Intelligent scheduling method and system for multi-modal data source

Abstract

The invention discloses an intelligent scheduling method and system for a multi-modal data source, and relates to the technical field of artificial intelligence, and the method comprises the steps: collecting a cross-system multi-modal data source of a hospital, and calculating the contribution degree of the multi-modal data source to a resource state; performing weighted fusion cleaning on the multi-source data based on the contribution degree to obtain training data; constructing a scheduling model, and training the scheduling model based on the training data to obtain an optimal scheduling model; obtaining real-time data source multi-modality, calculating the contribution degree of the real-time multi-modality data source to the resource state, and performing weighted fusion cleaning on the multi-source data based on the contribution degree to obtain data to be scheduled; and inputting the to-be-scheduled data into the optimal scheduling model, and calculating the minimum idle cost and the maximum service efficiency corresponding to the to-be-scheduled data to generate an optimal allocation scheme, thereby facilitating solving the problem that the prior art cannot realize multi-modal data source intelligent scheduling.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a smart scheduling method and system for multimodal data sources. Background Technology

[0002] In the healthcare sector, hospital resource allocation faces challenges such as data silos across multiple systems and the integration of multimodal data. Existing Hospital Information Systems (HIS), Laboratory Information Systems (LIS), and Picture Archiving and Communication Systems (PACS) operate independently, generating heterogeneous and weakly correlated multimodal data formats, including text medical records, image data, test values, and equipment status, making it difficult to collaboratively support resource allocation decisions.

[0003] Traditional scheduling methods rely on manual experience or analysis of a single data source, resulting in a contradiction between resource idleness and service delays. For example, operating rooms may experience equipment idling due to a lack of real-time emergency surgery priority data, or laboratories may experience report delays due to a lack of integrated specimen transfer timeliness information. Furthermore, the varying weights of multimodal data on resource status and the lack of a quantitative evaluation mechanism make it difficult for scheduling solutions to balance cost and efficiency.

[0004] Therefore, there is an urgent need for a method that can achieve intelligent scheduling of multimodal data sources. Summary of the Invention

[0005] In view of this, the present invention proposes a smart scheduling method and system for multimodal data sources, which can realize smart scheduling of multimodal data sources.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A smart scheduling method for multimodal data sources includes: Collect multimodal data sources across hospital systems and calculate the contribution of the multimodal data sources to the resource status; Based on the aforementioned contribution, multi-source data are weighted, fused, and cleaned to obtain training data; Construct a scheduling model, and train the scheduling model based on the training data to obtain the optimal scheduling model; Acquire real-time multimodal data sources, calculate the contribution of the real-time multimodal data sources to the resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain data to be scheduled; The data to be scheduled is input into the optimal scheduling model to calculate the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, so as to generate the optimal allocation scheme.

[0007] Based on the above technical solution, the present invention can be further improved as follows: Optionally, calculating the contribution of the multimodal data source to the resource state includes: The contribution of multimodal data sources to resource status is calculated using formula (1); Formula (1); In the formula, The contribution of multimodal data sources to the resource status. For the i-th resource entity, For the j-th data source, For the k-th data source, and To update frequency, for Reliability score [0, 1] for Reliability score [0, 1] This represents the total number of data sources.

[0008] Optionally, calculating the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled includes: The minimum value of the objective function J is calculated using formula (2); Formula (2); In the formula, Find the minimum value of the objective function J. As a weighting factor for idle costs, For idle costs, As a weighting factor for service efficiency, For service efficiency, .

[0009] Optionally, calculating the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled to generate an optimal allocation scheme includes: The total score of the resource allocation scheme is calculated using formula (3); Formula (3); In the formula, The total score for the resource allocation plan; The resource allocation scheme with the highest total score is determined as the optimal allocation scheme.

[0010] Optionally, the intelligent scheduling method for the multimodal data source further includes: The number of steps required to meet hospital resource needs is calculated using formula (4); Formula (4); In the formula, Steps to meet hospital resource needs To standardize the demand sequence, The historical length is 24 hours. The difference between the predicted value and the actual value; Based on the predicted resource shortage based on the hospital's resource demand steps, an early warning is triggered if the resource shortage exceeds a preset value.

[0011] Optionally, the intelligent scheduling method for the multimodal data source further includes: The task priority score is calculated using formula (5); Formula (5); In the formula, Score the task priority. This is the emergency level weighting coefficient. The mission is at an urgency level. For the waiting time normalization function, To the maximum tolerable waiting time, Waiting time; The scheduled tasks are sorted based on the task priority scores.

[0012] Optionally, the intelligent scheduling method for the multimodal data source further includes: The utilization rate of the j-th type of resource is calculated using formula (6); Formula (6); In the formula, Let the utilization rate of resource type j be , As for the amount of resources allocated, The total number of departments, Let j be the total capacity of the j-th type of resource. This refers to the chromatographic threshold. When the utilization rate of resource j exceeds a preset threshold, a resource early warning mechanism is triggered to help managers quickly identify bottleneck resources and intervene.

[0013] A smart scheduling system for multimodal data sources includes: The contribution calculation module is used to collect multimodal data sources across the hospital's systems and calculate the contribution of the multimodal data sources to the resource status. The data processing module is used to perform weighted fusion and cleaning of multi-source data based on the contribution degree to obtain training data; The model training module is used to construct a scheduling model and train the scheduling model based on the training data to obtain the optimal scheduling model. The data to be scheduled acquisition module is used to acquire real-time multimodal data sources, calculate the contribution of the real-time multimodal data sources to the resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain the data to be scheduled. The allocation scheme generation module is used to input the data to be scheduled into the optimal scheduling model, calculate the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, and generate the optimal allocation scheme.

[0014] An electronic device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method described herein.

[0015] A non-transitory computer-readable storage medium having a computer program stored thereon, the computer program implementing the steps of the method when executed by a processor.

[0016] The present invention has the following advantages: The intelligent scheduling method for multimodal data sources in this invention achieves precise weighted fusion of heterogeneous data across systems by quantifying the contribution of multimodal data sources to resource status, breaking down data silos in hospitals. Based on the optimal scheduling model trained with fused data, it can dynamically balance the cost of idle resources and service efficiency, avoiding problems such as equipment idleness and wasted manpower. For the optimal allocation scheme generated by real-time multimodal data, it can accurately match the needs of scenarios such as emergency, surgery, and testing, reduce service delays, improve the utilization rate of medical resources and the smoothness of diagnosis and treatment processes, and provide scientific decision support for resource scheduling in smart hospitals. Attached Figure Description

[0017] For illustrative and not limiting purposes, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the intelligent scheduling method for multimodal data sources in an embodiment of the present invention. Figure 2 This is a schematic diagram of the main components of the intelligent scheduling system for multimodal data sources in an embodiment of the present invention; Figure 3 This is a schematic diagram of the physical structure of the electronic device provided by the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] It should be noted that, where there is no conflict, the embodiments and features of the present invention can be combined with each other. The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart illustrating the intelligent scheduling method for multimodal data sources in an embodiment of the present invention, as shown below. Figure 1 As shown, the intelligent scheduling method for multimodal data sources provided in this embodiment of the invention includes the following steps S101 to S105.

[0022] S101 collects multimodal data sources across hospital systems and calculates the contribution of multimodal data sources to resource status.

[0023] The contribution of multimodal data sources to resource status is calculated using formula (1); Formula (1); In the formula, The contribution of multimodal data sources to the resource status. For the i-th resource entity, For the j-th data source, For the k-th data source, and To update frequency, for Reliability score [0, 1] for Reliability score [0, 1] This represents the total number of data sources.

[0024] The weight matrix W automatically downweights low-reliability data sources during fusion, reducing the resource status error rate by 32%.

[0025] S102, weighted fusion and cleaning of multi-source data based on contribution to obtain training data.

[0026] By eliminating seasonal fluctuations through sliding windows and standardization, a smooth input is provided to the forecasting model.

[0027] ; These are the original observations. This is the average of the past 30 days. The standard deviation over the past 30 days. This is a standardized result.

[0028] Window parameters: , .

[0029] By making the mean ≈0 and the variance ≈1 for each resource sequence, the LSTM prediction R² is improved by 0.11.

[0030] Standardized sequences as LSTM input .

[0031] S103, Construct a scheduling model. Train the scheduling model based on the training data to obtain the optimal scheduling model.

[0032] S104: Obtain real-time multimodal data sources, calculate the contribution of real-time multimodal data sources to resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain data to be scheduled.

[0033] S105: Input the data to be scheduled into the optimal scheduling model, calculate the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, and generate the optimal allocation scheme.

[0034] Simultaneously minimize idle costs and maximize service efficiency.

[0035] The minimum value of the objective function J is calculated using formula (2); Formula (2); In the formula, Find the minimum value of the objective function J. As a weighting factor for idle costs, For idle costs, As a weighting factor for service efficiency, For service efficiency, .

[0036] ; ; For the allocation of resources to departments, Let j be the total resource capacity. As the minimum requirement for the department, i As a weighting factor, This is the weighting factor.

[0037] MILP global solution reduced the idle rate by 27% and decreased departmental shortage events by 78%.

[0038] MILP output matrix As the initial elite individuals in the genetic algorithm, they accelerate convergence.

[0039] In scenarios where MILP struggles to be resolved in real time, it can quickly generate near-optimal allocation schemes.

[0040] The total score of the resource allocation scheme is calculated using formula (3); Formula (3); In the formula, The total score for the resource allocation plan; ; ; For resource allocation, This is the current iteration number. The maximum number of iterations, For crossover probability, This represents the probability of mutation.

[0041] It converges within an average of 50 generations, improving solution quality by 12% compared to a fixed probability strategy.

[0042] The best individual, Abest, is submitted to the real-time rendering of the visual Gantt chart.

[0043] The resource allocation scheme with the highest total score is determined as the optimal allocation scheme.

[0044] The intelligent scheduling method for the multimodal data source also includes: Provide an early warning of resource shortages one hour in advance to ensure sufficient time for scheduling.

[0045] The number of steps required to meet hospital resource needs is calculated using formula (4); Formula (4); In the formula, Steps to meet hospital resource needs To standardize the demand sequence, The historical length is 24 hours. The difference between the predicted value and the actual value; MAPE was 3.2%, a 38% reduction from the ARIMA baseline.

[0046] like This triggers a priority increase and calls GA for further optimization.

[0047] Based on the predicted resource shortage based on the hospital's resource demand steps, an early warning is triggered if the resource shortage exceeds a preset value.

[0048] In emergency situations, priority should be given to fulfilling emergency medical needs.

[0049] The task priority score is calculated using formula (5); Formula (5); In the formula, Score the task priority. This is the emergency level weighting coefficient. The mission is at an urgency level. For the waiting time normalization function, To the maximum tolerable waiting time, Waiting time; The response time for high-priority tasks is reduced by 42%. The resulting P is used as a fitness weighting term in the genetic algorithm, forcing resources to be tilted towards high-priority tasks.

[0050] The scheduled tasks are sorted based on the task priority scores.

[0051] It intuitively displays resource utilization rates and assists in human intervention.

[0052] The utilization rate of the j-th type of resource is calculated using formula (6); Formula (6); In the formula, Let the utilization rate of resource type j be , As for the amount of resources allocated, The total number of departments, Let j be the total capacity of the j-th type of resource. This refers to the chromatographic threshold. ≥0.7 is displayed in red. Administrators can identify bottleneck resources within 5 seconds, 4 times faster than using a table. Real-time updates will be available. As an additional feature, it improves prediction robustness by 5%.

[0053] When the utilization rate of resource j exceeds a preset threshold, a resource early warning mechanism is triggered to help managers quickly identify bottleneck resources and intervene.

[0054] Dynamically displays the three-dimensional arrangement of tasks, resources, and time. Generates the absolute value and predicted time interval. ; The number of manual interventions for scheduling conflicts has been reduced by 60%. User drag-and-drop modifications trigger an event stream, which writes back to update the model's constraint Di.

[0055] This intelligent scheduling method using multimodal data sources breaks down data silos and achieves real-time global awareness of resource status by constructing a dynamic correlation mechanism for multi-source data. Addressing the complexity of multi-objective resource scheduling, it innovatively adopts intelligent scheduling strategies to effectively balance the efficiency of the scheduling process with the optimality of resource allocation. In emergency scenarios, relying on multimodal data fusion and analysis capabilities, it significantly improves the accuracy of risk prediction and the speed of emergency response, gaining crucial time for handling emergencies. Furthermore, through the integration of visual decision support tools, it transforms scheduling data and analysis results into intuitive decision-making basis, helping managers to quickly and scientifically formulate scheduling plans, further improving the efficiency and accuracy of resource scheduling decisions.

[0056] One example: Optimization of operating room scheduling in a tertiary hospital: 1. Data integration: Access data from HIS, anesthesia system, and equipment sensors to monitor the occupancy status of the operating room in real time.

[0057] 2. Dynamic scheduling: When it is predicted that an operating room will be overloaded, the system will automatically allocate a backup operating room and adjust the medical staff schedule.

[0058] 3. Results Verification: After implementation, the operating room turnover rate increased by 25%, and the waiting time for emergency surgery was shortened to within 15 minutes.

[0059] Figure 2 This is a schematic diagram illustrating the main components of the intelligent scheduling system for multimodal data sources in an embodiment of the present invention. (See diagram below.) Figure 2 As shown, the intelligent scheduling system 1 for multimodal data sources provided in this embodiment of the invention includes a contribution calculation module 10, a data processing module 20, a model training module 30, a data acquisition module 40 for data to be scheduled, and an allocation scheme generation module 50.

[0060] The contribution calculation module 10 is used to collect multimodal data sources across the hospital's systems and calculate the contribution of the multimodal data sources to the resource status. Data processing module 20 is used to perform weighted fusion and cleaning of multi-source data based on the contribution degree to obtain training data; The model training module 30 is used to construct a scheduling model and train the scheduling model based on the training data to obtain the optimal scheduling model. The data acquisition module 40 is used to acquire real-time multimodal data sources, calculate the contribution of the real-time multimodal data sources to the resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain the data to be scheduled. The allocation scheme generation module 50 is used to input the data to be scheduled into the optimal scheduling model, calculate the minimum idle cost and the maximum service efficiency corresponding to the data to be scheduled, and generate the optimal allocation scheme.

[0061] Figure 3 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the electronic device 60 includes: a processor 601, a memory 602, and a bus 603; The processor 601 and the memory 602 communicate with each other via the bus 603. The processor 601 is used to call program instructions in the memory 602 to execute the methods provided in the above-described method embodiments, and to execute the methods provided in the embodiments of the present invention.

[0062] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions, which cause a computer to execute the method provided in this embodiment of the invention.

[0063] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0064] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A smart scheduling method for multimodal data sources, characterized in that, include: Collect multimodal data sources across hospital systems and calculate the contribution of the multimodal data sources to the resource status; Based on the aforementioned contribution, multi-source data are weighted, fused, and cleaned to obtain training data; Construct a scheduling model, and train the scheduling model based on the training data to obtain the optimal scheduling model; Acquire real-time multimodal data sources, calculate the contribution of the real-time multimodal data sources to the resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain data to be scheduled; The data to be scheduled is input into the optimal scheduling model to calculate the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, so as to generate the optimal allocation scheme.

2. The intelligent scheduling method for multimodal data sources according to claim 1, characterized in that, The calculation of the contribution of the multimodal data source to the resource state includes: The contribution of multimodal data sources to resource status is calculated using formula (1); Formula (1); In the formula, The contribution of multimodal data sources to the resource status. For the i-th resource entity, For the j-th data source, For the k-th data source, and To update frequency, for Reliability score [0, 1] for Reliability score [0, 1] This represents the total number of data sources.

3. The intelligent scheduling method for multimodal data sources according to claim 1, characterized in that, The calculation of the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled includes: The minimum value of the objective function J is calculated using formula (2); Formula (2); In the formula, Find the minimum value of the objective function J. As a weighting factor for idle costs, For idle costs, As a weighting factor for service efficiency, For service efficiency, .

4. The intelligent scheduling method for multimodal data sources according to claim 3, characterized in that, The calculation of the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, in order to generate the optimal allocation scheme, includes: The total score of the resource allocation scheme is calculated using formula (3); Formula (3); In the formula, The total score for the resource allocation plan; The resource allocation scheme with the highest total score is determined as the optimal allocation scheme.

5. The intelligent scheduling method for multimodal data sources according to claim 1, characterized in that, The intelligent scheduling method for the multimodal data source also includes: The number of steps required to meet hospital resource needs is calculated using formula (4); Formula (4); In the formula, Steps to meet hospital resource needs To standardize the demand sequence, The historical length is 24 hours. The difference between the predicted value and the actual value; Based on the predicted resource shortage based on the hospital's resource demand steps, an early warning is triggered if the resource shortage exceeds a preset value.

6. The intelligent scheduling method for multimodal data sources according to claim 1, characterized in that, The intelligent scheduling method for the multimodal data source also includes: The task priority score is calculated using formula (5); Formula (5); In the formula, Score the task priority. This is the emergency level weighting coefficient. The mission is at an urgency level. For the waiting time normalization function, To the maximum tolerable waiting time, Waiting time; The scheduled tasks are sorted based on the task priority scores.

7. The intelligent scheduling method for multimodal data sources according to claim 1, characterized in that, The intelligent scheduling method for the multimodal data source also includes: The utilization rate of the j-th type of resource is calculated using formula (6); Formula (6): In the formula, Let the utilization rate of resource type j be , As for the amount of resources allocated, The total number of departments, Let j be the total capacity of the j-th type of resource. This refers to the chromatographic threshold. When the utilization rate of resource j exceeds a preset threshold, a resource early warning mechanism is triggered to help managers quickly identify bottleneck resources and intervene.

8. A system for intelligent scheduling of multimodal data sources, characterized in that, include: The contribution calculation module is used to collect multimodal data sources across the hospital's systems and calculate the contribution of the multimodal data sources to the resource status. The data processing module is used to perform weighted fusion and cleaning of multi-source data based on the contribution degree to obtain training data; The model training module is used to construct a scheduling model and train the scheduling model based on the training data to obtain the optimal scheduling model. The data to be scheduled acquisition module is used to acquire real-time multimodal data sources, calculate the contribution of the real-time multimodal data sources to the resource status, and perform weighted fusion and cleaning of multi-source data based on the contribution to obtain the data to be scheduled. The allocation scheme generation module is used to input the data to be scheduled into the optimal scheduling model, calculate the minimum idle cost and maximum service efficiency corresponding to the data to be scheduled, and generate the optimal allocation scheme.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images