Rehabilitation nursing planning method and system for plastic burn of patient

Abstract

The invention provides a rehabilitation nursing planning method and system for plastic burn of a patient, and the method comprises the steps: collecting the physiological parameters, psychological state information and skin healing conditions of the patient, obtaining the basic information of the patient, and carrying out the analysis to generate a health condition score; based on the health condition score and the personalized rehabilitation target, a genetic algorithm is adopted to generate a rehabilitation activity scheme; constructing an uncertainty model by using a Bayesian network, and predicting result distribution of different treatment paths; performing global optimization on the rehabilitation activity scheme by using a mixed integer programming algorithm in combination with the optimization target, the constraint condition and the result distribution to obtain a rehabilitation activity combination implementation strategy; in the implementation process, a closed-loop feedback mechanism is used for regularly collecting and analyzing the actual reaction condition of a patient, and a deep reinforcement learning algorithm is used for adjustment to obtain a target rehabilitation activity combination implementation strategy; according to the invention, highly personalized rehabilitation nursing of the plastic burn patient is realized, the adaptability and the flexibility are high, and the rehabilitation experience and the treatment effect of the patient are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of medical and health technology, and in particular to a rehabilitation nursing planning method and system for patients with plastic burns. Background Art

[0002] The rehabilitation of patients with plastic burns is a long and complex process, which involves not only physical therapy, but also psychological support and social function recovery. With the development of medical technology and the improvement of patients' requirements for quality of life, the demand for personalized and precise rehabilitation care is growing.

[0003] Existing programs use multidisciplinary teamwork and standardized rehabilitation procedures. Multidisciplinary teamwork involves professional groups such as surgeons, nurses, physical therapists, and psychological counselors who jointly participate in the patient's rehabilitation process. Currently, many hospitals use standard rehabilitation processes based on clinical guidelines, which are usually developed based on extensive clinical research results. Existing programs are designed based on general guidelines and are difficult to fully consider the unique physical conditions and personal preferences of each patient. Existing follow-up methods often rely on patients' active reporting or regular return to the hospital for examinations, which may lead to the omission of key information and inaccurate evaluation of treatment effects. Information exchange between different departments of multidisciplinary teamwork may be delayed or incomplete, affecting the speed and accuracy of decision-making. Summary of the invention

[0004] The embodiments of the present invention provide a method and system for rehabilitation nursing planning for patients with plastic burns, so as to solve the problems in the prior art that it is difficult to fully consider the unique physical conditions and personal preferences of each patient, key information is omitted, the evaluation of treatment effects is inaccurate, and information exchange may be delayed or incomplete.

[0005] In a first aspect, an embodiment of the present invention provides a rehabilitation nursing planning method for a patient with plastic burns, comprising:

[0006] Receiving real-time data streams from different sources, the real-time data streams containing structured data and unstructured data;

[0007] Collect the patient's physiological parameters, psychological status information and skin healing status, use multimodal data fusion technology to obtain the patient's basic information, use an integrated learning algorithm to comprehensively analyze the patient's basic information and obtain a health status score;

[0008] Based on the health status score and the preset personalized rehabilitation goal, a rehabilitation activity plan with expected effect simulation is generated by using a genetic algorithm;

[0009] Using a Bayesian network to construct an uncertainty model in the patient's rehabilitation process, combined with the rehabilitation activity plan, using the uncertainty model to predict the distribution of results under different treatment pathways;

[0010] In combination with the preset optimization goal, constraint conditions and the result distribution, a mixed integer programming algorithm is applied to globally optimize the rehabilitation activity program to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination;

[0011] Implement the rehabilitation activity combination implementation strategy, use a closed-loop feedback mechanism during the implementation process, regularly collect and analyze patients' actual responses to the rehabilitation activity combination implementation strategy, and based on the actual responses, use a deep reinforcement learning algorithm to dynamically adjust the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy.

[0012] Optionally, the rehabilitation activity program is globally optimized by applying a mixed integer programming algorithm in combination with the preset optimization goal, constraint conditions and result distribution to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan of the optimal rehabilitation activity combination, including:

[0013] Determine the decision variables in the process of generating the rehabilitation activity combination implementation strategy, and define the objective function to determine the optimization target, wherein the decision variables include the execution time, frequency and intensity of each rehabilitation activity, and the objective function includes maximizing the rehabilitation effect, minimizing the treatment cost and improving the patient satisfaction;

[0014] defining constraints, wherein the constraints include actual medical resource constraints, patient preference constraints, and physiological constraints;

[0015] Based on the decision variables, the objective function and the constraint conditions, a mixed integer programming algorithm is applied to construct a mixed integer programming model;

[0016] Using a target mixed integer programming solver to solve the mixed integer programming model, to obtain a target solution that satisfies the constraint conditions and has an optimal objective function;

[0017] Based on the target solution and the result distribution, the expected effects of different rehabilitation activities within the target solution are evaluated to obtain expected effect evaluation results, and based on the expected effect evaluation results, a rehabilitation activity combination implementation strategy is generated; the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination, and the implementation plan includes the execution time, frequency and intensity of each rehabilitation activity in the optimal rehabilitation activity combination.

[0018] Optionally, the using a target mixed integer programming solver to solve the mixed integer programming model to obtain a target solution that satisfies the constraint conditions and has an optimal objective function includes:

[0019] Selecting an initial mixed integer programming solver, and setting parameters of the initial mixed integer programming solver according to actual needs to obtain a target mixed integer programming solver, wherein the solver parameters include a solution time limit and a solution accuracy requirement;

[0020] Based on the target mixed integer programming solver, the mixed integer programming model is processed by branch and bound method to obtain a preliminary feasible solution and a corresponding bound, wherein the bound is a lower bound or an upper bound;

[0021] Using the preliminary feasible solution and the limit, the mixed integer programming model is processed by a cutting plane method, and regions that do not contain integer solutions are excluded by adding constraint conditions to narrow the search space and obtain an optimized solution space;

[0022] Based on the optimized solution space, the mixed integer programming model is solved, and a target solution that satisfies all constraints and optimizes the objective function is selected. If the target solution cannot be selected within a preset time, the best feasible solution currently selected within the preset time is returned, and solution quality information of the best feasible solution is extracted. Based on the solution quality information, a target solution is generated, and the solution quality information includes a difference between the best feasible solution and the target solution.

[0023] Optionally, the use of a Bayesian network to construct an uncertainty model in the patient's rehabilitation process, combined with the rehabilitation activity plan, and using the uncertainty model to predict the distribution of results under different treatment pathways, includes:

[0024] Extracting features from the patient's basic information to obtain multimodal features, using a graph attention network and a structured learning algorithm to identify and define the multimodal features, generate nodes and edges, and obtain a directed acyclic graph based on the nodes and edges, wherein the nodes are various factors in the patient's rehabilitation process, including the physiological parameters, psychological state information, and skin healing conditions, and the edges are dependencies between the nodes;

[0025] Using Gaussian mixture model and expectation maximization algorithm, the conditional probability distribution of historical data of each node is estimated to obtain the preliminary conditional probability distribution;

[0026] Combining the variational inference method and the Bayesian optimization framework, adjusting the parameters of the Gaussian mixture model to optimize the preliminary conditional probability distribution and generate a target conditional probability distribution;

[0027] Based on the directed acyclic graph and the target conditional probability distribution, using a Bayesian network structure learning algorithm to construct an uncertainty model in the patient's rehabilitation process;

[0028] Based on the uncertain model and in combination with the rehabilitation activity program, the Bayesian reasoning method is used to perform reasoning calculations on different treatment pathways in a simulation environment to generate result distributions under different treatment pathways, where the treatment pathway is an implementation plan for the rehabilitation activity program.

[0029] Optionally, based on the uncertain model and in combination with the rehabilitation activity program, the Bayesian reasoning method is used to perform reasoning calculations on different treatment pathways in a simulation environment to generate result distributions under different treatment pathways, wherein the treatment pathway is an implementation plan of the rehabilitation activity program, including:

[0030] Based on the uncertainty model and the rehabilitation activity plan, defining a treatment pathway, wherein the treatment pathway includes the rehabilitation activity plan, an implementation sequence of the rehabilitation activity plan, and a timeline;

[0031] Initializing a simulation environment based on the multimodal features, setting the patient's health state as an initial state, the health state including the physiological parameters and the psychological state information;

[0032] Based on the simulation environment and the initial state, the treatment path is executed, and the state distribution of each step is predicted using the Bayesian reasoning method. Based on the state distribution, the patient's health status is updated, and the result distribution of each step is recorded to generate the result distribution under different treatment paths, wherein the result distribution includes the probability distribution of efficacy, cost, and patient satisfaction.

[0033] Optionally, the implementation of the rehabilitation activity combination implementation strategy uses a closed-loop feedback mechanism during the implementation process to regularly collect and analyze the actual response of patients to the rehabilitation activity combination implementation strategy, and based on the actual response, dynamically adjusts the rehabilitation activity combination implementation strategy using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy, including:

[0034] Implement the rehabilitation activity combination implementation strategy, regularly collect the actual feedback from patients, compare the actual feedback with the preset effect, identify the effect difference between the actual effect and the preset effect, evaluate the effect difference, and obtain the actual effect evaluation result, wherein the actual feedback includes the efficacy, side effects and patient satisfaction;

[0035] Initializing a deep reinforcement learning agent using a deep reinforcement learning algorithm based on the actual effect evaluation result;

[0036] The deep reinforcement learning agent is used to iteratively optimize the rehabilitation activity combination implementation strategy through experience replay and gradient descent method to obtain the target rehabilitation activity combination implementation strategy.

[0037] Optionally, the using the deep reinforcement learning agent to iteratively optimize the rehabilitation activity combination implementation strategy through experience replay and gradient descent method to obtain a target rehabilitation activity combination implementation strategy includes:

[0038] Utilizing the deep reinforcement learning agent to interact with the environment to generate experience data, and storing the experience data in an experience replay buffer, wherein the experience data includes states, actions, and rewards;

[0039] Randomly extracting target experience data from the experience replay buffer, calculating a target value of the target experience data, comparing the target value with a predicted value of a policy network in the deep reinforcement learning agent, and calculating a loss function;

[0040] Based on the loss function, the parameters of the strategy network are updated using the gradient descent method to obtain updated strategy network parameters, and based on the updated strategy network parameters, the rehabilitation activity combination implementation strategy is optimized to obtain the target rehabilitation activity combination implementation strategy.

[0041] In a second aspect, an embodiment of the present invention provides a rehabilitation nursing planning system for patients with plastic burns, comprising:

[0042] The acquisition module is used to collect the patient's physiological parameters, psychological status information and skin healing status, obtain the patient's basic information using multimodal data fusion technology, and use an integrated learning algorithm to comprehensively analyze the patient's basic information to obtain a health status score;

[0043] A generation module, for generating a rehabilitation activity program with expected effect simulation by using a genetic algorithm based on the health status score and the preset personalized rehabilitation goal;

[0044] A construction module is used to construct an uncertainty model in the patient's rehabilitation process using a Bayesian network, and in combination with the rehabilitation activity plan, use the uncertainty model to predict the distribution of results under different treatment paths;

[0045] an optimization module, for applying a mixed integer programming algorithm to globally optimize the rehabilitation activity program in combination with a preset optimization goal, constraint conditions and the result distribution, to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination;

[0046] The adjustment module is used to implement the rehabilitation activity combination implementation strategy. During the implementation process, a closed-loop feedback mechanism is used to regularly collect and analyze the actual responses of patients to the rehabilitation activity combination implementation strategy. Based on the actual responses, a deep reinforcement learning algorithm is used to dynamically adjust the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy.

[0047] In a third aspect, an embodiment of the present invention provides a computing device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a rehabilitation care planning method for patients with plastic burns as described in any one of the first aspects.

[0048] In a fourth aspect, an embodiment of the present invention provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement a rehabilitation care planning method for patients with plastic burns as described in any one of the first aspects.

[0049] In an embodiment of the present invention, the patient's physiological parameters, psychological state information and skin healing conditions are collected, the patient's basic information is obtained using multimodal data fusion technology, and the patient's basic information is comprehensively analyzed using an integrated learning algorithm to obtain a health status score; based on the health status score and a preset personalized rehabilitation goal, a genetic algorithm is used to generate a rehabilitation activity plan with an expected effect simulation; an uncertainty model in the patient's rehabilitation process is constructed using a Bayesian network, and combined with the rehabilitation activity plan, the uncertainty model is used to predict the result distribution under different treatment paths; combined with the preset optimization goal, constraints and the result distribution, a mixed integer programming algorithm is used to globally optimize the rehabilitation activity plan to obtain a rehabilitation activity combination implementation strategy, and the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination; the rehabilitation activity combination implementation strategy is implemented, and a closed-loop feedback mechanism is used during the implementation process to regularly collect and analyze the patient's actual response to the rehabilitation activity combination implementation strategy, and based on the actual response, a deep reinforcement learning algorithm is used to dynamically adjust the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy. The technical solution provided by the present invention realizes highly personalized rehabilitation care for plastic surgery burn patients, and at the same time has strong adaptability and flexibility, can effectively cope with various challenges in the rehabilitation process, effectively improve the patient's rehabilitation experience and improve the treatment effect. Among them, by configuring the solver parameters, it is ensured that a high-quality solution is obtained within a limited time, thereby improving the solution efficiency; using a method combining the branch and bound method and the cutting plane method, the search space is effectively narrowed, ensuring that the solution found is as close to the optimal solution as possible, thereby improving the optimization accuracy; the solver parameters are flexibly adjusted according to actual needs to adapt to problems of different scales and complexities, thereby enhancing the flexibility and applicability of the system; when the optimal solution cannot be found within the preset time, the current best feasible solution can be returned in time, and solution quality information can be provided to ensure that users always have available solutions.

[0050] These and other aspects of the present invention will become more apparent from the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 A flowchart of a rehabilitation nursing planning method for a patient with plastic burns provided by an embodiment of the present invention;

[0053] Figure 2 A schematic diagram of a rehabilitation nursing planning system for patients with plastic surgery burns provided by an embodiment of the present invention;

[0054] Figure 3 A schematic diagram of the structure of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to enable those skilled in the art to better understand the solutions of 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.

[0056] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0058] Figure 1 A flowchart of a method for planning rehabilitation nursing for patients with plastic burns is provided in an embodiment of the present invention, such as Figure 1 As shown, the method includes:

[0059] In the rehabilitation care process of patients with plastic burns, personalized and accurate treatment plans are crucial to improving rehabilitation effects, reducing treatment costs, and improving patient satisfaction. Traditional methods often rely on the doctor's experience and fixed patterns, and it is difficult to fully consider the unique situation of each patient. With the development of science and technology, it is possible to build a more intelligent and efficient rehabilitation care plan. Based on this, the present invention provides a rehabilitation care plan method for patients with plastic burns, such as Figure 1 ,include:

[0060] Step 101: collecting the patient's physiological parameters, psychological status information, and skin healing status, using multimodal data fusion technology to obtain the patient's basic information, and using an integrated learning algorithm to comprehensively analyze the patient's basic information to obtain a health status score;

[0061] In this step, physiological parameters refer to various quantitative indicators of the patient's body, such as blood pressure, heart rate, body temperature, blood oxygen saturation, etc. These parameters reflect the patient's health status and physical function. Psychological state information includes anxiety level, depression score, sleep quality, mood swings, etc., which are usually obtained through questionnaires or psychological assessment tools. Psychological state has an important impact on the rehabilitation effect. Skin healing refers to the progress of skin healing at the burn site, including wound size, healing speed, infection status, etc. Skin healing is one of the key indicators of burn rehabilitation. Multimodal data fusion technology refers to the integration of multidimensional data from different sources and types to generate a comprehensive data set to more comprehensively understand the patient's status. Ensemble learning algorithm refers to a machine learning method that improves prediction performance by combining multiple basic models.

[0062] This step collects the patient's physiological parameters, psychological status information, and skin healing conditions. Physiological parameters include blood pressure and heart rate, psychological status information includes anxiety level and depression score, and skin healing conditions include wound size and healing speed. Multimodal data fusion technology is used to integrate these different types of data to form a comprehensive data set, thereby more comprehensively describing the patient's current health status. Based on the above comprehensive data set, an integrated learning algorithm (such as random forest and gradient boosting machine) is applied to conduct a comprehensive analysis of the data. The algorithm generates a more accurate and stable health status score by training multiple basic models and combining their output results. This score not only reflects the patient's current health status, but also provides an important reference for subsequent rehabilitation plans.

[0063] Step 102: Based on the health status score and the preset personalized rehabilitation goal, a rehabilitation activity plan with expected effect simulation is generated using a genetic algorithm;

[0064] In this step, genetic algorithm refers to an optimization algorithm that imitates natural selection and genetic mechanisms to search for the optimal solution. In this case, it is used to generate a rehabilitation activity plan with a simulated expected effect.

[0065] This step uses a genetic algorithm to generate a variety of rehabilitation activity plans based on the patient's health status score and personalized rehabilitation goals, such as improving quality of life and reducing pain. The genetic algorithm gradually optimizes the plan by simulating the natural selection process, ensuring that each plan is accompanied by the expected effect simulation to better evaluate its feasibility and potential effect.

[0066] Step 103: constructing an uncertainty model of the patient's rehabilitation process using a Bayesian network, combining the rehabilitation activity plan, and using the uncertainty model to predict the distribution of results under different treatment pathways;

[0067] In this step, the Bayesian network refers to a probabilistic graphical model that represents the dependencies between variables. By constructing a directed acyclic graph (DAG) and conditional probability distribution, it can be used to model uncertainty problems. The result distribution refers to the probability distribution of different treatment paths and predicting possible results, such as efficacy, cost, and patient satisfaction, to help select the best treatment plan.

[0068] This step builds a Bayesian network model based on the patient's multimodal characteristics and rehabilitation activity plan. The model captures the uncertainty in the rehabilitation process by defining nodes and edges and estimating conditional probability distribution. Through the Bayesian reasoning method, different treatment pathways are inferred and calculated in a simulation environment to generate the result distribution under each pathway, including the probability distribution of efficacy, cost, and patient satisfaction.

[0069] Step 104: combining the preset optimization goal, constraint conditions and the result distribution, applying a mixed integer programming algorithm to globally optimize the rehabilitation activity program to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination;

[0070] In this step, the mixed integer programming algorithm refers to a mathematical optimization method used to solve optimization problems involving discrete variables. In this case, it is used to globally optimize the rehabilitation activity plan to ensure that the preset optimization objectives and constraints are met.

[0071] This step uses a mixed integer programming algorithm to globally optimize the rehabilitation activity plan based on the preset optimization goals, constraints, and result distribution, such as maximizing rehabilitation effects, minimizing treatment costs, and improving patient satisfaction, and constraints such as medical resource limitations, patient preferences, and physiological constraints. The algorithm constructs a mathematical model and solves it to find a combination of rehabilitation activities and its implementation plan that satisfies all constraints and optimizes the objective function. The resulting rehabilitation activity combination implementation strategy includes the best rehabilitation activity combination and a specific implementation plan.

[0072] Step 105: Implement the rehabilitation activity combination implementation strategy, use a closed-loop feedback mechanism during the implementation process, regularly collect and analyze the actual response of patients to the rehabilitation activity combination implementation strategy, and dynamically adjust the rehabilitation activity combination implementation strategy based on the actual response using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy;

[0073] In this step, the deep reinforcement learning algorithm refers to a method that combines deep learning and reinforcement learning. The agent continuously learns and optimizes its behavior strategy by interacting with the environment. In this case, it is used to dynamically adjust the implementation strategy of the rehabilitation activity combination.

[0074] In this step, specific rehabilitation activities are started according to the generated rehabilitation activity combination implementation strategy. During the implementation process, the actual response of the patient, such as changes in physiological parameters, improvement in psychological state, and progress in skin healing, is regularly collected and analyzed. Based on this feedback information, the deep reinforcement learning algorithm is used to dynamically adjust the rehabilitation activity combination implementation strategy to ensure that it can better adapt to the actual situation and needs of the patient. The target rehabilitation activity combination implementation strategy generated in the end can be continuously optimized in practice to achieve the best rehabilitation effect.

[0075] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0076] By comprehensively analyzing the patient's multimodal data, a personalized rehabilitation plan is generated to ensure that each patient can receive the most suitable treatment for him or her;

[0077] Utilize advanced algorithms and technologies to optimize rehabilitation activity plans and maximize rehabilitation effects, while taking into account the patient's psychological and physiological state to improve overall rehabilitation quality;

[0078] Through global optimization and dynamic adjustment, unnecessary treatment steps and resource waste can be reduced, thus lowering treatment costs;

[0079] Pay attention to the patient's psychological state and actual response, adjust the rehabilitation plan in time, and improve the patient's treatment experience and satisfaction;

[0080] Use closed-loop feedback mechanisms and deep reinforcement learning algorithms to achieve real-time adjustment and optimization of rehabilitation programs to ensure that the program always meets the patient's latest conditions;

[0081] Provide scientific decision support based on data and models to help doctors and nurses make more informed choices and improve the quality and efficiency of medical services.

[0082] After generating a preliminary rehabilitation activity plan, in order to ensure the feasibility and optimality of the plan, it must be globally optimized. In order to take into account the individual needs of patients, the limitations of medical resources, and the expected effects under different treatment pathways, based on this, the present invention provides a specific embodiment, wherein the step 104, in combination with the preset optimization objectives, constraints, and the result distribution, applies a mixed integer programming algorithm to globally optimize the rehabilitation activity plan, and obtains a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination, and specifically includes the following steps:

[0083] Step 401: determining decision variables in the process of generating a rehabilitation activity combination implementation strategy, and defining an objective function to determine an optimization target, wherein the decision variables include the execution time, frequency, and intensity of each rehabilitation activity, and the objective function includes maximizing rehabilitation effect, minimizing treatment cost, and improving patient satisfaction;

[0084] In this step, decision variables refer to the parameters that can be adjusted during the optimization process, which in this case include the duration, frequency, and intensity of each rehabilitation activity. These variables directly affect rehabilitation outcomes and resource utilization.

[0085] This step clarifies the decision variables that need to be adjusted when generating a rehabilitation activity combination implementation strategy, including the execution time, frequency, and intensity of each rehabilitation activity. The execution time is such as a specific time period per day, the frequency is such as a few times per week, and the intensity is such as the amount of exercise or drug dosage; define the objective function, that is, the goal of optimization. The objective functions in this case include maximizing the rehabilitation effect, minimizing the cost of treatment, and improving patient satisfaction, that is, improving the patient's health status score, reducing the use and cost of medical resources, and improving the patient's acceptance and comfort of the treatment process.

[0086] Step 402: defining constraints, wherein the constraints include actual medical resource constraints, patient preference constraints, and physiological constraints;

[0087] In this step, constraints limit the range of feasible solutions in the optimization process to ensure that the generated solutions are feasible in actual operation. The constraints in this case include actual medical resource constraints, patient preference constraints, and physiological constraints.

[0088] In order to ensure that the generated rehabilitation activity combination implementation strategy is feasible in actual operation, the following constraints need to be defined in this step: actual medical resource constraints, considering resource limitations such as available medical equipment and the number of medical staff; patient preference constraints, respecting patients' personal preferences and living habits, for example, some patients may prefer to perform rehabilitation activities in the morning; physiological constraints, ensuring that rehabilitation activities do not place excessive burden on patients' bodies, such as avoiding the negative impact of high-intensity activities on frail patients.

[0089] Step 403: Based on the decision variables, the objective function and the constraint conditions, a mixed integer programming algorithm is applied to construct a mixed integer programming model;

[0090] In this step, the mixed integer programming algorithm refers to a mathematical optimization method used to solve optimization problems involving discrete variables. By building a mathematical model and solving it, a solution that satisfies all constraints and makes the objective function optimal is found. The mixed integer programming model refers to a mathematical model that contains decision variables, objective functions, and constraints, and is used to describe and solve complex optimization problems.

[0091] In this step, a mixed integer programming model is constructed based on the determined decision variables, objective functions and constraints. The model takes the execution time, frequency and intensity of rehabilitation activities as decision variables, maximizes rehabilitation effects, minimizes treatment costs and improves patient satisfaction as objective functions, and considers various constraints. Through this model, complex optimization problems can be systematically described and solved.

[0092] Step 404: using a target mixed integer programming solver to solve the mixed integer programming model to obtain a target solution that satisfies the constraint conditions and has an optimal target function;

[0093] In this step, the mixed integer programming solver refers to a software tool specifically used to solve mixed integer programming problems, which can efficiently find a solution that satisfies the constraints and optimizes the objective function.

[0094] In this step, select a suitable mixed integer programming solver, such as Gurobi, CPLEX, etc., and set the solver parameters according to actual needs, such as solution time limit and solution accuracy requirements; use the solver to solve the constructed mixed integer programming model to find the target solution that meets all constraints and optimizes the objective function. This solution represents the best combination of rehabilitation activities and its specific implementation plan.

[0095] Step 405: Based on the target solution and the result distribution, the expected effects of different rehabilitation activities in the target solution are evaluated to obtain an expected effect evaluation result, and based on the expected effect evaluation result, a rehabilitation activity combination implementation strategy is generated; the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan of the optimal rehabilitation activity combination, and the implementation plan includes the execution time, frequency and intensity of each rehabilitation activity in the optimal rehabilitation activity combination;

[0096] This step evaluates the expected effect of each rehabilitation activity based on the target solution obtained by the solver and the previously predicted distribution of results for different treatment pathways. The evaluation includes aspects such as efficacy, cost, and patient satisfaction. Based on these evaluation results, the final rehabilitation activity combination implementation strategy is generated, including the optimal rehabilitation activity combination and its detailed implementation plan, such as the specific execution time, frequency, and intensity of each rehabilitation activity. This step ensures that the generated solution is not only optimal in theory, but also achieves the expected effect in practical application.

[0097] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0098] By comprehensively considering the multimodal data and individual needs of patients, we generate the most suitable rehabilitation activity combination implementation strategy for each patient to ensure the personalization and effectiveness of the program;

[0099] Using mixed integer programming algorithms, we can ensure that the rehabilitation effect is maximized and the treatment cost is minimized under limited medical resources, thus improving the efficiency of resource utilization;

[0100] Through global optimization and dynamic adjustment, the rehabilitation effect is maximized, while taking into account the patient's psychological and physiological state, and comprehensively improving the quality of rehabilitation;

[0101] Pay attention to the actual reactions and preferences of patients, adjust rehabilitation programs in a timely manner, and improve patients' treatment experience and satisfaction;

[0102] Provide scientific decision support based on data and models to help doctors and nurses make more informed choices and improve the quality and efficiency of medical services.

[0103] After constructing the mixed integer programming model, in order to find a solution that satisfies all constraints and optimizes the objective function, an efficient solution tool must be used. Based on this, the present invention provides a specific embodiment, wherein step 404 uses a target mixed integer programming solver to solve the mixed integer programming model to obtain a target solution that satisfies the constraints and optimizes the objective function, specifically comprising the following steps:

[0104] Step 411: selecting an initial mixed integer programming solver, and setting parameters of the initial mixed integer programming solver according to actual requirements to obtain a target mixed integer programming solver, wherein the solver parameters include a solution time limit and a solution accuracy requirement;

[0105] In this step, the initial mixed integer programming solver refers to the software tool initially selected for solving mixed integer programming problems, such as Gurobi, CPLEX, etc. These solvers provide a variety of optimization algorithms and parameter setting options. The target mixed integer programming solver refers to the solver after parameter adjustment, which is optimized and configured specifically for the current problem to meet specific practical needs. The solution time limit refers to the maximum time for the solver to run, usually in seconds or minutes. It is used to prevent the solution process from running indefinitely and ensure that the results are obtained within a reasonable time. The solution accuracy requirement refers to the allowable error range between the solution found by the solver and the optimal solution, which is usually expressed as a relative gap, that is, the gap between the objective function value of the current best feasible solution and the objective function value of the optimal solution.

[0106] In this step, an initial solver suitable for solving mixed integer programming problems is selected, such as Gurobi or CPLEX; the parameters of the solver are set according to actual needs, mainly including the solution time limit and solution accuracy requirements; the solution time limit is set, for example, the maximum solution time is set to 60 minutes to ensure that the result is obtained within a reasonable time; the solution accuracy requirements are set, for example, the relative gap is set to 0.01% to ensure that the accuracy of the solution meets the expected standard; after these parameters are configured, a target mixed integer programming solver with optimized configuration for the current problem is obtained to improve the solution efficiency and solution quality.

[0107] Step 412: Based on the target mixed integer programming solver, the mixed integer programming model is processed by branch and bound method to obtain a preliminary feasible solution and a corresponding bound, where the bound is a lower bound or an upper bound;

[0108] In this step, the Branch and Bound method refers to a commonly used method for solving integer programming problems. By systematically partitioning the problem space and evaluating the bounds of each subproblem, the search scope is gradually narrowed and the optimal solution is finally found. The preliminary feasible solution refers to the first solution found in the solution process that satisfies all constraints. Although it is not necessarily the optimal solution, it can be used as a reference benchmark. The bound, i.e., the lower bound or the upper bound, refers to the reference value used in the branch and bound method to evaluate the quality of the subproblem solution. The lower bound represents the minimum possible value of the currently known optimal solution, and the upper bound represents the maximum possible value of the currently known feasible solution.

[0109] This step uses the configured target mixed integer programming solver to perform branch and bound on the constructed mixed integer programming model to obtain preliminary feasible solutions and corresponding bounds. This method gradually narrows the search scope by systematically splitting the problem space and evaluating the bounds of each subproblem.

[0110] Step 413: using the preliminary feasible solution and the boundary, performing a cutting plane method on the mixed integer programming model, excluding regions that do not contain integer solutions by adding constraint conditions, so as to narrow the search space and obtain an optimized solution space;

[0111] In this step, the Cutting Plane Method refers to adding additional linear inequalities to exclude regions that do not contain integer solutions, thereby reducing the search space and speeding up the solution. The optimized solution space refers to the search space that has been reduced after the Cutting Plane Method, which only contains possible integer solutions.

[0112] This step further optimizes the solution space based on the preliminary feasible solution and bounds by applying the cutting plane method. By introducing additional linear inequalities, i.e., cutting planes, the search space is narrowed by excluding regions that do not contain integer solutions. This process significantly reduces unnecessary searches and improves the efficiency of the solution. The resulting optimized solution space only contains possible integer solutions, greatly increasing the possibility of finding the optimal solution.

[0113] More specifically, the embodiment of the present invention further provides a cutting plane constraint formula for generating a cutting plane to exclude regions that do not contain integer solutions. The specific calculation formula is as follows:

[0114] ∑ i∈l a i x i +∑ j∈J b j y j +∑ k∈K d k z k ≥c;

[0115] Where i∈I means that i is an element in the continuous variable set I, which usually represents a variable that can take any real value; a i Represents the continuous variable x i The coefficient of correlation; x i represents a continuous variable, which indicates a parameter or indicator in rehabilitation activities, such as treatment intensity, time, etc.; j∈J means that j is an element in the integer variable set J, and the integer variable can only take integer values; b j Represents the integer variable y j The coefficient of correlation; y j represents an integer variable, which indicates a parameter that needs to appear in the form of an integer in rehabilitation activities, such as the number of treatments; k∈K means that k is an element in the binary variable set K, and the binary variable can only take 0 or 1; d k Represents the binary variable z k The coefficient of correlation; z k represents a binary variable, which is used to indicate whether certain specific activities are executed, 0 means not executed, and 1 means executed; c represents the constant term on the right side of the cutting plane, which indicates the threshold of the constraint condition.

[0116] In the process of solving the mixed integer programming model, this formula adds additional constraints to exclude solutions that do not meet the integer requirements, thereby narrowing the search space and helping to find the optimal integer solution. This method improves the efficiency of the solution and ensures that the final solution is a practical and feasible integer solution.

[0117] Step 414: Based on the optimization solution space, the mixed integer programming model is solved, and a target solution that satisfies all constraints and optimizes the objective function is selected. If the target solution cannot be selected within the preset time, the best feasible solution currently selected within the preset time is returned, and the solution quality information of the best feasible solution is extracted. Based on the solution quality information, a target solution is generated, and the solution quality information includes the difference between the best feasible solution and the target solution.

[0118] In this step, the solution quality information includes the difference between the best feasible solution and the target solution, which is used to evaluate the quality of the solution. If the target solution cannot be found within a preset time, a suboptimal solution is generated based on the solution quality information.

[0119] In this step, the mixed integer programming model is solved in the optimization solution space to find the target solution that satisfies all constraints and optimizes the objective function. If such a solution is found within the preset time limit, the target solution is directly returned. If the target solution cannot be found, the best feasible solution found within the current preset time is returned; the solution quality information of the best feasible solution is extracted, including the difference between it and the theoretical optimal solution; a suboptimal solution is generated based on the solution quality information to ensure a high-quality solution within a limited time.

[0120] More specifically, the embodiment of the present invention also provides a formula of an objective function, which comprehensively considers multiple key factors to ensure that the final rehabilitation activity combination is not only effective, but also economical, safe and satisfactory to the patient. The specific calculation formula is as follows:

[0121]

[0122] Among them, f(x) is the objective function, which represents the goal of overall optimization; w 1 is the weight factor of the rehabilitation effect, which is used to adjust the importance of the rehabilitation effect in the overall goal; x 1 is the rehabilitation effect evaluation value corresponding to the rehabilitation activity program x; w 2 is the weight factor of treatment cost, which is used to adjust the importance of treatment cost in the overall goal; x 2 is the treatment cost assessment value corresponding to the rehabilitation activity plan x; w 3 is the weight factor of patient satisfaction, which is used to adjust the importance of patient satisfaction in the overall goal; x 3 is the patient satisfaction evaluation value corresponding to the rehabilitation activity program x; w 4 is the weight factor of the risk factor, which is used to adjust the importance of the risk factor in the overall goal; M is the number of risk factors considered; p m (x) is the probability of the mth risk factor occurring under rehabilitation activity plan x, which can be estimated based on historical data and clinical research knowledge; sm is the severity score of the mth risk factor, which can be determined based on medical criteria or patient feedback.

[0123] This objective function takes into account multiple key factors, including rehabilitation effect, treatment cost, patient satisfaction and potential risks, which helps to find the best solution to balance multiple goals and develop a more comprehensive and safe rehabilitation plan; the weight factor can be adjusted according to the specific application scenario and the specific needs of the patient, so as to achieve personalized and customized rehabilitation care; by maximizing the rehabilitation effect, it can ensure that the patient's physiological and psychological state is improved to the greatest extent; by minimizing the treatment cost, it can reduce the overall economic burden and improve resource utilization efficiency while ensuring the rehabilitation effect; by maximizing patient satisfaction, it can improve patient compliance and enthusiasm for rehabilitation, thereby further improving the rehabilitation effect; by minimizing risk factors, it can ensure the safety of the rehabilitation process and reduce potential adverse consequences.

[0124] More specifically, the embodiment of the present invention also provides a formula of an objective function, which comprehensively considers multiple key factors to ensure that the final rehabilitation activity combination is not only effective, but also economical, safe and satisfactory to the patient. The specific calculation formula is as follows:

[0125]

[0126] Among them, f(x) is the objective function, which represents the goal of overall optimization; w 1 is the weight factor of the rehabilitation effect, which is used to adjust the importance of the rehabilitation effect in the overall goal; x 1 is the rehabilitation effect evaluation value corresponding to the rehabilitation activity program x; w 2 is the weight factor of treatment cost, which is used to adjust the importance of treatment cost in the overall goal; x 2 is the treatment cost assessment value corresponding to the rehabilitation activity plan x; w 3 is the weight factor of patient satisfaction, which is used to adjust the importance of patient satisfaction in the overall goal; x 3 is the patient satisfaction evaluation value corresponding to the rehabilitation activity program x; w 4 is the weight factor of the risk factor, which is used to adjust the importance of the risk factor in the overall goal; M is the number of risk factors considered; p m (x) is the probability of the mth risk factor occurring under rehabilitation activity plan x, which can be estimated based on historical data and clinical research knowledge; s m is the severity score of the mth risk factor, which can be determined based on medical criteria or patient feedback.

[0127] This objective function takes into account multiple key factors, including rehabilitation effect, treatment cost, patient satisfaction and potential risks, which helps to find the best solution to balance multiple goals and develop a more comprehensive and safe rehabilitation plan; the weight factor can be adjusted according to the specific application scenario and the specific needs of the patient, so as to achieve personalized and customized rehabilitation care; by maximizing the rehabilitation effect, it can ensure that the patient's physiological and psychological state is improved to the greatest extent; by minimizing the treatment cost, it can reduce the overall economic burden and improve resource utilization efficiency while ensuring the rehabilitation effect; by maximizing patient satisfaction, it can improve patient compliance and enthusiasm for rehabilitation, thereby further improving the rehabilitation effect; by minimizing risk factors, it can ensure the safety of the rehabilitation process and reduce potential adverse consequences.

[0128] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0129] By configuring the solver parameters, we can ensure that high-quality solutions are obtained within a limited time, thus improving the solving efficiency.

[0130] The combination of branch and bound method and cutting plane method effectively narrows the search space, ensures that the solution found is as close to the optimal solution as possible, and improves the optimization accuracy;

[0131] The solver parameters can be flexibly adjusted according to actual needs to adapt to problems of different scales and complexities, thus enhancing the flexibility and applicability of the system;

[0132] When the optimal solution cannot be found within the preset time, the current best feasible solution can be returned in time and solution quality information can be provided to ensure that users always have available solutions.

[0133] In order to more accurately evaluate the effects of different treatment pathways and deal with the uncertainty in the rehabilitation process, based on this, the present invention provides a specific embodiment, in which the step 103 uses a Bayesian network to construct an uncertainty model in the patient's rehabilitation process, combines the rehabilitation activity plan, and uses the uncertainty model to predict the distribution of results under different treatment pathways, specifically including the following steps:

[0134] Step 301: extracting features from the basic information of the patient to obtain multimodal features, using a graph attention network and a structured learning algorithm to identify and define the multimodal features, generate nodes and edges, and obtain a directed acyclic graph based on the nodes and edges, wherein the nodes are various factors in the patient's rehabilitation process, including the physiological parameters, psychological state information, and skin healing conditions, and the edges are dependencies between the nodes;

[0135] In this step, a directed acyclic graph (DAG) refers to a graph structure in which nodes are connected by directed edges and there are no loops. In this case, it is used to represent the dependencies between different factors in the patient's recovery process.

[0136] This step collects the patient's physiological parameters, psychological state information, and skin healing status, and uses feature extraction technology to extract multimodal features from these data. These features can more comprehensively describe the patient's current health status and provide a basis for subsequent modeling; the extracted multimodal features are identified and defined using the graph attention network and structure learning algorithm, and nodes and edges are generated. Based on the nodes and edges, a directed acyclic graph is obtained: Among them, the graph attention network (GAT) identifies and defines the extracted multimodal features. The graph attention network can capture the complex relationship between different features and automatically learn which features are most important for predicting results; among them, the structure learning algorithm, such as the PC algorithm or the MMHC algorithm, generates nodes and edges based on the dependency between features; the nodes represent various factors in the patient's rehabilitation process, including physiological parameters, psychological state information, and skin healing; the edges represent the dependency or causal relationship between nodes, forming a directed acyclic graph (DAG). For example, changes in physiological parameters may affect the speed of skin healing, and changes in psychological state may affect the effect of rehabilitation activities.

[0137] Step 302: using a Gaussian mixture model and an expectation maximization algorithm, estimate the conditional probability distribution of the historical data of each node to obtain a preliminary conditional probability distribution;

[0138] In this step, Gaussian Mixture Model (GMM) refers to a commonly used probability model that can fit complex multimodal distributions.

[0139] In this step, for each node’s historical data, a Gaussian mixture model (GMM) is used to model its distribution, and the expectation-maximization (EM) algorithm is applied to iteratively optimize the model parameters to estimate the probability distribution of each node under different conditions and obtain the preliminary conditional probability distribution.

[0140] Step 303: Combining the variational inference method and the Bayesian optimization framework, adjusting the parameters of the Gaussian mixture model to optimize the preliminary conditional probability distribution and generate a target conditional probability distribution;

[0141] This step combines the variational inference method and the Bayesian optimization framework to further optimize the preliminary conditional probability distribution. Variational inference is used to approximate complex posterior distributions, while Bayesian optimization is used to efficiently search for optimal model parameters. Through this process, the parameters of the Gaussian mixture model are adjusted so that the conditional probability distribution more accurately reflects the characteristics of the actual data, and finally the target conditional probability distribution is generated.

[0142] Step 304: constructing an uncertainty model in the patient's rehabilitation process using a Bayesian network structure learning algorithm based on the directed acyclic graph and the target conditional probability distribution;

[0143] In this step, the Bayesian network structure learning algorithm refers to an algorithm used to construct a Bayesian network, which determines the connection mode and conditional probability distribution between nodes by analyzing the dependencies in the data. The uncertainty model refers to a mathematical model that describes the uncertainty in the system. In this case, it is used to capture various uncertainties in the patient's rehabilitation process.

[0144] This step uses directed acyclic graphs and target conditional probability distributions, and applies Bayesian network structure learning algorithms, such as the K2 algorithm or the BI C scoring method, to build an uncertainty model for the patient's rehabilitation process. This model not only describes the dependencies between various factors, but also quantifies the probabilistic dependencies between them, providing a basis for subsequent reasoning calculations.

[0145] Step 305: Based on the uncertain model and in combination with the rehabilitation activity program, different treatment pathways are inferred and calculated in a simulation environment by using a Bayesian reasoning method to generate result distributions under different treatment pathways, where the treatment pathway is an implementation plan for the rehabilitation activity program;

[0146] This step is based on the constructed uncertainty model and combined with the specific rehabilitation activity plan. Through Bayesian reasoning methods such as Gibbs sampling or Markov chain Monte Carlo method, different treatment pathways are inferred and calculated in a simulation environment. This step aims to predict the distribution of results under each treatment pathway, including efficacy, cost, and patient satisfaction. The generated result distribution provides a scientific basis for selecting the optimal rehabilitation plan.

[0147] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0148] By comprehensively considering the patient's multimodal data and personalized needs, the most suitable rehabilitation plan for each patient is generated to ensure the personalization and effectiveness of the plan;

[0149] Multiple methods improve the accuracy of predictions of future outcomes of different treatment pathways and reduce the risks associated with uncertainty;

[0150] The result distribution provides scientific decision support for doctors and nurses, which helps to select the best rehabilitation plan and improve the quality and efficiency of medical services.

[0151] In order to evaluate the actual effects of different rehabilitation activity programs and select the optimal treatment path, detailed reasoning calculations are required. Based on this, the present invention provides a specific embodiment. The step 305, based on the uncertain model and in combination with the rehabilitation activity program, performs reasoning calculations on different treatment paths in a simulation environment through a Bayesian reasoning method to generate result distributions under different treatment paths. The treatment path is an implementation plan for the rehabilitation activity program, and specifically includes the following steps:

[0152] Step 311: defining a treatment pathway based on the uncertainty model and the rehabilitation activity program, wherein the treatment pathway includes the rehabilitation activity program, an implementation sequence of the rehabilitation activity program, and a timeline;

[0153] This step defines the treatment pathway based on the constructed uncertainty model and the generated rehabilitation activity plan, that is, to clarify the specific content of each rehabilitation activity and its execution order and schedule. For example, it is determined that physical therapy will be conducted in the morning, psychological counseling will be conducted in the afternoon, and skin examination will be conducted once a week. This step ensures the detail and operability of the treatment pathway and provides clear guidance for subsequent simulations.

[0154] Step 312: Initializing a simulation environment based on the multimodal features, and setting the patient's health state to an initial state, wherein the health state includes the physiological parameters and the psychological state information;

[0155] In this step, the simulation environment refers to a virtual computing environment used to simulate the patient's rehabilitation process under different treatment pathways and evaluate the state changes at each step. The initial state refers to the patient's health status when he or she starts rehabilitation activities, including physiological parameters and psychological state information, as the starting point of the simulation environment.

[0156] This step uses the multimodal features of the patient to initialize a simulation environment. In this environment, the patient's initial health state is set, that is, their physical and psychological condition when they start rehabilitation activities. For example, the patient's blood pressure is 120 / 80 mmHg and the anxiety score is 5 points in the initial state. This step provides an accurate starting point for the simulation and ensures the reliability of the simulation results.

[0157] Step 313: Based on the simulation environment and the initial state, the treatment path is executed, and the state distribution of each step is predicted using the Bayesian reasoning method. Based on the state distribution, the health state of the patient is updated, and the result distribution of each step is recorded to generate the result distribution under different treatment paths, wherein the result distribution includes the probability distribution of efficacy, cost, and patient satisfaction.

[0158] In this step, the state distribution refers to the probability distribution of the patient's health state predicted based on the Bayesian inference method in each step, reflecting the possibility of different states.

[0159] In this step, various rehabilitation activities are performed step by step in the simulation environment according to the defined treatment pathway. For example, physical therapy is performed in the morning of the first day, and psychological counseling is performed in the afternoon; after each step is executed, the Bayesian inference method is used to predict the changes in the patient's health status and obtain the state distribution. For example, it is predicted that the patient's blood pressure may drop to 118 / 78mmHg and the anxiety score may drop to 4 points on the next morning; according to the predicted state distribution, the patient's health status is updated. For example, the patient's blood pressure and anxiety score are adjusted in the simulation environment; after each step is executed, the probability distribution of efficacy, cost, and patient satisfaction is recorded. These distributions reflect the possible effects under different treatment pathways; through multiple iterations of the above process, the result distribution under different treatment pathways is finally generated, including the probability distribution of efficacy, cost, and patient satisfaction. This step provides a scientific basis for selecting the optimal treatment pathway.

[0160] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0161] By defining the treatment pathway in detail and taking into account the patient's initial health status, the effects of different rehabilitation activities can be evaluated in a simulated environment to ensure that the generated rehabilitation program is highly personalized and meets the specific needs of each patient;

[0162] Using Bayesian reasoning methods to predict the state distribution of each step significantly improves the accuracy of predicting future health status changes and reduces the risks caused by uncertainty;

[0163] Based on the generated result distribution, doctors and nurses can more scientifically select the optimal rehabilitation plan and improve the quality and efficiency of medical services;

[0164] By recording the probability distribution of efficacy, cost, and patient satisfaction, the effects of different treatment pathways can be comprehensively evaluated, providing valuable data support for future rehabilitation plans.

[0165] In order to ensure the effectiveness and adaptability of the program, dynamic adjustments must be made in actual applications to ensure that the rehabilitation program can be optimized in real time according to the patient's latest condition, providing more personalized and effective treatment. Based on this, the present invention provides a specific embodiment, in which step 105 implements the rehabilitation activity combination implementation strategy, uses a closed-loop feedback mechanism during the implementation process, regularly collects and analyzes the patient's actual response to the rehabilitation activity combination implementation strategy, and dynamically adjusts the rehabilitation activity combination implementation strategy based on the actual response using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy, specifically including the following steps:

[0166] Step 501: Implement the rehabilitation activity combination implementation strategy, regularly collect the actual response of the patient, compare the actual response with the preset effect, identify the effect difference between the actual effect and the preset effect, evaluate the effect difference, and obtain the actual effect evaluation result, wherein the actual response includes the efficacy, side effects and patient satisfaction;

[0167] This step implements specific rehabilitation activities according to the generated rehabilitation activity combination implementation strategy. During the implementation process, the actual feedback from patients, including efficacy, side effects, and patient satisfaction, is collected regularly, such as weekly or monthly; these actual feedbacks are compared with the preset effects to identify the differences between the actual effects and the preset effects. For example, if the preset efficacy is that the wound heals completely within two weeks, and the actual feedback shows that the healing speed is slower, there is a difference in efficacy. Finally, these effect differences are evaluated to obtain the actual effect evaluation results, which provide a basis for subsequent optimization.

[0168] Step 502: Initializing a deep reinforcement learning agent using a deep reinforcement learning algorithm based on the actual effect evaluation result;

[0169] In this step, the deep reinforcement learning agent refers to an intelligent system trained by a deep reinforcement learning algorithm, which can adjust its behavioral strategy according to environmental feedback to optimize the implementation strategy of the rehabilitation activity combination.

[0170] In this step, a deep reinforcement learning agent using a deep reinforcement learning algorithm is initialized based on the actual effect evaluation results. The task of this agent is to continuously learn and optimize the implementation strategy of the combination of rehabilitation activities by interacting with the environment. The initialization process includes setting the state space, action space and reward function of the agent. The state space is such as the patient's physiological parameters and psychological state, the action space is such as different rehabilitation activities, and the reward function is such as improved efficacy, reduced side effects and improved patient satisfaction.

[0171] Step 503: using the deep reinforcement learning agent, through experience replay and gradient descent method, iteratively optimizing the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy;

[0172] In this step, experience replay refers to a memory mechanism where the agent stores past experiences in the experience replay buffer and then randomly extracts samples from it for learning to improve learning efficiency and stability. Gradient descent refers to an optimization algorithm that optimizes model performance by calculating the gradient of the loss function to the network parameters and adjusting the parameters in the opposite direction of the gradient to minimize the loss function.

[0173] This step uses the deep reinforcement learning agent to store its experience of interacting with the environment in the experience replay buffer, such as the change in the therapeutic effect after taking a certain rehabilitation activity; randomly extract samples from the experience replay buffer for learning, avoiding the overfitting problem caused by repeated learning of the same experience; by calculating the gradient of the loss function to the agent's strategy network parameters, and adjusting the parameters in the opposite direction of the gradient to minimize the loss function, the agent's behavior strategy is gradually optimized; the agent iterates the above process multiple times and continuously adjusts the rehabilitation activity combination implementation strategy until it finds the best strategy that can maximize the cumulative reward. The target rehabilitation activity combination implementation strategy obtained in the end is not only optimal in theory, but also can achieve the best effect in practical applications.

[0174] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0175] By collecting patients' actual feedback in real time and dynamically adjusting the implementation strategy of rehabilitation activity combinations, we ensure that each patient can get the most suitable personalized rehabilitation plan.

[0176] Using deep reinforcement learning algorithms, the intelligent agent can continuously optimize the implementation strategy of the combination of rehabilitation activities to maximize the rehabilitation effect, while taking into account the patient's physiological and psychological state to comprehensively improve the quality of rehabilitation;

[0177] Pay attention to the actual reactions and preferences of patients, adjust rehabilitation plans in a timely manner, improve patients' treatment experience and satisfaction, and increase patients' confidence and cooperation in treatment;

[0178] Using a closed-loop feedback mechanism and deep reinforcement learning algorithm, real-time adjustment and optimization of the rehabilitation plan can be achieved, ensuring that the plan always meets the patient's latest condition and is highly adaptable.

[0179] In order to further optimize the rehabilitation program and adapt to the individual differences of patients, it is necessary to introduce an intelligent learning mechanism. Based on this, the present invention provides a specific embodiment, wherein step 503 uses the deep reinforcement learning agent to iteratively optimize the rehabilitation activity combination implementation strategy through experience replay and gradient descent method to obtain the target rehabilitation activity combination implementation strategy, which specifically includes the following steps:

[0180] Step 511: using the deep reinforcement learning agent to interact with the environment to generate experience data, and storing the experience data in an experience replay buffer, wherein the experience data includes state, action, and reward;

[0181] In this step, the experience replay buffer refers to a storage mechanism for saving the experience data of the agent's interaction with the environment. By randomly sampling these experience data for learning, the learning efficiency and stability can be improved.

[0182] In this step, the deep reinforcement learning agent interacts with the environment. During each interaction, the agent selects an action, i.e., a specific rehabilitation activity, based on the current state, such as the patient's physiological parameters and psychological state information; records the changes in the environment after the action is performed, records the new state, the action taken, and the reward obtained, such as improved efficacy, reduced side effects, and improved patient satisfaction; these data are stored in the experience playback buffer as an experience data to provide material for subsequent learning.

[0183] Step 512: randomly extracting target experience data from the experience replay buffer, calculating a target value of the target experience data, comparing the target value with a predicted value of the policy network in the deep reinforcement learning agent, and calculating a loss function;

[0184] In this step, the target value refers to the ideal value calculated based on the target experience data, which is used to evaluate the quality of the agent's current strategy and guide the update of the strategy network parameters.

[0185] This step randomly extracts several pieces of target experience data from the experience replay buffer. For each piece of target experience data, calculate its target value, that is, the ideal result. For example, if the target experience data shows that a certain rehabilitation activity significantly improves the therapeutic effect, the target value reflects this positive change; compare the target value with the predicted value of the state by the policy network in the agent, and calculate the difference between the two, that is, the loss function. The loss function measures the quality of the agent's current strategy and provides a basis for subsequent parameter updates.

[0186] Step 513: Based on the loss function, the parameters of the strategy network are updated using the gradient descent method to obtain updated strategy network parameters, and based on the updated strategy network parameters, the rehabilitation activity combination implementation strategy is optimized to obtain the target rehabilitation activity combination implementation strategy;

[0187] In this step, the policy network refers to the neural network in the deep reinforcement learning agent, which is used to decide the best action to take in a given state, that is, the recovery activity, and continuously optimizes to maximize the long-term reward.

[0188] This step uses the gradient descent method to calculate the gradient of the loss function to the policy network parameters based on the calculated loss function, and adjusts the parameters in the opposite direction of the gradient to minimize the loss function. This step gradually optimizes the agent's behavior strategy; based on the updated policy network parameters, the agent re-evaluates and optimizes the rehabilitation activity combination implementation strategy. For example, the agent may adjust the execution time, frequency, or intensity of certain rehabilitation activities to better adapt to the patient's latest condition and ensure that the program achieves the best effect in practical application; after multiple iterations of the above process, the target rehabilitation activity combination implementation strategy generated is not only theoretically optimal, but also can be continuously optimized in practical applications to meet the personalized needs of patients.

[0189] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0190] Through the interaction and real-time feedback between the intelligent body and the environment, the implementation strategy of the rehabilitation activity combination is dynamically adjusted to ensure that each patient can obtain the most suitable personalized rehabilitation plan;

[0191] Using deep reinforcement learning algorithms, the intelligent agent can continuously optimize the implementation strategy of the combination of rehabilitation activities to maximize the rehabilitation effect, while taking into account the patient's physiological and psychological state to comprehensively improve the quality of rehabilitation;

[0192] Pay attention to the patient's actual reactions and preferences, adjust the rehabilitation plan in a timely manner, improve the patient's treatment experience and satisfaction, and increase the patient's confidence and cooperation in treatment.

[0193] Real-time adjustment and optimization of the rehabilitation program is achieved to ensure that the program always meets the patient's latest situation and is highly adaptable.

[0194] Figure 2 The present invention provides a structural diagram of a rehabilitation nursing planning system for patients with plastic burns, such as Figure 2 As shown, the system includes:

[0195] The acquisition module 21 is used to collect the patient's physiological parameters, psychological state information and skin healing status, obtain the patient's basic information using multimodal data fusion technology, and use an integrated learning algorithm to comprehensively analyze the patient's basic information to obtain a health status score;

[0196] A generating module 22, for generating a rehabilitation activity program with expected effect simulation by using a genetic algorithm based on the health status score and the preset personalized rehabilitation goal;

[0197] A construction module 23 is used to construct an uncertainty model in the patient's rehabilitation process using a Bayesian network, and in combination with the rehabilitation activity plan, use the uncertainty model to predict the distribution of results under different treatment paths;

[0198] An optimization module 24 is used to apply a mixed integer programming algorithm to globally optimize the rehabilitation activity program in combination with a preset optimization goal, constraint conditions and the result distribution, so as to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan of the optimal rehabilitation activity combination;

[0199] The adjustment module 25 is used to implement the rehabilitation activity combination implementation strategy. During the implementation process, a closed-loop feedback mechanism is used to regularly collect and analyze the actual responses of patients to the rehabilitation activity combination implementation strategy. Based on the actual responses, the rehabilitation activity combination implementation strategy is dynamically adjusted using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy.

[0200] Figure 2 The rehabilitation nursing planning system for patients with plastic burns can be implemented Figure 1 The implementation principle and technical effect of the rehabilitation nursing planning method for patients with plastic burns described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the rehabilitation nursing planning system for patients with plastic burns in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0201] In one possible design, Figure 2 A rehabilitation care planning system for a patient with plastic burns according to the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0202] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0203] The processing component 32 is used to: collect the patient's physiological parameters, psychological state information and skin healing status, use multimodal data fusion technology to obtain the patient's basic information, use an integrated learning algorithm to comprehensively analyze the patient's basic information to obtain a health status score; based on the health status score and the preset personalized rehabilitation goal, use a genetic algorithm to generate a rehabilitation activity plan with an expected effect simulation; use a Bayesian network to construct an uncertainty model in the patient's rehabilitation process, and combine the rehabilitation activity plan to use the uncertainty model to predict the result distribution under different treatment paths; combine the preset optimization goals, constraints and the result distribution, use a mixed integer programming algorithm to globally optimize the rehabilitation activity plan to obtain a rehabilitation activity combination implementation strategy, and the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination; implement the rehabilitation activity combination implementation strategy, use a closed-loop feedback mechanism during the implementation process, regularly collect and analyze the patient's actual response to the rehabilitation activity combination implementation strategy, and dynamically adjust the rehabilitation activity combination implementation strategy based on the actual response using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy.

[0204] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0205] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0206] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0207] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0208] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0209] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0210] The embodiment of the present invention further provides a computer storage medium storing a computer program, which can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for planning rehabilitation care for patients with plastic burns.

[0211] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0212] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0213] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0214] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rehabilitation nursing planning method for patients with plastic burns, characterized in that: include: Collect the patient's physiological parameters, psychological status information and skin healing status, use multimodal data fusion technology to obtain the patient's basic information, use an integrated learning algorithm to comprehensively analyze the patient's basic information and obtain a health status score; Based on the health status score and the preset personalized rehabilitation goal, a rehabilitation activity plan with expected effect simulation is generated by using a genetic algorithm; Using a Bayesian network to construct an uncertainty model in the patient's rehabilitation process, combined with the rehabilitation activity plan, using the uncertainty model to predict the distribution of results under different treatment pathways; In combination with the preset optimization goal, constraint conditions and the result distribution, a mixed integer programming algorithm is applied to globally optimize the rehabilitation activity program to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination; Implement the rehabilitation activity combination implementation strategy, use a closed-loop feedback mechanism during the implementation process, regularly collect and analyze patients' actual responses to the rehabilitation activity combination implementation strategy, and based on the actual responses, use a deep reinforcement learning algorithm to dynamically adjust the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy.

2. The method according to claim 1, characterized in that The method combines the preset optimization goal, constraint conditions and result distribution, applies a mixed integer programming algorithm to globally optimize the rehabilitation activity program, and obtains a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan of the optimal rehabilitation activity combination, including: Determine the decision variables in the process of generating the rehabilitation activity combination implementation strategy, and define the objective function to determine the optimization target, wherein the decision variables include the execution time, frequency and intensity of each rehabilitation activity, and the objective function includes maximizing the rehabilitation effect, minimizing the treatment cost and improving the patient satisfaction; defining constraints, wherein the constraints include actual medical resource constraints, patient preference constraints, and physiological constraints; Based on the decision variables, the objective function and the constraint conditions, a mixed integer programming algorithm is applied to construct a mixed integer programming model; Using a target mixed integer programming solver to solve the mixed integer programming model, to obtain a target solution that satisfies the constraint conditions and has an optimal objective function; Based on the target solution and the result distribution, the expected effects of different rehabilitation activities within the target solution are evaluated to obtain expected effect evaluation results, and based on the expected effect evaluation results, a rehabilitation activity combination implementation strategy is generated; the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination, and the implementation plan includes the execution time, frequency and intensity of each rehabilitation activity in the optimal rehabilitation activity combination.

3. The method according to claim 2, characterized in that The method of using a target mixed integer programming solver to solve the mixed integer programming model to obtain a target solution that satisfies the constraint conditions and has an optimal target function includes: Selecting an initial mixed integer programming solver, and setting parameters of the initial mixed integer programming solver according to actual needs to obtain a target mixed integer programming solver, wherein the solver parameters include a solution time limit and a solution accuracy requirement; Based on the target mixed integer programming solver, the mixed integer programming model is processed by branch and bound method to obtain a preliminary feasible solution and a corresponding bound, wherein the bound is a lower bound or an upper bound; Using the preliminary feasible solution and the limit, the mixed integer programming model is processed by a cutting plane method, and regions that do not contain integer solutions are excluded by adding constraint conditions to narrow the search space and obtain an optimized solution space; Based on the optimized solution space, the mixed integer programming model is solved, and a target solution that satisfies all constraints and optimizes the objective function is selected. If the target solution cannot be selected within a preset time, the best feasible solution currently selected within the preset time is returned, and solution quality information of the best feasible solution is extracted. Based on the solution quality information, a target solution is generated, and the solution quality information includes a difference between the best feasible solution and the target solution.

4. The method according to claim 1, characterized in that: The use of the Bayesian network to construct an uncertainty model in the patient's rehabilitation process, combined with the rehabilitation activity plan, and using the uncertainty model to predict the distribution of results under different treatment paths, includes: Extracting features from the patient's basic information to obtain multimodal features, using a graph attention network and a structured learning algorithm to identify and define the multimodal features, generate nodes and edges, and obtain a directed acyclic graph based on the nodes and edges, wherein the nodes are various factors in the patient's rehabilitation process, including the physiological parameters, psychological state information, and skin healing conditions, and the edges are dependencies between the nodes; Using Gaussian mixture model and expectation maximization algorithm, the conditional probability distribution of historical data of each node is estimated to obtain the preliminary conditional probability distribution; Combining the variational inference method and the Bayesian optimization framework, adjusting the parameters of the Gaussian mixture model to optimize the preliminary conditional probability distribution and generate a target conditional probability distribution; Based on the directed acyclic graph and the target conditional probability distribution, using a Bayesian network structure learning algorithm to construct an uncertainty model in the patient's rehabilitation process; Based on the uncertain model and in combination with the rehabilitation activity program, the Bayesian reasoning method is used to perform reasoning calculations on different treatment pathways in a simulation environment to generate result distributions under different treatment pathways, where the treatment pathway is an implementation plan for the rehabilitation activity program.

5. The method according to claim 4, characterized in that Based on the uncertain model and in combination with the rehabilitation activity program, the Bayesian reasoning method is used to perform reasoning calculations on different treatment pathways in a simulation environment to generate result distributions under different treatment pathways, wherein the treatment pathway is an implementation plan for the rehabilitation activity program, including: Based on the uncertainty model and the rehabilitation activity plan, defining a treatment pathway, wherein the treatment pathway includes the rehabilitation activity plan, an implementation sequence of the rehabilitation activity plan, and a timeline; Initializing a simulation environment based on the multimodal features, setting the patient's health state as an initial state, the health state including the physiological parameters and the psychological state information; Based on the simulation environment and the initial state, the treatment path is executed, and the state distribution of each step is predicted using the Bayesian reasoning method. Based on the state distribution, the patient's health status is updated, and the result distribution of each step is recorded to generate the result distribution under different treatment paths, wherein the result distribution includes the probability distribution of efficacy, cost, and patient satisfaction.

6. The method according to claim 1, characterized in that The implementation of the rehabilitation activity combination implementation strategy uses a closed-loop feedback mechanism during the implementation process to regularly collect and analyze the actual response of patients to the rehabilitation activity combination implementation strategy, and based on the actual response, dynamically adjusts the rehabilitation activity combination implementation strategy using a deep reinforcement learning algorithm to obtain a target rehabilitation activity combination implementation strategy, including: Implement the rehabilitation activity combination implementation strategy, regularly collect the actual feedback from patients, compare the actual feedback with the preset effect, identify the effect difference between the actual effect and the preset effect, evaluate the effect difference, and obtain the actual effect evaluation result, wherein the actual feedback includes the efficacy, side effects and patient satisfaction; Initializing a deep reinforcement learning agent using a deep reinforcement learning algorithm based on the actual effect evaluation result; The deep reinforcement learning agent is used to iteratively optimize the rehabilitation activity combination implementation strategy through experience replay and gradient descent method to obtain the target rehabilitation activity combination implementation strategy.

7. The method according to claim 6, characterized in that The deep reinforcement learning agent is used to iteratively optimize the rehabilitation activity combination implementation strategy through experience replay and gradient descent method to obtain the target rehabilitation activity combination implementation strategy, including: Utilizing the deep reinforcement learning agent to interact with the environment to generate experience data, and storing the experience data in an experience replay buffer, wherein the experience data includes states, actions, and rewards; Randomly extracting target experience data from the experience replay buffer, calculating a target value of the target experience data, comparing the target value with a predicted value of a policy network in the deep reinforcement learning agent, and calculating a loss function; Based on the loss function, the parameters of the strategy network are updated using the gradient descent method to obtain updated strategy network parameters, and based on the updated strategy network parameters, the rehabilitation activity combination implementation strategy is optimized to obtain the target rehabilitation activity combination implementation strategy.

8. A rehabilitation nursing planning system for patients with plastic burns, characterized in that: include: The acquisition module is used to collect the patient's physiological parameters, psychological status information and skin healing status, obtain the patient's basic information using multimodal data fusion technology, and use an integrated learning algorithm to comprehensively analyze the patient's basic information to obtain a health status score; A generation module, for generating a rehabilitation activity program with expected effect simulation by using a genetic algorithm based on the health status score and the preset personalized rehabilitation goal; A construction module is used to construct an uncertainty model in the patient's rehabilitation process using a Bayesian network, and in combination with the rehabilitation activity plan, use the uncertainty model to predict the distribution of results under different treatment paths; an optimization module, for applying a mixed integer programming algorithm to globally optimize the rehabilitation activity program in combination with a preset optimization goal, constraint conditions and the result distribution, to obtain a rehabilitation activity combination implementation strategy, wherein the rehabilitation activity combination implementation strategy includes an optimal rehabilitation activity combination and an implementation plan for the optimal rehabilitation activity combination; The adjustment module is used to implement the rehabilitation activity combination implementation strategy. During the implementation process, a closed-loop feedback mechanism is used to regularly collect and analyze the actual responses of patients to the rehabilitation activity combination implementation strategy. Based on the actual responses, a deep reinforcement learning algorithm is used to dynamically adjust the rehabilitation activity combination implementation strategy to obtain a target rehabilitation activity combination implementation strategy.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a rehabilitation nursing planning method for patients with plastic burns as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a rehabilitation nursing planning method for patients with plastic burns as described in any one of claims 1 to 7 is implemented.

Images