneous graph structure effectively integrates real-time and historical data, providing the model with more comprehensive patient health information. By learning node features through methods such as graph convolutional networks, it can deeply explore and understand the potential correlation between each physiological parameter and the patient's health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a structural diagram of a remote medical service task scheduling system for brain disease patients adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0016] Example 1, a method for scheduling remote medical service tasks for patients with brain diseases, comprising the following steps: Acquire real-time physiological data from patients with brain diseases; analyze the real-time physiological data and a medical service evaluation model for patients with brain diseases to obtain physiological indicators of patients with brain diseases, signals for remote medical services for brain diseases, and preliminary screening criteria; real-time physiological data is acquired from sensors worn by patients with brain diseases; if the patient actively makes a remote service call, the patient's real-time physiological data is automatically acquired and subsequent steps are taken; The medical service evaluation model for brain disease patients includes a real-time physiological parameter analysis layer, a historical indicator analysis layer, a medical service evaluation layer, and a medical service fast matching layer. It is constructed in conjunction with a stacked graph convolution layer to perform real-time physiological condition assessment on brain disease patients. The physiological parameter real-time analysis layer is used to preprocess real-time physiological data to obtain preprocessed real-time physiological data; perform feature extraction on the preprocessed real-time physiological data to obtain features of the preprocessed real-time physiological data; by preprocessing and extracting features from real-time physiological data through the physiological parameter real-time analysis layer, the patient's current physiological condition can be obtained in a timely and accurate manner, which enables the medical team to quickly respond to the patient's physiological changes, predict possible health problems in advance, and provide more refined telemedicine services; real-time processing and analysis of physiological data ensures that the patient's physiological condition can be continuously tracked and monitored, making it easier to identify potential health risks and take appropriate medical measures; The historical indicator analysis layer is used to obtain historical medical service data corresponding to brain disease patients; feature extraction is performed on historical medical service data to obtain historical medical service data characteristics; through the historical indicator analysis layer, combined with the historical medical service data and treatment records of brain disease patients, it can effectively supplement and improve real-time data analysis, providing historical comparison and support for medical service decision-making; using historical data to provide contextual information for real-time physiological data, it can more accurately assess the patient's health change trends, support future medical interventions, and form more personalized and refined treatment plans; The medical service assessment layer is used to perform feature matching based on pre-processed real-time physiological data features and historical medical service data features to obtain physiological condition indicators for patients with brain diseases. By matching the real-time physiological data features with the historical medical service data features, the medical service assessment layer can accurately assess the patient's health status and ensure that the assessment results are based on comprehensive data rather than relying solely on the current physiological condition. The medical service rapid matching layer is used to perform quantitative assessments based on the physiological indicators of brain disease patients to obtain brain disease telemedicine service signals. Preliminary screening conditions are matched based on the physiological indicators of brain disease patients. Quantitative assessments rely on preset quantitative indicators to send signals, and automatic judgments are made by setting abnormal thresholds. Preliminary screening conditions are obtained by remotely sending brain disease telemedicine service signals and are mainly used to screen nearby idle medical resources. Telemedicine service signals can help medical service platforms make rapid decisions among multiple treatment options, optimize the allocation of medical resources, and improve service efficiency, especially when resources are limited. The specific steps for training the medical service assessment layer include: Construct 2N graph convolution layers G k As the basic structure of the medical service evaluation layer, k=1, 2, …, n, …, 2N; The graph convolution layer G 2n-1 and graph convolution layer G 2n Parallelize to get the merged graph convolution layer G m , m=1,2,…,N, n=1,2,…,N; where the graph convolution layer G 2n-1 Represented as 2N graph convolution layers G k The 2n-1th graph convolution layer in the graph convolution layer G 2n Represented as 2N graph convolution layers G k The 2nth graph convolutional layer in ; Merge all graph convolution layers G m The initial medical service evaluation layer is obtained by stacking them. Every two graph convolution layers are merged in parallel to obtain a merged graph convolution layer. This design reduces the repeated transmission of information between layers and optimizes the computational complexity. By processing in parallel, the amount of computation can be reduced while retaining information diversity. This reduces the redundant information of a single-layer graph convolution, making the feature representation more compact and improving the computational efficiency of the model. This not only accelerates the model training and inference process, but also reduces the consumption of computing resources. The use of stacked graph convolutional networks enables physiological condition assessment to deeply capture the complex relationships and interdependencies between different physiological characteristics. By constructing a heterogeneous graph of physiological parameters of patients with brain diseases, the graph convolutional layer can effectively learn the correlations between nodes, which is particularly important for modeling the multidimensional and nonlinear relationships that often appear in medical data. By stacking multiple graph convolutional layers, the model can perform multi-level and in-depth feature learning on patients' physiological data and historical data, thereby improving the accuracy and reliability of disease prediction. Collecting several sets of continuous physiological condition indicator training samples; each set of physiological condition indicator training samples includes an initial physiological parameter map and a predicted physiological parameter map; using the predicted physiological parameter map as a target value in the physiological condition indicator training samples; combining several sets of continuous physiological condition indicator training samples to obtain a physiological condition indicator training set; Inputting the physiological condition indicator training set into the initial medical service evaluation layer for training with the target value as the target to obtain a pre-trained medical service evaluation layer; performing model evaluation on the pre-trained medical service evaluation layer to obtain a model evaluation result of the pre-trained medical service evaluation layer; if the model evaluation result of the pre-trained medical service evaluation layer is passed, the pre-trained medical service evaluation layer is used as the medical service evaluation layer in the medical service evaluation model for brain disease patients; otherwise, continuing model training using the physiological condition indicator training set; In the medical service evaluation layer, the pre-processed real-time physiological data features and historical medical service data features are used to construct a graph structure to obtain a heterogeneous graph of physiological parameters of brain disease patients. Node features are learned from the heterogeneous graph of physiological parameters of brain disease patients to obtain the relationship weights of physiological parameters of brain disease patients. Based on the relationship weights of physiological parameters of brain disease patients, the physiological state embedding vector of brain disease patients is obtained. Predictive analysis is performed based on the physiological state embedding vector of brain disease patients to obtain physiological status indicators of brain disease patients. By combining the features of real-time physiological data with those of historical medical service data, a heterogeneous graph of physiological parameters for patients with brain diseases is constructed. This allows for comprehensive consideration of data from different sources. This heterogeneous graph structure effectively integrates real-time and historical data, providing the model with more comprehensive patient health information. Learning node features through methods such as graph convolutional networks allows for in-depth exploration and understanding of the potential correlation between each physiological parameter and the patient's health status. For example, physiological signals such as real-time heart rate, blood pressure, and brain waves can be captured through node feature learning to understand their relationship with the patient's health status. After receiving the brain disease telemedicine service signal, obtain all the medical resources to be screened that meet the preliminary screening conditions H i , i=1, 2, ..., I, I is the total number of all medical resources to be screened; medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Medical staff resources P i 、Medical equipment resources B i Random allocation can be performed. In actual use, it may happen that the medical equipment resources at the same location do not meet the needs of brain disease patients, but the medical staff resources do. For example, in service center A, there are 0 idle medical equipment and 5 medical staff; in service center D, there are 5 idle medical equipment and 2 medical staff. In this case, there may be a resource overlap. Based on the medical resources to be screened H i , screening the physiological condition indicators of brain disease patients and the brain disease telemedicine service scheduling model to obtain a telemedicine service scheduling result; providing medical services to the brain disease patients corresponding to the brain disease telemedicine service signals sent based on the telemedicine service scheduling result; By combining the physiological indicators of brain disease patients with medical service signals, it is possible to accurately match them according to the patient's needs and select medical resources to be screened that meet the initial screening criteria. This method can ensure the accurate allocation of medical resources, avoid resource waste, and improve resource utilization; medical staff resources and medical equipment resources can be flexibly and randomly allocated, so that when a certain resource is not available in a specific service center, another resource can still meet the patient's needs. This dynamic resource scheduling can cope with the diversity and uncertainty of medical resources and improve the responsiveness of medical services; by coordinating the allocation of medical staff resources and medical equipment resources, it can effectively avoid situations such as "equipment is idle but medical staff is insufficient" or "medical staff is sufficient but equipment is tight". This flexible scheduling mechanism can intelligently allocate resources among multiple service centers according to demand, improving overall service efficiency; The telemedicine service scheduling model for brain diseases includes a medical resource analysis layer, a telemedicine service decision layer, and a telemedicine service scheduling layer. In the telemedicine service decision layer, the improved NSGAII algorithm is used to perform multi-objective screening and task scheduling. The medical resource analysis layer is used to extract features based on the physiological condition indicators of brain disease patients and obtain the characteristics of patients receiving telemedicine services; The telemedicine service decision layer is used to determine the characteristics of telemedicine patients and all medical resources to be screened. i Perform multi-objective optimization screening to obtain the telemedicine service scheduling results; multi-objective optimization screening is combined with the improved NSGAII algorithm to construct; The telemedicine service scheduling layer is used to output telemedicine service scheduling results; The specific steps for multi-objective optimization screening in the telemedicine service decision-making layer include: Construct Q medical service decision-making individuals C q ,q=1,2,…,Q; each medical service decision-making individual C q There is a set of solutions for telemedicine service scheduling; Q medical service decision individuals C q Combine and obtain the medical service decision initialization population; set the maximum number of iterations, initial mutation probability and SAA initial parameters; the maximum number of iterations, initial mutation probability and SAA initial parameters are set by professional technicians according to actual conditions; Calculate the individual C of medical service decision q The fitness ofq , specific steps: Using the function F(C q )=W1*min(max(Tq))+W2*min(L q )+W3*min(R q ) as the medical service decision optimization function F(C q );F(C q ) is taken as the fitness S q ; Among them, T q Indicates the estimated completion time, L q Represents the expected dispatch distance, R q Indicates the number of resource scheduling times; W1, W2, and W3 are preset parameter weight values; W1, W2, and W3 are set by professional technicians based on actual conditions; Perform population iteration on the medical service decision initialization population and retain the optimal Pareto medical service decision solution set; During population iteration, the elite retention strategy is used to select the initial population of medical service decision-making, and the initial population of medical service decision-making is cross-recombined according to the initial mutation probability to obtain an updated initial population of medical service decision-making. The SAA initial parameters are used to perform a local search on the optimal Pareto medical service decision solution set to obtain a new optimal Pareto medical service decision solution set. When the maximum number of iterations is reached, the current optimal Pareto medical service decision solution set and the medical service decision individual with the maximum fitness in the updated medical service decision initialization population are output, which is the optimal medical service decision individual; based on the optimal medical service decision individual, the telemedicine service scheduling result is obtained; The telemedicine service decision-making layer uses the improved NSGA-II algorithm to calculate the estimated completion time (T q ), dispatch distance (L q ) and resource usage times (R q ) as the optimization objective, and combined with different weights (W1, W2, W3) for comprehensive consideration, a comprehensive multi-objective optimization system is formed. This optimization method ensures that the scheduling results can achieve the best balance among multiple objectives. After multiple iterations, the NSGA-II algorithm can obtain the optimal Pareto solution set, ensuring the balance of decision results among multiple objectives, improving the accuracy and practicality of task scheduling, and meeting the complex needs of brain disease telemedicine. Local search of the optimal Pareto solution set through SAA initial parameters helps to further improve the quality of the solution, avoid the model from falling into the local optimal solution, and ensure that the final scheduling solution is optimized with higher accuracy. This local search method improves the robustness of the model, making the scheduling results highly adaptable under different circumstances. The elite retention strategy ensures that the individuals with the best fitness are not lost during each iteration, thus ensuring the continuity of high-quality solutions and improving the stability and reliability of the algorithm. This is particularly important for complex telemedicine task scheduling models and can ensure that the model maintains high accuracy during multi-objective optimization. Based on the scheduling results of the optimal medical service decision-making individuals, the telemedicine system can quickly respond to patient needs, especially in emergency situations, and provide rapid and efficient services. This optimized response mechanism ensures that patients with brain diseases can receive timely diagnosis and treatment, effectively expanding the coverage of telemedicine services. Through multi-objective optimization screening, conflicts between different scheduling tasks can be avoided, ensuring the rational allocation of resources for each task, avoiding scheduling conflicts caused by duplicate or insufficient resources, and reducing resource waste. When a telemedicine service task scheduling conflict occurs, a conflicting telemedicine service scheduling task is obtained; the conflicting telemedicine service scheduling task includes conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; the number of conflicting brain disease patient medical tasks is two or more; task scheduling is performed based on the conflicting telemedicine service scheduling task and the telemedicine service task conflict model, and medical resources are reallocated for the conflicting telemedicine service scheduling task; the conflicting telemedicine service scheduling task occurs when two different brain disease patients select the same telemedicine service scheduling in the same time period, which is a conflicting telemedicine service scheduling resource; The telemedicine service task conflict model includes a medical service demand assessment layer, a medical service conflict redistribution layer, and a medical service task conflict scheduling layer, and is constructed based on the BP neural network model. The medical service demand assessment layer is used to prioritize conflicting medical tasks for patients with brain diseases in conflicting telemedicine service scheduling tasks, and obtain the priority of conflicting medical tasks for patients with brain diseases; The medical service conflict reallocation layer is used to input the conflicting brain disease patients' medical tasks with lower priority among the conflicting brain disease patients' medical tasks into the brain disease telemedicine service scheduling model to reschedule the tasks and obtain updated telemedicine service scheduling resources; The medical service task conflict scheduling layer is used to reschedule tasks based on updated telemedicine service scheduling resources and conflicting telemedicine service scheduling resources; The specific steps for training the medical service needs assessment layer include: Collecting several sets of training samples of medical task priorities for patients with brain diseases; each set of training samples of medical task priorities for patients with brain diseases contains an evaluated quantitative level of medical tasks and corresponding medical task characteristics of patients with brain diseases; combining the several sets of training samples of medical task priorities for patients with brain diseases to obtain a training set of medical task priorities for patients with brain diseases; using the evaluated quantitative level of medical tasks as a target value; Inputting the priority training set of medical tasks for patients with brain diseases into the telemedicine service task conflict model, the medical service demand assessment layer is trained with the target value as the target to obtain an initial medical service demand assessment layer; performing model evaluation on the initial medical service demand assessment layer, if the initial medical service demand assessment layer passes the model evaluation, the initial medical service demand assessment layer is used as the medical service demand assessment layer in the telemedicine service task conflict model; otherwise, continuing model training using the priority training set of medical tasks for patients with brain diseases; Through the medical service demand assessment layer, the priority of the medical tasks of brain disease patients in the conflicting remote medical service scheduling tasks can be matched to ensure that high-priority tasks are handled first, thereby improving the response speed and efficiency of medical services; the medical service conflict redistribution layer re-inputs the tasks with lower priority into the remote medical service scheduling model to achieve optimal resource allocation, reduce resource waste, and improve service efficiency; the medical service task conflict scheduling layer re-schedules tasks according to the updated resources, so that medical service tasks can be more reasonably allocated, thereby improving the overall efficiency of medical services; by optimizing task scheduling and resource allocation, the quality of medical services can be ultimately improved, ensuring that patients can receive timely and effective treatment and enhancing patient satisfaction.

[0017] Example 2, a remote medical service task scheduling system for patients with brain diseases, see Figure 1 As shown, including: The brain disease condition assessment module includes a data acquisition unit and a data analysis unit; the data acquisition unit is used to obtain real-time physiological data of brain disease patients; the data analysis unit is used to analyze the real-time physiological data and the medical service assessment model for brain disease patients to obtain physiological condition indicators of brain disease patients, brain disease telemedicine service signals and preliminary screening conditions; The brain disease medical service scheduling module includes a task scheduling unit; the task scheduling unit is used to obtain all the medical resources to be screened that meet the preliminary screening conditions after receiving the brain disease remote medical service signal. i , i=1, 2, ..., I, I is the total number of all medical resources to be screened; medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Based on the medical resources to be screened H i, screening the physiological condition indicators of brain disease patients and the brain disease telemedicine service scheduling model to obtain a telemedicine service scheduling result; providing medical services to the brain disease patients corresponding to the brain disease telemedicine service signals sent based on the telemedicine service scheduling result; The brain disease medical service conflict module includes a conflict task acquisition unit and a task evaluation unit; the conflict task acquisition unit is used to obtain conflicting telemedicine service scheduling tasks when a telemedicine service task scheduling conflict occurs; the conflicting telemedicine service scheduling tasks include conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; the number of conflicting brain disease patient medical tasks is two or more; the task evaluation unit is used to perform task scheduling based on the conflicting telemedicine service scheduling tasks and the telemedicine service task conflict model, and reallocate medical resources for the conflicting telemedicine service scheduling tasks.

[0018] It should be understood that those skilled in the art may make improvements or modifications based on the above description, and all such improvements and modifications shall fall within the scope of protection of the appended claims. Any portion of this specification not described in detail is prior art known to those skilled in the art.

Claims

1. A method for scheduling remote medical service tasks for patients with brain diseases, characterized in that: The following steps are involved: Obtain real-time physiological data from patients with brain diseases; Analyze real-time physiological data and the medical service evaluation model for brain disease patients to obtain physiological status indicators of brain disease patients, brain disease telemedicine service signals and preliminary screening conditions; The medical service evaluation model for brain disease patients includes a real-time physiological parameter analysis layer, a historical indicator analysis layer, a medical service evaluation layer, and a medical service fast matching layer. It is constructed in conjunction with a stacked graph convolution layer to perform real-time physiological condition assessment on brain disease patients. After receiving the brain disease telemedicine service signal, obtain all the medical resources to be screened that meet the preliminary screening conditions H i , i=1, 2,…, I, I is the total number of all medical resources to be screened; Medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Based on the medical resources to be screened H i , physiological condition indicators of brain disease patients and brain disease telemedicine service scheduling model are screened to obtain telemedicine service scheduling results; Providing medical services to the brain disease patient corresponding to the brain disease telemedicine service signal based on the telemedicine service scheduling result; The telemedicine service scheduling model for brain diseases includes a medical resource analysis layer, a telemedicine service decision layer, and a telemedicine service scheduling layer. In the telemedicine service decision layer, the improved NSGAII algorithm is used to perform multi-objective screening and task scheduling. When a telemedicine service task scheduling conflict occurs, obtaining the conflicting telemedicine service scheduling task; The conflicting telemedicine service scheduling tasks include conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; The number of conflicting medical tasks for patients with brain diseases is two or more; task scheduling is performed based on conflicting telemedicine service scheduling tasks and telemedicine service task conflict models, and medical resources are reallocated for conflicting telemedicine service scheduling tasks; The telemedicine service task conflict model includes a medical service demand assessment layer, a medical service conflict redistribution layer and a medical service task conflict scheduling layer, and is constructed based on the BP neural network model.

2. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 1, characterized in that: The medical service evaluation model for patients with brain diseases includes a real-time physiological parameter analysis layer, a historical indicator analysis layer, a medical service evaluation layer, and a medical service rapid matching layer. The physiological parameter real-time analysis layer is used to pre-process the real-time physiological data to obtain pre-processed real-time physiological data; Performing feature extraction on the pre-processed real-time physiological data to obtain features of the pre-processed real-time physiological data; The historical indicator analysis layer is used to obtain historical medical service data corresponding to patients with brain diseases; extract features from the historical medical service data to obtain the features of the historical medical service data; The medical service evaluation layer is used to perform feature matching based on pre-processed real-time physiological data features and historical medical service data features to obtain physiological status indicators of brain disease patients; The medical service rapid matching layer is used to conduct quantitative evaluation based on the physiological condition indicators of brain disease patients to obtain brain disease telemedicine service signals; and to match preliminary screening conditions based on the physiological condition indicators of brain disease patients.

3. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 2, characterized in that: The specific steps for training the medical service assessment layer include: Construct 2N graph convolution layers G k As the basic structure of the medical service evaluation layer, k=1, 2, …, n, …, 2N; The graph convolution layer G 2n-1 and graph convolution layer G 2n Parallelize to get the merged graph convolution layer G m , m=1,2,…,N, n=1,2,…,N; where the graph convolution layer G 2n-1 Represented as 2N graph convolution layers G k The 2n-1th graph convolution layer in the graph convolution layer G 2n Represented as 2N graph convolution layers G k The 2nth graph convolutional layer in ; Merge all graph convolution layers G m Stacking is performed to obtain the initial medical service evaluation layer; Collecting several sets of continuous physiological condition indicator training samples; each set of physiological condition indicator training samples includes an initial physiological parameter map and a predicted physiological parameter map; using the predicted physiological parameter map as a target value in the physiological condition indicator training samples; combining several sets of continuous physiological condition indicator training samples to obtain a physiological condition indicator training set; Inputting the physiological condition indicator training set into the initial medical service evaluation layer for training with the target value as the target to obtain a pre-trained medical service evaluation layer; performing model evaluation on the pre-trained medical service evaluation layer to obtain a model evaluation result of the pre-trained medical service evaluation layer; if the model evaluation result of the pre-trained medical service evaluation layer is passed, the pre-trained medical service evaluation layer is used as the medical service evaluation layer in the medical service evaluation model for brain disease patients; otherwise, continuing model training using the physiological condition indicator training set; In the medical service evaluation layer, the pre-processed real-time physiological data features and historical medical service data features are used to construct a graph structure to obtain a heterogeneous graph of physiological parameters of brain disease patients; the node features of the heterogeneous graph of physiological parameters of brain disease patients are learned to obtain the relationship weights of physiological parameters of brain disease patients; aggregation is performed based on the relationship weights of physiological parameters of brain disease patients to obtain the physiological state embedding vector of brain disease patients; predictive analysis is performed based on the physiological state embedding vector of brain disease patients to obtain physiological condition indicators of brain disease patients.

4. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 3, characterized in that: The telemedicine service scheduling model for brain diseases includes medical resource analysis layer, telemedicine service decision layer and telemedicine service scheduling layer; The medical resource analysis layer is used to extract features based on the physiological condition indicators of brain disease patients and obtain the characteristics of patients receiving telemedicine services; The telemedicine service decision layer is used to determine the characteristics of telemedicine patients and all medical resources to be screened. i Perform multi-objective optimization screening to obtain the telemedicine service scheduling results; multi-objective optimization screening is combined with the improved NSGAII algorithm to construct; The telemedicine service scheduling layer is used to output telemedicine service scheduling results.

5. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 4, characterized in that: The specific steps for multi-objective optimization screening in the telemedicine service decision-making layer include: Construct Q medical service decision-making individuals C q ,q=1,2,…,Q; each medical service decision-making individual C q There is a set of solutions for telemedicine service scheduling; Q medical service decision individuals C q Combine and obtain the medical service decision initialization population; set the maximum number of iterations, initial mutation probability and SAA initial parameters; Calculate the individual C of medical service decision q The fitness of q , specific steps: Using the function F(C q )=W1*min(max(Tq))+W2*min(L q )+W3*min(R q ) as the medical service decision optimization function F(C q );F(C q ) is taken as the fitness S q ; Among them, T q Indicates the estimated completion time, L q Represents the expected dispatch distance, R q Indicates the number of resource scheduling times; W1, W2 and W3 are preset parameter weight values; Perform population iteration on the medical service decision initialization population and retain the optimal Pareto medical service decision solution set; During population iteration, the elite retention strategy is used to select the initial population of medical service decision-making, and the initial population of medical service decision-making is cross-recombined according to the initial mutation probability to obtain an updated initial population of medical service decision-making. The SAA initial parameters are used to perform a local search on the optimal Pareto medical service decision solution set to obtain a new optimal Pareto medical service decision solution set. When the maximum number of iterations is reached, the current optimal Pareto medical service decision solution set and the medical service decision individual with the maximum fitness in the updated medical service decision initialization population are output, which is the optimal medical service decision individual; based on the optimal medical service decision individual, the telemedicine service scheduling result is obtained.

6. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 5, characterized in that: The telemedicine service task conflict model includes the medical service demand assessment layer, the medical service conflict redistribution layer and the medical service task conflict scheduling layer; The medical service demand assessment layer is used to prioritize conflicting medical tasks for patients with brain diseases in conflicting telemedicine service scheduling tasks, and obtain the priority of conflicting medical tasks for patients with brain diseases; The medical service conflict reallocation layer is used to input the conflicting brain disease patients' medical tasks with lower priority among the conflicting brain disease patients' medical tasks into the brain disease telemedicine service scheduling model to reschedule the tasks and obtain updated telemedicine service scheduling resources; The medical service task conflict scheduling layer is used to reschedule tasks based on updated telemedicine service scheduling resources and conflicting telemedicine service scheduling resources.

7. A method for scheduling remote medical service tasks for patients with brain diseases according to claim 6, characterized in that: The specific steps for training the medical service needs assessment layer include: Collecting several sets of training samples of medical task priorities for patients with brain diseases; each set of training samples of medical task priorities for patients with brain diseases contains an evaluated quantitative level of medical tasks and corresponding medical task characteristics of patients with brain diseases; combining the several sets of training samples of medical task priorities for patients with brain diseases to obtain a training set of medical task priorities for patients with brain diseases; using the evaluated quantitative level of medical tasks as a target value; The training set of medical task priorities for patients with brain diseases is input into the telemedicine service task conflict model to train the medical service demand assessment layer with the target value as the target, and obtain the initial medical service demand assessment layer; the initial medical service demand assessment layer is evaluated; if the initial medical service demand assessment layer passes the model evaluation, the initial medical service demand assessment layer is used as the medical service demand assessment layer in the telemedicine service task conflict model; otherwise, the model training is continued using the training set of medical task priorities for patients with brain diseases.

8. A remote medical service task scheduling system for patients with brain diseases, characterized in that: The system applies a method for scheduling remote medical service tasks for patients with brain diseases as described in any one of claims 1 to 7, comprising: The brain disease condition assessment module includes a data acquisition unit and a data analysis unit; the data acquisition unit is used to obtain real-time physiological data of brain disease patients; the data analysis unit is used to analyze the real-time physiological data and the medical service assessment model for brain disease patients to obtain physiological condition indicators of brain disease patients, brain disease telemedicine service signals and preliminary screening conditions; The brain disease medical service scheduling module includes a task scheduling unit; the task scheduling unit is used to obtain all the medical resources to be screened that meet the preliminary screening conditions after receiving the brain disease remote medical service signal. i , i=1, 2, ..., I, I is the total number of all medical resources to be screened; medical resources to be screened H i Contains medical staff resources P i 、Medical equipment resources B i ; Based on the medical resources to be screened H i , screening the physiological condition indicators of brain disease patients and the brain disease telemedicine service scheduling model to obtain a telemedicine service scheduling result; providing medical services to the brain disease patients corresponding to the brain disease telemedicine service signals sent based on the telemedicine service scheduling result; The brain disease medical service conflict module includes a conflict task acquisition unit and a task evaluation unit; the conflict task acquisition unit is used to obtain conflicting telemedicine service scheduling tasks when a telemedicine service task scheduling conflict occurs; the conflicting telemedicine service scheduling tasks include conflicting brain disease patient medical tasks and corresponding conflicting telemedicine service scheduling resources; the number of conflicting brain disease patient medical tasks is two or more; the task evaluation unit is used to perform task scheduling based on the conflicting telemedicine service scheduling tasks and the telemedicine service task conflict model, and reallocate medical resources for the conflicting telemedicine service scheduling tasks.

Images