Scheduling method and system under hospital training task, storage medium and equipment

Abstract

The invention discloses a scheduling method and system under a hospital training task, a storage medium and equipment, and relates to the technical field of scheduling, and the method comprises the steps: collecting medical personnel data, hospital data, and determinacy data and uncertainty data in a scheduling rule; using a neural network to predict the uncertainty data; based on the deterministic data and the predicted uncertainty data, taking professional skill matching and workload balancing of medical staff as algorithm targets, and using a reinforcement learning algorithm to generate a scheduling scheme; detecting conflicts in the scheduling scheme, and providing a solution; analyzing the scheduling scheme data, generating a feedback value of the algorithm target, and optimizing the scheduling scheme according to the feedback value of the algorithm target; and sending a scheduling notification, confirmation information and change reminding to the user. The scheduling efficiency, the scheduling reasonability and the scheduling balance can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of shift scheduling, and in particular to a shift scheduling method, system, storage medium and equipment for hospital training tasks. Background Art

[0002] The scheduling system under hospital training is a complex system designed specifically to support the training needs of medical staff. It not only meets the scheduling needs of daily medical services, but also ensures that medical staff can receive adequate training to improve their professional skills and the quality of medical services. The system needs to integrate the training needs of medical staff, including but not limited to basic medical knowledge, professional skills training, clinical practice, emergency response training, etc. This means that the scheduling system must not only consider the personnel needs of medical services, but also reserve time for medical staff to participate in training. The scheduling system should be flexible and be able to dynamically adjust shifts according to the training progress and needs of medical staff. For example, new medical staff may need more time for basic training, while experienced medical staff may need to participate in more advanced clinical training or seminars. The system needs to make effective use of training resources, such as senior doctors as mentors, to ensure that each training link is fully supported. By optimizing resource allocation, resource waste can be avoided and training efficiency can be improved. On the premise of ensuring that medical services are not affected, the system needs to balance the needs of training and services. This may mean that more experienced medical staff need to be arranged in certain time periods or departments to free up some personnel for training. The system should take into account the personal development and career planning of medical staff, and provide them with a variety of training opportunities, such as domestic and international academic conferences, professional certification courses, etc., to support their career growth. The scheduling system should include a feedback and evaluation mechanism to regularly collect feedback from medical staff on training, evaluate the effectiveness of training, and how training affects the quality of medical services and user satisfaction. Through continuous improvement, ensure that the content of training is consistent with actual needs. When arranging training and scheduling, the system must comply with relevant labor laws and medical industry standards to ensure that the working hours and rest time of medical staff are reasonably arranged to avoid overwork.

[0003] In the standardized training of resident doctors, reasonable department rotation arrangements are crucial for the comprehensive development of doctors and the improvement of their professional skills. However, the traditional manual scheduling method is inefficient and prone to unreasonable and unbalanced scheduling, which cannot meet the high-quality requirements of training. Summary of the invention

[0004] The purpose of the present invention is to solve the problems of low efficiency, unreasonable and unbalanced scheduling of hospitals under traditional training, and to propose a scheduling method under hospital training tasks, comprising the following steps:

[0005] S1. Collect the deterministic and uncertain data in medical staff data, hospital data, and scheduling rules;

[0006] S2. Use neural networks to predict uncertain data. Based on deterministic data and predicted uncertain data, match professional skills of medical staff and balance workload as algorithm goals, and use reinforcement learning algorithms to generate scheduling plans.

[0007] S3, detect conflicts in the scheduling scheme and provide solutions;

[0008] S4. Analyze the scheduling plan data, generate feedback values ​​of the algorithm targets, and optimize the scheduling plan according to the feedback values ​​of the algorithm targets;

[0009] S5. Send shift notifications, confirmation information, and change reminders to users.

[0010] Furthermore, determined data include: professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in each department, rotation cycle, rest time, working hours, and user satisfaction; uncertain data include: patient flow and department needs.

[0011] Furthermore, reinforcement learning is used to generate scheduling plans, generate environment states, action spaces, and rewards for reinforcement learning;

[0012] Environmental status includes: current time, professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in the department, department needs, rotation cycle, rest time, working hours, and patient flow;

[0013] Action space includes: assigning employees for training, adjusting employees' work status, and emergency events

[0014] The elements of the reward include: the degree of professional skill matching between medical staff, workload balance, and customer complaints.

[0015] Furthermore, the reward function is designed as:

[0016] r=α1a+α2b+α3c-α4d

[0017] Among them, α1, α2, α3, and α4 all represent weight coefficients, and are all positive values. a represents the utilization rate of personnel, b represents the degree of professional skill matching of medical staff, c represents the workload balance, and d represents the customer complaint rate.

[0018] Furthermore, based on the environment state, action space, and rewards of reinforcement learning, a deep learning model is constructed using the policy gradient method A3C. The A3C model consists of two main parts: the policy network and the value network;

[0019] The policy network is used to predict the scheduling decision under a given state;

[0020] The value network is used to evaluate the value of the current scheduling solution of the strategy network.

[0021] Furthermore, the training process of the A3C model includes:

[0022] Initialize the global policy network and value network;

[0023] Multiple agents independently perform tasks in their own copies of the environment, collecting states, actions, and rewards at each moment;

[0024] The agent sends the collected states, actions, and rewards at each moment to the global network for updating model parameters;

[0025] The updated global network parameters are broadcast to all agents for the next round of learning.

[0026] The present invention also provides an automatic rotation scheduling system for hospital training, comprising:

[0027] Data collection and storage module, used to collect medical staff data, hospital data, deterministic data and uncertain data in scheduling rules;

[0028] The scheduling algorithm module is used to use neural networks to predict uncertain data. Based on deterministic data and predicted uncertain data, the algorithm takes the professional skills matching and workload balance of medical staff as the algorithm goals, and uses the reinforcement learning algorithm to generate a scheduling plan.

[0029] Conflict detection and resolution module, used to detect conflicts in the scheduling plan and provide solutions;

[0030] The data analysis module is used to analyze the scheduling plan data, generate feedback values ​​of the algorithm goals, and optimize the scheduling plan according to the feedback values ​​of the algorithm goals;

[0031] The notification module is used to send shift notifications, confirmation information, and change reminders to users.

[0032] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the scheduling method under the above-mentioned hospital training task is implemented.

[0033] The present invention also proposes an electronic device, including a processor and a memory, wherein the processor and the memory are interconnected, wherein the memory is used to store a computer program, the computer program includes computer-readable instructions, and the processor is configured to call the computer-readable instructions to execute the scheduling method under the above-mentioned hospital training task.

[0034] The beneficial effects brought by the technical solution provided by the present invention are:

[0035] The present invention first uses a neural network to predict the collected uncertainty data to generate a more accurate scheduling plan, takes the professional skills matching of medical staff and workload balance as the algorithm goals, uses a reinforcement learning algorithm to generate a scheduling plan, and optimizes the scheduling plan according to the feedback value of the algorithm goal, thereby improving the scheduling efficiency, scheduling rationality, and scheduling balance, and maximizing user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of a scheduling method under a hospital training task according to an embodiment of the present invention;

[0037] Figure 2 It is a block diagram of an electronic device in an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0039] The flowchart of the scheduling method for hospital training tasks according to an embodiment of the present invention is as follows: Figure 1 , specifically including the following steps:

[0040] S1. Collect deterministic and uncertain data from medical staff data, hospital data, and scheduling rules.

[0041] Medical staff data include: professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in each department, and user satisfaction; hospital data include: patient flow and department needs; scheduling rules include: rotation cycle, rest time, and working hours; and the collected data will be stored.

[0042] Professional skills refer to the areas in which medical staff are proficient in diseases; experience and qualifications refer to the level and experience of medical staff; training needs specify the experience and qualifications of lecturers and trainees, and the number of participants; personal training preferences refer to the courses that medical staff prefer to attend based on the training courses in their career development plans; the number of departments and the number of medical staff in each department can be used to determine the rotation cycle and the number of people scheduled for each shift in each department; user satisfaction is the user's score of satisfaction with each completed shift.

[0043] Statistics on the professional skills of medical staff, the number of departments, and the number of medical staff in each department are used to allocate on-duty personnel based on patient flow, department needs, and scheduling rules. Statistics on the professional skills, personal training preferences, experience qualifications, and the number of medical staff in each department are used to generate training learners and training instructors based on training needs and scheduling rules.

[0044] Determined data refers to data that will not change normally in the future, including: professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in each department, rotation cycles, rest time, working hours, and user satisfaction; uncertain data refers to data that cannot be determined in the future, including: patient flow and department needs.

[0045] S2. Before scheduling, it is necessary to predict the performance of uncertain data in the future and use neural networks to predict uncertain data. Based on the historical data of patient flow and department needs, neural networks are used to predict the future performance of uncertain data. Based on deterministic data and predicted uncertain data, the professional skills matching and workload balance of medical staff are taken as algorithm goals, and a reinforcement learning algorithm is used to generate a scheduling plan.

[0046] Workload balance means that the working hours of each medical staff should not differ too much. i -t| <t0,t i It represents the working hours of the i-th medical staff, t is the average working hours of the medical staff, and when the working hours of each medical staff are less than the set threshold, the workload is considered balanced.

[0047] The professional skills matching of medical staff refers to the matching of professional skills of each medical staff's arranged job position; the training courses are the non-professional skills of the trainees; the experience and qualifications of the training instructors; and the matching of personal training preferences and arranged training.

[0048] First, generate the environment state, action space, and rewards for reinforcement learning.

[0049] Environmental status includes: current time, professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in the department, department needs, rotation cycle, rest time, working hours, and patient flow.

[0050] The action space includes: assigning employees to training, adjusting employees’ work status, and emergencies.

[0051] The elements of the reward include: the degree of professional skill matching between medical staff, workload balance, and customer complaints.

[0052] The algorithmic goals of matching medical staff’s professional skills and balancing their workload are used as positive rewards, and customer complaints are used as negative rewards.

[0053] The reward function is designed as:

[0054] r=α1a+α2b+α3c-α4d

[0055] Among them, α1, α2, α3, and α4 all represent weight coefficients, and are all positive values. a represents the utilization rate of personnel, b represents the degree of professional skill matching of medical staff, c represents the workload balance, and d represents the customer complaint rate.

[0056] Based on the environment state, action space, and reward of reinforcement learning, a deep learning model is constructed using the policy gradient method A3C. The A3C model consists of two main parts: the policy network and the value network.

[0057] The policy network is used to predict the scheduling decision under a given state; the value network is used to evaluate the value of the current scheduling plan of the policy network.

[0058] The training process of the A3C model includes:

[0059] 1. Initialize the global strategy network and value network;

[0060] 2. Multiple agents independently perform tasks in their own copies of the environment, collecting states, actions, and rewards at each moment;

[0061] 3. The agent sends the collected states, actions, and rewards at each moment to the global network for updating model parameters;

[0062] 4. The updated global network parameters are broadcast to all agents for the next round of learning.

[0063] S3. Detect conflicts in the scheduling plan and provide solutions. Automatically detect conflicts in the scheduling, such as overlapping personnel, insufficient departmental staff, policy conflicts, etc., and provide solutions, such as automatically adjusting the schedule, reminding management personnel to intervene, etc., to ensure the accuracy and feasibility of the scheduling.

[0064] S4. Analyze the scheduling data and generate feedback values ​​for the algorithm goals, such as the matching degree of professional skills of medical staff and the balance of workload, and optimize the scheduling plan based on the feedback values ​​of the algorithm goals. Managers can also use the system's data analysis function to optimize the scheduling strategy based on historical data and current operating conditions.

[0065] S5. Send shift notifications, confirmation information, and change reminders to users. This is usually achieved through email, text messages, mobile app notifications, etc.

[0066] The present invention also provides an automatic rotation scheduling system for hospital training, comprising:

[0067] Data collection and storage module, used to collect medical staff data, hospital data, deterministic data and uncertain data in scheduling rules;

[0068] The scheduling algorithm module is used to use neural networks to predict uncertain data. Based on deterministic data and predicted uncertain data, the algorithm takes the professional skills matching and workload balance of medical staff as the algorithm goals, and uses the reinforcement learning algorithm to generate a scheduling plan.

[0069] Conflict detection and resolution module, used to detect conflicts in the scheduling plan and provide solutions;

[0070] The data analysis module is used to analyze the scheduling plan data, generate feedback values ​​of the algorithm goals, and optimize the scheduling plan according to the feedback values ​​of the algorithm goals;

[0071] The notification module is used to send shift notifications, confirmation information, and change reminders to users.

[0072] In an exemplary embodiment, a computer-readable storage medium is included, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the scheduling method under the above-mentioned hospital training task is implemented.

[0073] See also Figure 2 In an exemplary embodiment, an electronic device is also included, including at least one processor, at least one memory, and at least one communication bus.

[0074] Wherein, a computer program is stored in the memory, and the computer program includes computer-readable instructions. The processor calls the computer-readable instructions stored in the memory through the communication bus to execute the scheduling method under the above-mentioned hospital training task.

[0075] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A scheduling method for hospital training tasks, characterized in that: include: S1. Collect the deterministic and uncertain data in medical staff data, hospital data, and scheduling rules; S2. Use neural networks to predict uncertain data; Based on deterministic data and predicted uncertainty data, the algorithm targets matching medical staff's professional skills and balancing their workload, and uses a reinforcement learning algorithm to generate a scheduling plan. S3, detect conflicts in the scheduling scheme and provide solutions; S4. Analyze the scheduling plan data, generate feedback values ​​of the algorithm targets, and optimize the scheduling plan according to the feedback values ​​of the algorithm targets; S5. Send shift notifications, confirmation information, and change reminders to users.

2. The method for scheduling shifts under hospital training tasks according to claim 1, characterized in that: Determined data include: professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in each department, rotation cycles, rest time, working hours, and user satisfaction; uncertain data include: patient flow and department needs.

3. The method for scheduling shifts under hospital training tasks according to claim 2, characterized in that: Use reinforcement learning to generate scheduling plans, generate reinforcement learning environment states, action spaces, and rewards; Environmental status includes: current time, professional skills, training needs, personal training preferences, experience and qualifications, number of departments, number of medical staff in the department, department needs, rotation cycle, rest time, working hours, and patient flow; Action space includes: assigning employees for training, adjusting employees' work status, and emergency events The elements of the reward include: the degree of professional skill matching between medical staff, workload balance, and customer complaints.

4. The method for scheduling shifts under hospital training tasks according to claim 3, characterized in that: The reward function is designed as: r=α1a+α2b+α3c-α4d Among them, α1, α2, α3, and α4 all represent weight coefficients, and are all positive values. a represents the utilization rate of personnel, b represents the degree of professional skill matching of medical staff, c represents the workload balance, and d represents the customer complaint rate.

5. The method for scheduling shifts under hospital training tasks according to claim 4, characterized in that: Based on the environment state, action space, and reward of reinforcement learning, a deep learning model is constructed using the policy gradient method A3C. The A3C model consists of two main parts: the policy network and the value network. The policy network is used to predict the scheduling decision under a given state; The value network is used to evaluate the value of the current scheduling solution of the strategy network.

6. The method for scheduling shifts under hospital training tasks according to claim 5, characterized in that: The training process of the A3C model includes: Initialize the global policy network and value network; Multiple agents independently perform tasks in their own copies of the environment, collecting states, actions, and rewards at each moment; The agent sends the collected states, actions, and rewards at each moment to the global network for updating model parameters; The updated global network parameters are broadcast to all agents for the next round of learning.

7. An automatic rotation scheduling system for hospital training, characterized in that: include: Data collection and storage module, used to collect medical staff data, hospital data, deterministic data and uncertain data in scheduling rules; Scheduling algorithm module, used to predict uncertain data using neural networks; Based on deterministic data and predicted uncertainty data, the algorithm targets the matching of professional skills of medical staff and the balance of workload, and uses the reinforcement learning algorithm to generate a scheduling plan; Conflict detection and resolution module, used to detect conflicts in the scheduling plan and provide solutions; The data analysis module is used to analyze the scheduling plan data, generate feedback values ​​of the algorithm goals, and optimize the scheduling plan according to the feedback values ​​of the algorithm goals; The notification module is used to send shift notifications, confirmation information, and change reminders to users.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the processor and the memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises computer-readable instructions, and the processor is configured to call the computer-readable instructions to execute the method according to any one of claims 1 to 6.

Images