A hospital appointment registration resource dynamic allocation method and system and a storage medium

Abstract

The application discloses a hospital appointment registration resource dynamic allocation method and system and a storage medium, and comprises the following steps: establishing a registration resource item and generating resource state data according to scheduling, number source configuration, historical appointment data, demand estimation parameters and scheme evaluation parameters; estimating the appointment demand in combination with the current appointment request volume, and generating a candidate resource allocation scheme according to the reception capacity and adjustment limit; calculating the demand satisfaction volume, the number source utilization volume and the resource adjustment volume, screening a target resource allocation scheme and executing appointment allocation; generating demand feedback results and evaluation feedback results according to actual appointment results, correcting the demand estimation parameters and the scheme evaluation parameters by using an improved SPSA algorithm with the CRITIC weighting method, and generating a target resource allocation scheme in the next allocation period. The application forms a parameter feedback closed loop based on actual appointment results, so that the registration resource allocation is updated according to the supply and demand state changes.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, system and storage medium for dynamic allocation of hospital appointment registration resources. Background Technology

[0002] Hospital appointment registration resource allocation falls under the cross-application field of medical informatics and computer resource scheduling, primarily focusing on the allocation of appointment slots and the processing of appointment requests among outpatient departments, doctors, and appointment times. Existing systems mostly allocate available appointment slots for different departments, doctors, and time slots based on scheduling plans, doctor capacity, historical appointment records, and current appointment requests, and process appointments using reserved slots, waitlist slots, and temporary adjustment rules. With the increase in appointment channels and the greater fluctuation in patient demand, appointment resource allocation has gradually shifted from static allocation under fixed schedules to a process of periodically assessing demand and adjusting the number of available slots.

[0003] Existing technologies primarily employ a combination of demand forecasting and solution evaluation. The demand forecasting stage estimates future appointment demand based on historical appointment volumes and current request volumes. The solution generation stage generates different appointment allocation results based on patient capacity and adjustment constraints. The solution evaluation stage scores each allocation result according to demand satisfaction, appointment utilization, and adjustment costs, selecting the highest-scoring solution for implementation. Some systems also record actual appointment volumes and remaining appointment slots for operational statistics and manual adjustments; however, demand estimation parameters and solution evaluation parameters are mostly pre-set during the system deployment phase.

[0004] In actual operation, the appointment demand of different departments, doctors, and time periods changes with outpatient schedules and patient choices. When demand estimation parameters are not adjusted based on actual request and appointment volumes, the demand estimation results in subsequent cycles are prone to deviating from the true supply and demand situation. When scheme evaluation parameters are not adjusted in conjunction with the demand gap and remaining appointment slots after the implementation of the target scheme, the candidate scheme scoring will still use existing evaluation preferences. The current process mainly utilizes actual appointment results for statistical display, and has not yet formed a parameter correction link that simultaneously feeds back to the demand estimation process and the scheme evaluation process, making it difficult for the resource allocation scheme in the next allocation cycle to be continuously adjusted according to the actual supply and demand situation.

[0005] Therefore, how to provide a method, system, and storage medium for the dynamic allocation of hospital appointment registration resources is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a method, system, and storage medium for dynamic allocation of hospital appointment registration resources. This invention utilizes feedback from actual appointment results to correct demand estimation parameters and scheme evaluation parameters, so that the registration resource scheme in subsequent allocation cycles is continuously adjusted according to the actual supply and demand status.

[0007] A method for dynamically allocating hospital appointment registration resources according to an embodiment of the present invention includes: Read the schedule, appointment allocation, historical appointment data, demand estimation parameters and scheme evaluation parameters, and establish registration resource items by department, doctor and time period to obtain resource status data; Collect the current number of reservation requests, and calculate the reservation demand estimate for each registration resource item based on resource status data and demand estimation parameters; Read the patient reception capacity and adjustment limits, calculate the number of available appointment slots, compare the estimated appointment demand with the number of available appointment slots, and perform supply and demand status marking to obtain candidate resource allocation schemes; Based on the estimated reservation demand, the demand fulfillment, appointment utilization, and resource adjustment of each candidate resource allocation scheme are calculated. Evaluation values ​​are calculated and filtered through scheme evaluation parameters to obtain the target resource allocation scheme and expected usage results. Update the appointment slot configuration and process appointment requests according to the target resource allocation plan, collect actual appointment results, compare the actual appointment results with the estimated appointment demand results and expected usage results, and obtain demand feedback results and evaluation feedback results. Based on the demand feedback and evaluation feedback results, the improved SPSA algorithm is used to perturb the demand estimation parameters and scheme evaluation parameters to obtain updated demand estimation parameters and updated scheme evaluation parameters.

[0008] Demand estimation is performed using updated demand estimation parameters. After supply and demand status marking and candidate resource allocation schemes are generated, scheme evaluation is performed using updated scheme evaluation parameters to obtain the target resource allocation scheme for the next allocation cycle.

[0009] Optionally, the step of reading the schedule, appointment resource configuration, historical appointment data, demand estimation parameters, and scheme evaluation parameters, and establishing appointment resource items by department, doctor, and time period, specifically involves: Extract the department, doctor, start time and end time of the time period from the scheduling and appointment allocation. Then, collect the start time and end time of the time period by department and doctor and sort them to obtain the time period boundary sequence. The scheduling and appointment allocation are divided according to adjacent times in the time period boundary sequence. The segmented intervals that are simultaneously located in the scheduling time period and the appointment allocation time period are retained to obtain the effective time period sequence. The number of number sources is divided according to the proportion of the effective time period to the number source allocation time period in the effective time period sequence. The result of the division is rounded and the number of unallocated number sources is added in order of the effective time period from longest to shortest to obtain the segmented number source allocation. Filter historical appointment data with the same department and doctor, and write the appointment time in the historical appointment data into the corresponding valid time period to obtain segmented historical appointment data; The resource item record is obtained by matching demand estimation parameters and scheme evaluation parameters with department, doctor and effective time period, and writing them together with the segmented appointment source configuration and segmented historical appointment data into the effective time period sequence. Based on the resource item records, registration resource items are created and arranged in order of department, doctor, and time period to obtain resource status data.

[0010] Optionally, the step of collecting the current number of reservation requests and calculating the reservation demand estimation results for each registration resource item based on resource status data and demand estimation parameters specifically involves: Extract the effective time period and segmented historical reservation data of each registered resource item from the resource status data, count the number of historical reservations and calculate the duration of the effective time period to obtain the historical reservation quantity sequence and the time period length sequence. Collect current appointment requests and write them into the corresponding registration resource items according to department, doctor and appointment time. Count the current appointment request volume of each registration resource item to obtain the current request volume sequence. Divide the quantities in the historical reservation volume sequence and the current request volume sequence by the corresponding duration in the time period length sequence to obtain the historical reservation intensity sequence and the current request intensity sequence, respectively. Compare the current request intensity sequence with the historical reservation intensity sequence, determine the direction of demand change according to the sign of the difference, and determine the amount of demand change according to the absolute value of the difference, thus obtaining the demand change sequence; Based on the demand estimation parameters, the historical reservation intensity sequence, the current request intensity sequence, and the demand change sequence are weighted and synthesized, and the weighted result corresponding to the demand change is corrected in the same direction according to the direction of demand change to obtain the demand intensity estimation sequence. The estimated demand intensity sequence is multiplied by the corresponding duration in the time period length sequence, and then written into the resource status data in the order of the registered resource items to obtain the reservation demand estimation result.

[0011] Optionally, the steps of reading the patient reception capacity and adjusting the limits, calculating the number of available appointment slots, comparing the estimated appointment demand with the number of available appointment slots, and performing supply and demand status marking are as follows: Extract the segmented appointment slot configuration and appointment demand estimation results from the resource status data, collect the patient reception capacity and adjustment limits, and arrange them according to the order of registration resource items to obtain the resource constraint sequence; Extract the patient capacity from the resource constraint sequence item by item, compare the number of appointment slots in the segmented appointment slot configuration with the corresponding patient capacity, determine the number of available appointment slots, and obtain the sequence of available appointment slots. Subtract the corresponding number in the available appointment slot sequence from the estimated appointment demand, and determine the number of appointment slots to be transferred in and the number of appointment slots to be transferred out. Select the registration resource items with a quantity of available appointment slots and a quantity of available appointment slots that are greater than 0, and sort them in descending order of the quantity of available appointment slots and the quantity of available appointment slots, respectively, to obtain the sequence of available appointment resource items and the sequence of available appointment resource items. According to the adjustment restrictions in the resource constraint sequence, the sequence of resource items to be transferred in and the sequence of resource items to be transferred out are matched, and the number of source numbers for this adjustment is determined; Based on the adjusted number of available appointment slots, the corresponding number of available appointment slots to be transferred in and the number of available appointment slots to be transferred out are deducted respectively. For registration resource items with values ​​greater than 0, pairing continues until there are no more pairings that meet the adjustment restrictions, and the first candidate resource allocation scheme is obtained. The sequence of resources to be transferred in and the sequence of resources to be transferred out are rearranged according to the start time of the effective time period to obtain the second candidate resource allocation scheme, which is then merged with the first candidate resource allocation scheme to obtain the candidate resource allocation scheme.

[0012] Optionally, obtaining the target resource allocation scheme and expected usage results specifically involves: According to the order of department, doctor and effective time period, extract the available number of appointment slots, the estimated appointment demand, and the adjusted number of appointment slots in each candidate resource allocation scheme to obtain the scheme time sequence record; By recording the estimated reservation demand and the number of available appointment slots item by item along the time sequence of the plan, the cumulative demand estimation sequence and the cumulative available slot sequence are obtained. The initial demand gap is determined based on the cumulative demand estimation sequence and the cumulative allocable sequence, thus obtaining the initial demand gap sequence. For each candidate resource allocation scheme, the number of available slots is accumulated and adjusted along the time sequence of the scheme, and compared with the cumulative demand estimation sequence to obtain the candidate demand gap sequence and the candidate remaining slot sequence; Calculate the demand fulfillment and utilization of license plates based on the initial demand gap sequence, the candidate demand gap sequence, and the adjusted license plate availability. Summarize the number of available slots adjusted in each candidate resource allocation scheme to obtain the resource adjustment amount, and write the demand fulfillment amount, slot utilization amount, and resource adjustment amount into the corresponding candidate resource allocation scheme, and determine the allocation scheme to be evaluated; The evaluation values ​​of each allocation scheme to be evaluated are calculated by using scheme evaluation parameters and then filtered to obtain the target resource allocation scheme. The candidate demand gap sequence and candidate remaining number source sequence corresponding to the target resource allocation scheme are then determined as the expected usage results.

[0013] Optionally, obtaining the demand feedback results and evaluation feedback results specifically includes: According to the target resource allocation plan, the adjusted number of available appointment slots is written into the appointment slot configuration of the corresponding registration resource item. The appointment requests are processed and the actual number of requests, actual number of appointments, and actual number of remaining appointment slots are counted to obtain the actual appointment results. The actual appointment results are grouped according to department and doctor, and then arranged according to the start time of the valid time period within each group to obtain the actual usage time sequence record; By sequentially adding up the actual request volume, actual reservation volume, and actual remaining number of slots along the actual usage timeline, we can obtain the cumulative actual request sequence, the cumulative actual reservation sequence, and the cumulative actual remaining sequence. The cumulative actual request sequence corresponding to the actual reservation results is compared with the cumulative demand estimation sequence corresponding to the reservation demand estimation results to obtain the demand feedback results; The actual demand gap is determined based on the cumulative actual request sequence and the cumulative actual reservation sequence, thus obtaining the actual demand gap sequence. Candidate demand gap sequences and candidate remaining number source sequences are extracted from the expected usage results, and compared with the actual demand gap sequence and the cumulative actual remaining sequence respectively to obtain the gap deviation sequence and the remaining number source deviation sequence; The evaluation feedback results are obtained by combining the gap deviation sequence and the remaining number source deviation sequence according to the order of the registration resource items.

[0014] Optionally, the improved SPSA algorithm is specifically as follows: Extract the demand deviation direction and demand deviation amount from the demand feedback results, and extract the gap deviation sequence and remaining number source deviation sequence from the evaluation feedback results to obtain the feedback sequence; The absolute value of the deviation is calculated based on the feedback sequence and then standardized to obtain the standardized feedback sequence. The CRITIC weighting method is used to calculate the feedback weights based on the dispersion of each deviation and the degree of inter-item correlation in the standardized feedback sequence; Set up parameter indexes according to the sequence numbers of demand estimation parameters and scheme evaluation parameters, record the number of parameter corrections, and establish an initial disturbance symbol sequence; Summarize the signed values ​​of demand deviation, gap deviation sequence, and remaining source deviation sequence, and determine the feedback sign; Set a base perturbation amplitude and a lower limit for the perturbation amplitude that are greater than 0, and calculate the perturbation amplitude sequence based on the feedback weights; The corrected perturbation symbol sequence is obtained by multiplying each initial perturbation symbol in the initial perturbation symbol sequence with its corresponding feedback symbol. According to the corrected disturbance symbol sequence and disturbance amplitude sequence, positive and negative disturbances are applied to the demand estimation parameters and the scheme evaluation parameters respectively to obtain the first disturbance parameter set and the second disturbance parameter set; The first and second disturbance parameter sets are used to back-calculate the current allocation period, and the obtained reservation demand estimation results and expected usage results are compared with the actual reservation results and summarized according to the feedback weight to obtain the first disturbance feedback value and the second disturbance feedback value. The difference between the first disturbance feedback value and the second disturbance feedback value is divided by the difference between the corresponding parameters in the first disturbance parameter group and the second disturbance parameter group to obtain the parameter correction amount; According to the parameter index, subtract the corresponding parameter correction amount from each parameter in the demand estimation parameters and the scheme evaluation parameters, and perform normalization respectively to obtain the updated demand estimation parameters and the updated scheme evaluation parameters.

[0015] Optionally, obtaining the target resource allocation scheme for the next allocation cycle specifically involves: The resource status data is updated based on the actual reservation results, the target resource allocation plan, the updated demand estimation parameters, and the updated plan evaluation parameters to obtain the resource status data for the next cycle. Collect the current reservation request volume for the next allocation cycle, and perform demand estimation based on the resource status data for the next cycle to obtain the reservation demand estimation result for the next cycle; Based on the estimated appointment demand for the next cycle, the segmented appointment allocation, the patient capacity, and the adjustment restrictions, the supply and demand status is marked to obtain the expected resource item sequence for the next week and the resource item sequence for the next cycle. Pair the sequence of resource items expected to be added in the next week with the sequence of resource items that can be added out in the next cycle to obtain the candidate resource allocation scheme for the next cycle. Based on the estimated reservation demand for the next cycle, calculate the demand satisfaction, appointment utilization, and resource adjustment for each candidate resource allocation scheme for the next cycle to obtain the expected evaluation allocation scheme for the next week. By updating the evaluation parameters of the scheme, calculating the evaluation value of each expected evaluation allocation scheme for the next week, and filtering them, the target resource allocation scheme for the next allocation cycle is obtained.

[0016] A dynamic allocation system for hospital appointment registration resources includes: The resource status construction module is used to establish registration resource items and generate resource status data based on the schedule, appointment configuration, historical reservation data, demand estimation parameters and scheme evaluation parameters. The demand estimation module is used to generate reservation demand estimation results based on the current reservation request volume, resource status data, and demand estimation parameters. The resource allocation module is used to perform supply and demand status marking based on the patient capacity, adjustment restrictions, and appointment demand estimation results, and generate candidate resource allocation plans. The scheme evaluation and screening module is used to calculate the demand satisfaction, number of slots utilized, and resource adjustment of each candidate resource allocation scheme, and to generate the target resource allocation scheme and expected usage results through scheme evaluation parameters. The appointment execution feedback module is used to update the appointment resource configuration according to the target resource allocation plan and process appointment requests, and generate demand feedback results and evaluation feedback results based on the actual appointment results; The parameter correction module is used to perturb the demand estimation parameters and scheme evaluation parameters using the improved SPSA algorithm, calculate the feedback weights, and generate updated demand estimation parameters and updated scheme evaluation parameters. The periodic allocation module is used to update resource status data based on actual reservation results, updated demand estimation parameters, and updated scheme evaluation parameters, and to generate the target resource allocation scheme for the next allocation cycle.

[0017] A computer storage medium storing a computer program, which, when executed by a processor, implements a method for dynamically allocating hospital appointment registration resources.

[0018] The beneficial effects of this invention are: (1) This invention establishes registration resource items according to department, doctor and effective time period, performs supply and demand status marking in combination with the reception capacity, adjustment restrictions and appointment demand estimation results, and generates candidate resource allocation schemes based on the number of appointments to be transferred in and the number of appointments that can be transferred out, thereby realizing the joint constraint between registration resource allocation and actual scheduling time period, reception capacity and appointment demand, thereby reducing the adjustment of appointments that do not conform to business constraints across departments, across doctors and different time periods.

[0019] (2) This invention collects the actual number of requests, the actual number of reservations and the actual number of remaining appointment slots, compares the actual reservation results with the reservation demand estimation results and the expected usage results, generates demand feedback results and evaluation feedback results, and feeds back the updated demand estimation parameters and updated scheme evaluation parameters to the next allocation cycle, thereby realizing dual feedback correction of the demand estimation process and the scheme evaluation process, so that the registration resource allocation scheme is continuously updated according to the actual supply and demand status.

[0020] (3) The improved SPSA algorithm in this invention generates a perturbation symbol sequence by parameter index and parameter correction number, uses feedback symbols to correct the perturbation direction, and uses CRITIC weighting method to calculate feedback weight according to the dispersion of feedback deviation and the degree of inter-item correlation, and adjusts the perturbation amplitude according to the feedback weight; at the same time, a lower limit of perturbation amplitude is set and the parameter correction amount is calculated by using positive perturbation and reverse perturbation, thereby reducing parameter update fluctuations caused by random perturbation, avoiding the corresponding perturbation parameter difference being 0, and maintaining parameter update efficiency under two-sided perturbation. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for dynamically allocating hospital appointment registration resources proposed in this invention; Figure 2This is a system structure diagram of a dynamic allocation system for hospital appointment registration resources proposed in this invention; Figure 3 The flowchart of the improved SPSA algorithm for a dynamic allocation method of hospital appointment registration resources proposed in this invention is shown below. Figure 4 Chart showing the supply and demand status of registration resources and the allocation and flow of appointment slots; Figure 5 A two-layer deviation matrix diagram between expected usage results and actual reservation results. Detailed Implementation

[0022] Example 1: Reference Figures 1-5 A method for dynamically allocating hospital appointment registration resources, comprising: Read the schedule, appointment allocation, historical appointment data, demand estimation parameters and scheme evaluation parameters, and establish registration resource items by department, doctor and time period to obtain resource status data; Collect the current number of reservation requests, and calculate the reservation demand estimate for each registration resource item based on resource status data and demand estimation parameters; Read the patient reception capacity and adjustment limits, calculate the number of available appointment slots, compare the estimated appointment demand with the number of available appointment slots, and perform supply and demand status marking to obtain candidate resource allocation schemes; Based on the estimated reservation demand, the demand fulfillment, appointment utilization, and resource adjustment of each candidate resource allocation scheme are calculated. Evaluation values ​​are calculated and filtered through scheme evaluation parameters to obtain the target resource allocation scheme and expected usage results. Update the appointment slot configuration and process appointment requests according to the target resource allocation plan, collect actual appointment results, compare the actual appointment results with the estimated appointment demand results and expected usage results, and obtain demand feedback results and evaluation feedback results. Based on the demand feedback and evaluation feedback results, the improved SPSA algorithm is used to perturb the demand estimation parameters and the scheme evaluation parameters. Specifically, the CRITIC weighting method is used to calculate the feedback weights, and the perturbation direction of the improved SPSA algorithm is corrected according to the feedback sign. The perturbation amplitude of the improved SPSA algorithm is adjusted according to the feedback weights to obtain the updated demand estimation parameters and the updated scheme evaluation parameters.

[0023] Demand estimation is performed using updated demand estimation parameters. After supply and demand status marking and candidate resource allocation schemes are generated, scheme evaluation is performed using updated scheme evaluation parameters to obtain the target resource allocation scheme for the next allocation cycle.

[0024] In this embodiment, the step of reading the schedule, appointment allocation, historical appointment data, demand estimation parameters, and scheme evaluation parameters, and establishing appointment resource items by department, doctor, and time period, specifically involves: Extract the department, doctor, start time and end time of the time period from the scheduling and appointment allocation. Then, collect the start time and end time of the time period by department and doctor and sort them to obtain the time period boundary sequence. The scheduling and appointment allocation are divided according to adjacent times in the time period boundary sequence. The segmented intervals that are simultaneously located in the scheduling time period and the appointment allocation time period are retained to obtain the effective time period sequence. In Example 1, the start and end times of the scheduled shifts, the start and end times of the appointment allocation periods, and the end times of the appointment allocation periods for the same department and the same doctor are summarized and arranged in chronological order. Two adjacent times form a segmentation interval. Each segmentation interval is checked to see if the start and end times fall within the corresponding scheduled shift and appointment allocation periods. If both conditions are met, the corresponding segmentation interval is retained; if neither condition is met, the corresponding segmentation interval is deleted. The retained segmentation intervals are arranged in chronological order, and the corresponding department, doctor, start and end times are recorded to form a valid time period sequence. The number of number source configurations is divided according to the proportion of the effective time period to the number source configuration time period in the effective time period sequence. The result of the division is rounded down, and the number of unallocated number sources is added in order of the effective time period from the longest to the shortest to obtain the segmented number source configuration. In Example 1, for the same number source configuration period, the duration of each valid period and the duration of the number source configuration period are calculated. The ratio obtained by dividing the duration of each valid period by the duration of the number source configuration period is multiplied by the number of number sources to obtain the initial number of segments. Each initial number of segments is rounded down, all rounding results are summarized, and the summation result is subtracted from the number of number sources to obtain the number of unallocated number sources. The valid periods are arranged from longest to shortest duration, and if the durations are the same, they are arranged from earliest to latest start time. Starting from the first position, one unallocated number source is added one by one. After reaching the last position, the process returns to the first position and continues until the number of unallocated number sources is 0. The number of number sources for each valid period after the addition is written into the corresponding valid period to obtain the segmented number source configuration. Filter historical appointment data with the same department and doctor, and write the appointment time in the historical appointment data into the corresponding valid time period to obtain segmented historical appointment data; The resource item record is obtained by matching demand estimation parameters and scheme evaluation parameters with department, doctor and effective time period, and writing them together with the segmented appointment source configuration and segmented historical appointment data into the effective time period sequence. In Example 1, the demand estimation parameters are set as the weights of historical appointment volume and current appointment request volume, and the scheme evaluation parameters are set as the weights of demand fulfillment volume, appointment utilization volume, and resource adjustment volume. The sum of the two demand estimation parameters and the sum of the three scheme evaluation parameters are both 1. When there is a previous allocation cycle, the demand estimation parameters and scheme evaluation parameters are updated. When the system is cold-started and historical appointment data exists, the two demand estimation parameters are both set to 0.5, and the three scheme evaluation parameters are both set to the value obtained by dividing 1 by 3. When historical appointment data is missing, the weight of historical appointment volume is set to 0, and the weight of current appointment request volume is set to 1. The selected parameters are used for back-calculation of historical appointment data. When the difference between the appointment demand estimation result and the historical actual appointment volume does not exceed the set deviation threshold, and the candidate resource allocation scheme meets the reception capacity and adjustment restrictions, the parameters are deemed feasible. If not, the two demand estimation parameters are adjusted according to the direction of the difference, and the weights of the evaluation indicators that cause the exceedance are reduced until the judgment conditions are met. Based on the resource item records, registration resource items are created and arranged in order of department, doctor, and time period to obtain resource status data.

[0025] In Example 1, the department, doctor, valid time slot, segmented appointment allocation, segmented historical appointment data, demand estimation parameters, and scheme evaluation parameters are extracted from each resource item record. The department, doctor, and valid time slot are used as the identification content of the registration resource item, and the segmented appointment allocation, segmented historical appointment data, demand estimation parameters, and scheme evaluation parameters are written into the registration resource item. When the department, doctor, and valid time slot are exactly the same, the resource item records are merged, and the merged single registration resource item is retained. All registration resource items are arranged in the order of department, doctor, and start time of valid time slot to form resource status data.

[0026] In this embodiment, the step of collecting the current reservation request volume and calculating the reservation demand estimation results for each registration resource item based on resource status data and demand estimation parameters specifically involves: Extract the effective time period and segmented historical reservation data of each registered resource item from the resource status data, count the number of historical reservations and calculate the duration of the effective time period to obtain the historical reservation quantity sequence and the time period length sequence. Collect current appointment requests and write them into the corresponding registration resource items according to department, doctor and appointment time. Count the current appointment request volume of each registration resource item to obtain the current request volume sequence. Divide the quantities in the historical reservation volume sequence and the current request volume sequence by the corresponding duration in the time period length sequence to obtain the historical reservation intensity sequence and the current request intensity sequence, respectively. Compare the current request intensity sequence with the historical reservation intensity sequence, determine the direction of demand change according to the sign of the difference, and determine the amount of demand change according to the absolute value of the difference, thus obtaining the demand change sequence; In Example 1, the corresponding values ​​in the current request intensity sequence and the historical reservation intensity sequence are extracted item by item according to the order of the registered resource items. The current request intensity is subtracted from the historical reservation intensity to obtain the intensity difference. When the intensity difference is positive, the direction of demand change is recorded as increasing; when the intensity difference is negative, the direction of demand change is recorded as decreasing; when the intensity difference is 0, the direction of demand change is recorded as unchanged. The intensity difference is converted into a non-negative value to obtain the demand change amount. The demand change direction and demand change amount are combined according to the order of the registered resource items to form the demand change sequence. Based on the demand estimation parameters, the historical reservation intensity sequence, the current request intensity sequence, and the demand change sequence are weighted and synthesized, and the weighted result corresponding to the demand change is corrected in the same direction according to the direction of demand change to obtain the demand intensity estimation sequence. In Example 1, the historical reservation intensity, current request intensity, demand change direction, demand change amount, historical reservation volume weight, and current reservation request volume weight are extracted item by item according to the order of the registered resource items. The historical reservation intensity is multiplied by the historical reservation volume weight, and the current request intensity is multiplied by the current reservation request volume weight. The two products are added together to obtain the basic demand intensity. The demand change amount is multiplied by the current reservation request volume weight to obtain the demand adjustment amount. The demand change amount is not assigned a separate weight. When the demand change direction is increasing, the demand adjustment amount is added to the basic demand intensity. When the demand change direction is decreasing, the demand adjustment amount is subtracted from the basic demand intensity. When the demand change direction is unchanged, the basic demand intensity is retained. The adjusted values ​​are arranged according to the order of the registered resource items to form a demand intensity estimation sequence. The estimated demand intensity sequence is multiplied by the corresponding duration in the time period length sequence, and then written into the resource status data in the order of the registered resource items to obtain the reservation demand estimation result.

[0027] In this embodiment, the steps of reading the reception capacity and adjustment limits, calculating the number of available appointment slots, comparing the estimated appointment demand with the number of available appointment slots, and performing supply and demand status marking are as follows: Extract the segmented appointment slot configuration and appointment demand estimation results from the resource status data, collect the patient reception capacity and adjustment limits, and arrange them according to the order of registration resource items to obtain the resource constraint sequence; In Example 1, the adjustment restrictions are set before the current allocation cycle is executed. For each registration resource item, the number of appointment slots in the segmented appointment slot configuration is subtracted from the number of historical appointments in the segmented historical appointment data. If the difference is greater than 0, the difference is set as the upper limit of the allowed number of appointment slots to be transferred out; if the difference is not greater than 0, the upper limit of the allowed number of appointment slots to be transferred out is set to 0. The reception capacity is subtracted from the number of appointment slots in the segmented appointment slot configuration. If the difference is greater than 0, the difference is set as the upper limit of the allowed number of appointment slots to be transferred in; if the difference is not greater than 0, the upper limit of the allowed number of appointment slots to be transferred in is set to 0. Registration resource items within the same department are set as allowed pairing ranges, and the upper limit of the allowed number of appointment slots to be transferred out, the upper limit of the allowed number of appointment slots to be transferred in, and the allowed pairing range are written into the adjustment restrictions in the order of registration resource items. When matching the resource items to be transferred in with the resource items that can be transferred out, the smaller value between the corresponding upper limit of the allowed number of appointment slots to be transferred in and the upper limit of the allowed number of appointment slots to be transferred out is determined as the allowed number of appointment slots to be adjusted. Extract the patient capacity from the resource constraint sequence item by item, compare the number of appointment slots in the segmented appointment slot configuration with the corresponding patient capacity, and determine the item with the smaller value as the allocable appointment slot quantity to obtain the allocable appointment slot quantity sequence. Subtract the corresponding quantity in the available appointment slot sequence from the estimated appointment demand, and determine the number of appointment slots to be transferred in and the number of appointment slots to be transferred out. When the difference is positive, mark the shortage state and record the difference as the number of appointment slots to be transferred in. When the difference is negative, mark the surplus state and record the absolute value of the difference as the number of appointment slots to be transferred out. When the difference is 0, mark the balance state. Select the registration resource items with a quantity of available appointment slots and a quantity of available appointment slots that are greater than 0, and sort them in descending order of the quantity of available appointment slots and the quantity of available appointment slots, respectively, to obtain the sequence of available appointment resource items and the sequence of available appointment resource items. According to the adjustment restrictions in the resource constraint sequence, the sequence of resource items to be transferred in and the sequence of resource items to be transferred out are matched, and the minimum value among the number of resource items to be transferred in, the number of resource items to be transferred out, and the number of resource items allowed to be adjusted according to the adjustment restrictions is determined as the number of resource items to be adjusted this time. Based on the adjusted number of available appointment slots, the corresponding number of available appointment slots to be transferred in and the number of available appointment slots to be transferred out are deducted respectively. For registration resource items with values ​​greater than 0, pairing continues until there are no more pairings that meet the adjustment restrictions, and the first candidate resource allocation scheme is obtained. The sequence of resources to be transferred in and the sequence of resources to be transferred out are rearranged according to the start time of the effective time period. Repeated pairing and number source deduction are performed to obtain the second candidate resource allocation scheme, which is then merged with the first candidate resource allocation scheme to obtain the candidate resource allocation scheme.

[0028] In this embodiment, obtaining the target resource allocation scheme and the expected usage results specifically refers to: According to the order of department, doctor and effective time period, extract the available number of appointment slots, the estimated appointment demand, and the adjusted number of appointment slots in each candidate resource allocation scheme to obtain the scheme time sequence record; In Example 1, the time-series records of the plan are grouped according to department and doctor. Records corresponding to the same department and the same doctor form a time-series record. Within each time-series record, the records are arranged from earliest to latest according to the start time of the effective time period. Starting from the first effective time period in the group, the estimated appointment demand, the number of available appointment slots, and the adjusted number of appointment slots are accumulated item by item. When the department or doctor corresponding to the current record is different from the previous record, the accumulated value is reset to 0, and the accumulation starts again from the current record. Each group of time-series records forms a corresponding cumulative demand estimation sequence, cumulative available slot sequence, candidate demand gap sequence, and candidate remaining appointment slot sequence. By recording the estimated reservation demand and the number of available appointment slots item by item along the time sequence of the plan, the cumulative demand estimation sequence and the cumulative available slot sequence are obtained. The initial demand gap is determined based on the cumulative demand estimation sequence and the cumulative allocable sequence, resulting in the initial demand gap sequence. The value in the cumulative demand estimation sequence is subtracted from the corresponding value in the cumulative allocable sequence. If the difference is greater than 0, the difference is recorded as the initial demand gap; if the difference is not greater than 0, the initial demand gap is recorded as 0. For each candidate resource allocation scheme, the number of available slots is accumulated and adjusted along the time sequence of the scheme, and compared with the cumulative demand estimation sequence to obtain the candidate demand gap sequence and the candidate remaining slot sequence; The demand fulfillment amount and the number of number slots are calculated based on the initial demand gap sequence, the candidate demand gap sequence, and the adjusted number of number slots. Specifically, the demand fulfillment amount is obtained by subtracting the corresponding values ​​in the initial demand gap sequence from the corresponding values ​​in the candidate demand gap sequence and summing them up. The number of number slots is obtained by subtracting the summed value of the remaining candidate number slot sequence from the summed value of the adjusted number of number slots. Summarize the number of available slots adjusted in each candidate resource allocation scheme to obtain the resource adjustment amount. Write the demand fulfillment amount, slot utilization amount, and resource adjustment amount into the corresponding candidate resource allocation scheme and determine the allocation scheme to be evaluated. Among them, delete the candidate resource allocation schemes whose demand fulfillment amount and slot utilization amount are not higher than another candidate resource allocation scheme and whose resource adjustment amount is not lower than another candidate resource allocation scheme. The allocation scheme to be evaluated is obtained. The evaluation values ​​of each allocation scheme to be evaluated are calculated by using scheme evaluation parameters and then filtered to obtain the target resource allocation scheme. The candidate demand gap sequence and candidate remaining number source sequence corresponding to the target resource allocation scheme are then determined as the expected usage results.

[0029] In Example 1, the demand satisfaction amount, number of available slots, and resource adjustment amount are extracted according to the order of the allocation schemes to be evaluated. The weights of demand satisfaction amount, number of available slots, and resource adjustment amount in the scheme evaluation parameters are also extracted. The products of demand satisfaction amount and demand satisfaction amount weight, number of available slots and number of available slots weight, and resource adjustment amount and resource adjustment amount weight are calculated respectively. The first two products are added together and the third product is subtracted to obtain the evaluation value of each allocation scheme to be evaluated. The allocation schemes to be evaluated are arranged in descending order of evaluation value. If the evaluation values ​​are the same, the allocation scheme with the smaller resource adjustment amount is selected. If the resource adjustment amounts are still the same, the allocation scheme with the larger number of available slots is selected to obtain the target resource allocation scheme. The candidate demand gap sequence and candidate remaining number of available slots corresponding to the target resource allocation scheme are written into the expected usage results.

[0030] In this embodiment, obtaining the demand feedback result and the evaluation feedback result specifically refers to: According to the target resource allocation plan, the adjusted number of available appointment slots is written into the appointment slot configuration of the corresponding registration resource item. The appointment requests are processed and the actual number of requests, actual number of appointments, and actual number of remaining appointment slots are counted to obtain the actual appointment result. The actual number of remaining appointment slots is obtained by subtracting the actual number of appointments from the adjusted number of available appointment slots. In Example 1, an appointment request is received and the department, doctor, and appointment time are extracted. The appointment request is written into the corresponding registration resource item according to the department, doctor, and appointment time. The appointment requests are processed one by one according to the arrival order. When the number of appointments already booked in a registration resource item is less than the adjusted number of available appointments, the appointment is recorded as successful and the number of appointments is increased by 1. When the number of appointments already booked reaches the adjusted number of available appointments, the appointment is recorded as unsuccessful. The number of appointment requests received by each registration resource item is counted as the actual number of requests. The number of successful appointments is counted as the actual number of appointments. The adjusted number of available appointments is subtracted from the actual number of appointments to obtain the actual number of available appointments. The actual appointment results are grouped according to department and doctor, and then arranged according to the start time of the valid time period within each group to obtain the actual usage time sequence record; In Example 1, the department, doctor, valid time slot, actual number of requests, actual number of appointments, and actual number of remaining appointment slots are extracted from each actual appointment result. Only when the department name and doctor name are the same, the corresponding actual appointment results are grouped into the same group; otherwise, they are grouped into different groups. Within the same group, the actual number of requests, actual number of appointments, and actual number of remaining appointment slots are retained according to the valid time slot, forming actual appointment result groups divided by department and doctor. The actual number of requests, actual number of appointments, and actual number of remaining appointment slots are accumulated item by item along the actual usage time sequence to obtain the cumulative actual request sequence, the cumulative actual appointment sequence, and the cumulative actual remaining sequence. The cumulative value is reset when the department or doctor changes. The cumulative actual request sequence corresponding to the actual reservation result is compared with the cumulative demand estimation sequence corresponding to the reservation demand estimation result to obtain the demand feedback result. The demand feedback result records the demand deviation direction according to the positive or negative value of the corresponding numerical difference, and records the demand deviation amount according to the absolute value of the difference. The actual demand gap is determined based on the cumulative actual request sequence and the cumulative actual reservation sequence, and the actual demand gap sequence is obtained. The actual demand gap is the positive difference obtained by subtracting the cumulative actual reservation from the cumulative actual request, and is recorded as 0 when the difference is not greater than 0. Candidate demand gap sequences and candidate remaining number source sequences are extracted from the expected usage results, and compared with the actual demand gap sequence and the cumulative actual remaining sequence respectively to obtain the gap deviation sequence and the remaining number source deviation sequence; In Example 1, the gap deviation sequence is the sequence obtained by subtracting the corresponding value in the candidate demand gap sequence from the value in the actual demand gap sequence, and the remaining number source deviation sequence is the sequence obtained by subtracting the corresponding value in the candidate remaining number source sequence from the value in the cumulative actual remaining sequence. The evaluation feedback results are obtained by combining the gap deviation sequence and the remaining number source deviation sequence according to the order of the registration resource items.

[0031] In this embodiment, the improved SPSA algorithm is specifically as follows: Extract the demand deviation direction and demand deviation amount from the demand feedback results, and extract the gap deviation sequence and remaining number source deviation sequence from the evaluation feedback results to obtain the feedback sequence; The absolute value of the deviation is calculated and standardized based on the demand deviation, gap deviation sequence and remaining number source deviation sequence in the feedback sequence to obtain the standardized feedback sequence. In Example 1, the demand deviation, gap deviation sequence, and remaining number source deviation sequence values ​​are extracted item by item according to the order of the registered resource items. Each value is converted into a non-negative number to obtain three types of absolute deviation value sequences. The maximum and minimum values ​​in each type of absolute deviation value sequence are determined. For each absolute deviation value, the minimum value of the same type is subtracted from the corresponding value, and the quotient is calculated by using the difference between the maximum and minimum values ​​of the same type as the divisor to obtain a standardized value. When the maximum and minimum values ​​of the same type are the same, the corresponding standardized value is recorded as 0. The three types of standardized values ​​are combined according to the order of the registered resource items to obtain a standardized feedback sequence. The CRITIC weighting method is used to calculate the feedback weights based on the dispersion of each deviation and the degree of inter-item correlation in the standardized feedback sequence; In Example 1, a standardized feedback sequence is divided according to demand deviation, gap deviation, and remaining registration source deviation, and the standardized values ​​of each type of deviation in all registration resource items are extracted. The average value of the standardized value of each type of deviation is calculated, the square of the difference between each standardized value and the average value is summed, the quotient is calculated by dividing by the number of registration resource items, and the square root is taken to obtain the dispersion. The correlation coefficients of the standardized values ​​of the three types of deviations are calculated pairwise, and the values ​​obtained by subtracting the correlation coefficients from 1 are summed to obtain the degree of conflict between items. The dispersion of each type of deviation is multiplied by the degree of conflict between items to obtain the amount of feedback information. The quotient of each amount of feedback information is calculated by dividing by the sum of all feedback information amounts to obtain the feedback weight. Set up parameter indexes according to the sequence numbers of demand estimation parameters and scheme evaluation parameters, record the number of parameter corrections, and establish an initial disturbance symbol sequence with a value of 1 or -1 based on the parity relationship between parameter indexes and the number of parameter corrections. Summarize the signed values ​​of the demand deviation, the gap deviation sequence, and the remaining source deviation sequence, and determine the feedback sign. The feedback sign when the summed value is not less than 0 is recorded as 1, and the feedback sign when the summed value is less than 0 is recorded as -1. Set a base disturbance amplitude and a lower limit of disturbance amplitude that are greater than 0, and calculate the disturbance amplitude sequence based on the feedback weight. The calculation is to multiply the larger value between the feedback weight and the lower limit of disturbance amplitude by the base disturbance amplitude. The corrected perturbation symbol sequence is obtained by multiplying each initial perturbation symbol in the initial perturbation symbol sequence with its corresponding feedback symbol. According to the corrected disturbance symbol sequence and disturbance amplitude sequence, positive and negative disturbances are applied to the demand estimation parameters and the scheme evaluation parameters respectively to obtain the first disturbance parameter group and the second disturbance parameter group. The difference between the corresponding parameters in the two groups is twice the corresponding disturbance amplitude. In Example 1, the demand estimation parameters, scheme evaluation parameters, corrected disturbance sign sequence, and disturbance amplitude sequence are matched item by item according to the parameter index. When the corrected disturbance sign is 1, the corresponding disturbance amplitude is added to the corresponding parameter to form the first disturbance parameter, and the corresponding disturbance amplitude is subtracted from the corresponding parameter to form the second disturbance parameter. When the corrected disturbance sign is -1, the corresponding disturbance amplitude is subtracted from the corresponding parameter to form the first disturbance parameter, and the corresponding disturbance amplitude is added to the corresponding parameter to form the second disturbance parameter. All the first and second disturbance parameters are arranged according to the parameter index to obtain the first disturbance parameter group and the second disturbance parameter group, respectively. The first and second disturbance parameter sets are used to back-calculate the current allocation period, and the obtained reservation demand estimation results and expected usage results are compared with the actual reservation results and summarized according to the feedback weight to obtain the first disturbance feedback value and the second disturbance feedback value. In Example 1, keeping the resource status data, reception capacity, and adjustment limits of the current allocation cycle unchanged, the demand estimation parameters and scheme evaluation parameters in the first and second disturbance parameter groups are substituted into the corresponding processing steps to recalculate the reservation demand estimation results, target resource allocation scheme, and expected usage results. The reservation demand estimation results obtained from each backcalculation are compared with the cumulative actual request sequence in the actual reservation results to obtain the absolute value of the demand deviation. The candidate demand gap sequence obtained from each backcalculation is compared with the actual demand gap sequence, and the candidate remaining appointment sequence is compared with the cumulative actual remaining sequence to obtain the absolute value of the gap deviation and the absolute value of the remaining appointment deviation. The three types of absolute deviation values ​​are multiplied by the corresponding feedback weights and summed to obtain the first disturbance feedback value and the second disturbance feedback value. The difference between the first disturbance feedback value and the second disturbance feedback value is divided by the difference between the corresponding parameters in the first disturbance parameter group and the second disturbance parameter group to obtain the parameter correction amount; According to the parameter index, subtract the corresponding parameter correction amount from each parameter in the demand estimation parameters and the scheme evaluation parameters, and perform normalization respectively to obtain the updated demand estimation parameters and the updated scheme evaluation parameters.

[0032] In Example 1, the improved SPSA algorithm takes demand feedback results, evaluation feedback results, demand estimation parameters, and scheme evaluation parameters as inputs, and updates demand estimation parameters and scheme evaluation parameters as outputs. The demand feedback results are derived from a comparison between the cumulative actual request sequence and the cumulative demand estimation sequence. The evaluation feedback results are derived from a comparison between the actual demand gap sequence and the candidate demand gap sequence, and a comparison between the cumulative actual remaining sequence and the candidate remaining number source sequence. The cumulative actual request sequence, actual demand gap sequence, and cumulative actual remaining sequence are all generated from the actual reservation results of the current allocation period. When there are updated demand estimation parameters and updated scheme evaluation parameters from the previous allocation period... The update result of the previous allocation cycle is used as the initial value for this iteration. If there is no update result from the previous allocation cycle, parameters of the same type are set to non-negative, equal weights with a sum of 1. The demand deviation direction and amount are extracted from the demand feedback results, and the gap deviation sequence and remaining number source deviation sequence are extracted from the evaluation feedback results. The demand deviation amount, gap deviation value, and remaining number source deviation value are converted into absolute values. Interval standardization is performed using the maximum and minimum values ​​of the absolute values ​​of the same type of deviation. When the maximum and minimum values ​​are the same, the corresponding standardized value is recorded as 0, resulting in a standardized feedback sequence. The dispersion of the three types of standardized deviations is calculated using the CRITIC weighting method. The inter-item correlation of the three types of standardized deviations is calculated. The values ​​obtained by subtracting the correlation coefficient from 1 are summarized as the inter-item conflict level. The dispersion of each type of deviation is multiplied by the inter-item conflict level to obtain the feedback information quantity. Each type of feedback information quantity is then divided by the sum of all feedback information quantities to obtain the demand deviation feedback weight, gap deviation feedback weight, and remaining source deviation feedback weight. The demand estimation parameters use the demand deviation feedback weight, and the scheme evaluation parameters use the sum of the gap deviation feedback weight and the remaining source deviation feedback weight, establishing a fixed correspondence between the feedback weights and the corresponding parameter groups. Parameter indices are set according to the sequence numbers of the demand estimation parameters and scheme evaluation parameters, and the number of parameter corrections is recorded. An initial disturbance symbol sequence with values ​​of 1 or -1 is generated based on the parity relationship between the parameter index and the sum of the parameter correction times. Demand feedback symbols are determined for the demand estimation parameter group according to the direction of demand deviation. The signed values ​​in the gap deviation sequence and the remaining source deviation sequence are summarized to determine the evaluation feedback symbols for the scheme evaluation parameter group. When the summarized value is not less than 0, the corresponding feedback symbol is recorded as 1; when the summarized value is less than 0, the corresponding feedback symbol is recorded as -1. Each initial disturbance symbol in the initial disturbance symbol sequence is multiplied by the corresponding parameter group's feedback symbol to obtain the corrected disturbance symbol sequence. Within the normalized parameter value range of 0 to 1, the basic disturbance amplitude is set to 0.1, and the lower limit of the disturbance amplitude is set to 0.05. Multiply the larger of the feedback weight corresponding to the parameter and the lower limit of the disturbance amplitude by the basic disturbance amplitude to obtain the disturbance amplitude sequence; when the disturbance sign is corrected to 1, increase the disturbance amplitude on the corresponding parameter to form the first disturbance parameter, and decrease the disturbance amplitude on the corresponding parameter to form the second disturbance parameter; when the disturbance sign is corrected to -1, perform the opposite process to obtain the first disturbance parameter group and the second disturbance parameter group, so that the difference between the two groups of corresponding parameters is twice the corresponding disturbance amplitude, and the lower limit of the disturbance amplitude greater than 0 ensures that the difference between the two groups of corresponding parameters is not 0; when backcalculating the current allocation cycle, fix the resource status data, reception capacity, adjustment restrictions, actual request volume and registration resource item arrangement order, and substitute the first disturbance parameter group and the second disturbance parameter group into the demand estimation, supply and demand status marking, candidate resource allocation scheme generation and scheme evaluation process, respectively, to regenerate the reservation demand estimation result, target resource allocation scheme and expected usage result; compare the reservation demand estimation results obtained from the two backcalculations with the cumulative actual request sequence, and then... The candidate demand gap sequence obtained from each back-calculation is compared with the actual demand gap sequence, and the candidate remaining source sequence obtained from the two back-calculations is compared with the cumulative actual remaining sequence. The absolute values ​​of demand deviation, gap deviation, and remaining source deviation are calculated. The three types of absolute deviation values ​​are multiplied by their corresponding feedback weights and summed to obtain the first disturbance feedback value and the second disturbance feedback value. The difference between the first disturbance feedback value and the second disturbance feedback value is divided by the non-zero difference between the corresponding parameters in the first disturbance parameter group and the second disturbance parameter group to obtain the correction amount for each parameter. According to the parameter index, the corresponding parameter correction amount is subtracted from each parameter in the demand estimation parameters and the scheme evaluation parameters, and negative numbers are truncated to 0. Then, they are divided by the sum of the parameters of the same type to complete the normalization, so as to obtain the updated demand estimation parameters and the updated scheme evaluation parameters. After each parameter update, the parameter correction count is increased by 1, and the disturbance symbol generation, two-sided disturbance, current cycle back-calculation, and parameter correction are repeatedly performed using the updated parameters. The parameter correction threshold is set to 0.001. The maximum number of parameter corrections is set to 20. Iteration stops when the absolute value of all parameter corrections is not greater than the parameter correction threshold, or when the number of parameter corrections reaches the maximum number of parameter corrections. After stopping iteration, the parameters obtained from the last normalization are determined as the updated demand estimation parameters and updated scheme evaluation parameters, and are used for demand estimation and scheme evaluation in the next allocation cycle. Compared with the existing SPSA algorithm that uses random disturbance symbols and uniform disturbance amplitude, this algorithm uses parameter index and parameter correction count to generate a reproducible initial disturbance symbol sequence, reducing parameter update fluctuations caused by random disturbances. The actual reservation feedback symbol is used to correct the disturbance direction, making the parameters... The search direction is adjusted according to the current supply-demand deviation and the deviation in scheme execution; the CRITIC weighting method is used to allocate feedback weights based on the dispersion and inter-item correlation of the feedback data, so that feedback data with high differences and less repetitive information receives a larger perturbation effect; the perturbation amplitude is adjusted by the feedback weights and a lower limit for the perturbation amplitude is set to take into account the adaptability of different feedback items and avoid the difference between the two-sided perturbation parameters being 0; two sets of perturbation parameters are used to complete the estimation of the correction amount of all parameters, avoiding the increase of back-calculations with the number of parameters, thus maintaining the computational complexity characteristics of the SPSA algorithm, and the demand estimation parameters and scheme evaluation parameters are corrected in a closed loop through actual reservation results.

[0033] In this embodiment, obtaining the target resource allocation scheme for the next allocation cycle specifically involves: The resource status data is updated based on the actual reservation results, the target resource allocation plan, the updated demand estimation parameters, and the update plan evaluation parameters to obtain the resource status data for the next cycle. The update involves writing the actual reservation volume into the segmented historical reservation data and writing the adjusted number of available slots into the segmented number of available slots configuration. Collect the current reservation request volume for the next allocation cycle, and perform demand estimation based on the resource status data for the next cycle to obtain the reservation demand estimation result for the next cycle. The demand estimation adopts the updated demand estimation parameters to perform weighted synthesis of historical reservation intensity, current request intensity and demand change sequence. Based on the estimated appointment demand for the next cycle, the segmented appointment allocation, the patient capacity, and the adjustment restrictions, the supply and demand status is marked to obtain the expected resource items to be transferred in the next week and the resource items that can be transferred out in the next cycle. Pair the expected resource item sequence for the next week with the available resource item sequence for the next cycle to obtain the candidate resource allocation scheme for the next cycle. The pairing order is determined according to the number of number items to be transferred in, the number of number items available for transfer, and the start time of the effective period. Based on the estimated reservation demand for the next cycle, the demand satisfaction, appointment utilization, and resource adjustment of each candidate resource allocation scheme for the next cycle are calculated to obtain the expected evaluation allocation scheme for the next week. The expected evaluation allocation scheme for the next week is obtained by deleting schemes whose demand satisfaction and appointment utilization are not higher than another scheme and whose resource adjustment is not lower than another scheme. The evaluation values ​​of each expected evaluation allocation scheme for the next week are calculated by updating the scheme evaluation parameters and then filtered to obtain the target resource allocation scheme for the next allocation cycle. The filtering is carried out by arranging the evaluation values ​​from largest to smallest, and selecting the expected evaluation allocation scheme for the next week with the smaller resource adjustment amount when the evaluation values ​​are the same.

[0034] Example 2: A dynamic allocation system for hospital appointment registration resources, comprising: The resource status construction module is used to establish registration resource items and generate resource status data based on the schedule, appointment configuration, historical reservation data, demand estimation parameters and scheme evaluation parameters. The demand estimation module is used to generate reservation demand estimation results based on the current reservation request volume, resource status data, and demand estimation parameters. The resource allocation module is used to perform supply and demand status marking based on the patient capacity, adjustment restrictions, and appointment demand estimation results, and generate candidate resource allocation plans. The scheme evaluation and screening module is used to calculate the demand satisfaction, number of slots utilized, and resource adjustment of each candidate resource allocation scheme, and to generate the target resource allocation scheme and expected usage results through scheme evaluation parameters. The appointment execution feedback module is used to update the appointment resource configuration according to the target resource allocation plan and process appointment requests, and generate demand feedback results and evaluation feedback results based on the actual appointment results; The parameter correction module is used to perturb the demand estimation parameters and scheme evaluation parameters using the improved SPSA algorithm, calculate the feedback weights using the CRITIC weighting method, and generate updated demand estimation parameters and updated scheme evaluation parameters. The periodic allocation module is used to update resource status data based on actual reservation results, updated demand estimation parameters, and updated scheme evaluation parameters, and to generate the target resource allocation scheme for the next allocation cycle.

[0035] Example 3: This example provides a computer storage medium on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of any of the methods described above.

[0036] The computer storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or any combination thereof. More specific examples of computer storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), registers, hard disks, optical fibers, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof, or any other form of computer storage medium in the art.

[0037] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside within an Application Specific Integrated Circuit (ASIC). In embodiments of the invention, the computer storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device.

[0038] Example 4:

[0039] To verify the feasibility of this invention in practice, it was applied to a dynamic allocation test scenario for outpatient appointment registration resources. The application targets were appointment resource items categorized by department, doctor, and valid time slot. The implementation environment consisted of an appointment registration business server, an appointment resource configuration database, and an appointment request interface. The computer program was stored on computer storage media and executed by a processor. The test scenario included six appointment resource items, collecting scheduling data, appointment resource configurations, historical appointment data, and current appointment request volume for 20 allocation cycles, and configuring the patient capacity. Existing processing methods do not utilize actual appointment results to correct demand estimation parameters and scheme evaluation parameters.

[0040] During implementation, departments, doctors, start and end times of time slots are extracted from the scheduling and appointment allocation. These are then grouped and sorted by department and doctor to form a time slot boundary sequence. The scheduling and appointment allocation are further segmented according to adjacent times, retaining the segments that simultaneously fall within both scheduling and appointment allocation time slots, resulting in an effective time slot sequence. The number of appointment slots is allocated based on the proportion of effective time slots to the total appointment allocation time slot. The initial allocation is rounded down, and unallocated appointment slots are added sequentially according to the duration of the effective time slots, from longest to shortest, forming segmented appointment allocations. Historical appointment data, demand estimation parameters, and scheme evaluation parameters are written into the corresponding effective time slots to form resource status data. The initial demand estimation parameters are set with a weight of 0.5 for historical appointment volume and 0.5 for current appointment request volume. Demand changes are adjusted in the same direction using the current appointment request volume weight. The weights for demand fulfillment, appointment utilization, and resource adjustment are all set to 0.3333.

[0041] For each registration resource item, the historical reservation quantity and current reservation request quantity are statistically analyzed, and then divided by the duration of the effective period to obtain the historical reservation intensity sequence and the current request intensity sequence. The difference between the two sequences is used to determine the direction and amount of demand change. The product of historical reservation intensity and historical reservation quantity weight, and the product of current request intensity and current reservation request quantity weight are added together. Then, the product of demand change amount and current reservation request quantity weight is increased or decreased according to the direction of demand change to obtain the demand intensity estimation sequence. This sequence is then multiplied by the corresponding duration of the effective period to form the reservation demand estimation result.

[0042] Adjustment restrictions are set before calculating the available appointment slots. The number of appointment slots in the segmented appointment slot configuration is subtracted from the number of historical appointments in the segmented historical appointment data; the positive difference is recorded as the upper limit of the allowed number of appointment slots to be transferred out. The reception capacity is subtracted from the number of appointment slots in the segmented appointment slot configuration; the positive difference is recorded as the upper limit of the allowed number of appointment slots to be transferred in. Registration resources within the same department are set as allowed pairing ranges. The available appointment slots are the smaller value between the number of appointment slots in the segmented appointment slot configuration and the reception capacity. The estimated appointment demand is compared with the available appointment slots to determine a shortage, surplus, or balanced state. The system performs pairing according to the order of appointment slot size and the start time of the effective time period, respectively. The minimum value among the number of appointment slots to be transferred in, the number of appointment slots to be transferred out, and the number of appointment slots allowed for adjustment is determined as the number of appointment slots to be adjusted this time, resulting in a candidate resource allocation scheme.

[0043] During scheme evaluation, a scheme time sequence record is generated according to department, doctor, and effective time period. Within the same department and doctor's corresponding group, the estimated appointment demand, available appointment slots, and adjusted appointment slots are accumulated item by item. The accumulated values ​​are reset when the department or doctor changes. An initial demand gap sequence is formed from the accumulated demand estimation sequence and the accumulated available slot sequence. A candidate demand gap sequence and a candidate remaining appointment slot sequence are formed from the accumulated adjusted appointment slots and the accumulated demand estimation sequence. The system calculates the demand fulfillment rate, appointment slot utilization rate, and resource adjustment rate. Candidate resource allocation schemes whose demand fulfillment rate and appointment slot utilization rate are both no higher than another scheme, and whose resource adjustment rate is no lower than another scheme, are deleted. Then, an evaluation value is calculated based on the scheme evaluation parameters. The scheme with the highest evaluation value is selected as the target resource allocation scheme, and the corresponding candidate demand gap sequence and candidate remaining appointment slot sequence are determined as the expected usage results.

[0044] The system writes the adjusted number of available appointment slots into the corresponding slot configuration and processes appointment requests according to their arrival order. If the number of already booked slots is less than the adjusted number, the appointment is recorded as successful; if the adjusted number is reached, the appointment is recorded as unsuccessful. The number of appointment requests received by each registration resource item is recorded as the actual request quantity, and the number of successful appointments is recorded as the actual appointment quantity. The actual remaining number of appointment slots is obtained by subtracting the actual appointment quantity from the adjusted number of available slots. The actual appointment results are grouped by department and doctor, and then arranged within each group according to the effective time slot. These results are accumulated to form a cumulative actual request sequence, a cumulative actual appointment sequence, and a cumulative actual remaining sequence. The cumulative actual request sequence is compared with the cumulative demand estimation sequence to form a demand feedback result. The positive difference between the cumulative actual request quantity and the cumulative actual appointment quantity forms the actual demand gap sequence. The actual demand gap sequence is compared with the candidate demand gap sequence to form a gap deviation sequence. The cumulative actual remaining sequence is compared with the candidate remaining slot sequence to form a remaining slot deviation sequence. These are combined to obtain the evaluation feedback result.

[0045] The improved SPSA algorithm converts demand deviation, gap deviation sequence, and remaining appointment source deviation sequence into absolute deviation values ​​and performs interval standardization. When the maximum and minimum values ​​of the same category are the same, the standardized value is recorded as 0. The CRITIC weighting method calculates feedback weights based on the dispersion and inter-item correlation of the three types of standardized deviations. The algorithm generates an initial disturbance symbol sequence with values ​​of 1 or -1 according to the parity relationship between parameter indices and parameter correction times. Feedback symbols are determined based on the sum of signed deviations and multiplied to obtain the corrected disturbance symbol sequence. The basic disturbance amplitude is set to 0.1, and the lower limit of the disturbance amplitude is set to 0.05. The disturbance amplitude sequence is calculated based on the feedback weights. The system performs forward and reverse disturbances, forming the first and second disturbance parameter sets, and recalculates the current allocation cycle. During the recalculation, fixed resource status data, reception capacity, adjustment limits, and actual request volume are used. The recalculation results are compared with the actual reservation results and summarized according to the feedback weights to obtain the first and second disturbance feedback values. Then, the parameter correction amount is calculated and the parameters are normalized. The parameter correction threshold is set to 0.001, the maximum number of parameter corrections is set to 20, and the iteration ends when any stopping condition is met.

[0046] Table 1 Current Allocation Cycle and Parameter Update Results

[0047] In Table 1, the actual request volumes for R1, R3, and R5 are higher than the estimated reservation demand results of 5, 4, and 4, respectively, resulting in actual demand gaps of 3, 1, and 1. After parameter correction, the estimated reservation demand results for the next cycle are adjusted to 38, 24, and 21, respectively, with the correction direction consistent with the demand feedback results. R2, R4, and R6 have one actual remaining appointment slot, and the estimated reservation demand results for the next cycle are correspondingly reduced to 26, 30, and 24. The improved SPSA algorithm meets the stopping condition after the 12th parameter correction. The demand estimation parameters are updated from 0.5 and 0.5 to 0.42 and 0.58, and the scheme evaluation parameters are updated to 0.39, 0.37, and 0.24. All parameters remain non-negative, and the sum of similar parameters is 1. After re-executing demand estimation, supply and demand status marking, candidate resource allocation scheme generation, and scheme evaluation using the updated parameters, the total absolute deviation of the reservation demand estimation for the six registered resource items decreases from 19 to 6, the total actual demand gap decreases from 9 to 4, and the total actual remaining appointment slot decreases from 7 to 3.

[0048] This embodiment demonstrates that resource status construction, demand estimation, candidate resource allocation scheme generation, scheme evaluation, reservation execution feedback, and parameter correction can form a continuous processing chain, and the actual reservation results can be simultaneously fed back to the demand estimation parameters and scheme evaluation parameters. The improved SPSA algorithm uses feedback signs to correct the perturbation direction and combines the CRITIC weighting method to adjust the perturbation amplitude, so that the target resource allocation scheme for the next allocation cycle is updated according to changes in the actual request volume, actual reservation volume, and actual remaining number resources.

[0049] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for dynamically allocating hospital appointment registration resources, characterized in that, include: Read the schedule, appointment allocation, historical appointment data, demand estimation parameters and scheme evaluation parameters, and establish registration resource items by department, doctor and time period to obtain resource status data; Collect the current number of reservation requests, and calculate the reservation demand estimate for each registration resource item based on resource status data and demand estimation parameters; Read the patient reception capacity and adjustment limits, calculate the number of available appointment slots, compare the estimated appointment demand with the number of available appointment slots, and perform supply and demand status marking to obtain candidate resource allocation schemes; Based on the estimated reservation demand, the demand fulfillment, appointment utilization, and resource adjustment of each candidate resource allocation scheme are calculated. Evaluation values ​​are calculated and filtered through scheme evaluation parameters to obtain the target resource allocation scheme and expected usage results. Update the appointment slot configuration and process appointment requests according to the target resource allocation plan, collect actual appointment results, compare the actual appointment results with the estimated appointment demand results and expected usage results, and obtain demand feedback results and evaluation feedback results. Based on the demand feedback and evaluation feedback results, the improved SPSA algorithm is used to perturb the demand estimation parameters and scheme evaluation parameters to obtain updated demand estimation parameters and updated scheme evaluation parameters. Demand estimation is performed using updated demand estimation parameters. After supply and demand status marking and candidate resource allocation schemes are generated, scheme evaluation is performed using updated scheme evaluation parameters to obtain the target resource allocation scheme for the next allocation cycle.

2. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The process involves reading scheduling data, appointment slot configuration, historical appointment data, demand estimation parameters, and scheme evaluation parameters, and establishing appointment resource items by department, doctor, and time slot. Specifically: Extract the department, doctor, start time and end time of the time period from the scheduling and appointment allocation. Then, collect the start time and end time of the time period by department and doctor and sort them to obtain the time period boundary sequence. The scheduling and appointment allocation are divided according to adjacent times in the time period boundary sequence. The segmented intervals that are simultaneously located in the scheduling time period and the appointment allocation time period are retained to obtain the effective time period sequence. The number of number sources is divided according to the proportion of the effective time period to the number source allocation time period in the effective time period sequence. The result of the division is rounded and the number of unallocated number sources is added in order of the effective time period from longest to shortest to obtain the segmented number source allocation. Filter historical appointment data with the same department and doctor, and write the appointment time in the historical appointment data into the corresponding valid time period to obtain segmented historical appointment data; The resource item record is obtained by matching demand estimation parameters and scheme evaluation parameters with department, doctor and effective time period, and writing them together with the segmented appointment source configuration and segmented historical appointment data into the effective time period sequence. Based on the resource item records, registration resource items are created and arranged in order of department, doctor, and time period to obtain resource status data.

3. The method for dynamically allocating hospital appointment registration resources according to claim 1, characterized in that, The process of collecting the current number of reservation requests and calculating the reservation demand estimate for each registration resource item based on resource status data and demand estimation parameters is as follows: Extract the effective time period and segmented historical reservation data of each registered resource item from the resource status data, count the number of historical reservations and calculate the duration of the effective time period to obtain the historical reservation volume sequence and the time period length sequence. Collect current appointment requests and write them into the corresponding registration resource items according to department, doctor and appointment time. Count the current appointment request volume of each registration resource item to obtain the current request volume sequence. Divide the quantities in the historical reservation volume sequence and the current request volume sequence by the corresponding duration in the time period length sequence to obtain the historical reservation intensity sequence and the current request intensity sequence, respectively. Compare the current request intensity sequence with the historical reservation intensity sequence, determine the direction of demand change according to the sign of the difference, and determine the amount of demand change according to the absolute value of the difference, thus obtaining the demand change sequence; Based on the demand estimation parameters, the historical reservation intensity sequence, the current request intensity sequence, and the demand change sequence are weighted and synthesized, and the weighted result corresponding to the demand change is corrected in the same direction according to the direction of demand change to obtain the demand intensity estimation sequence. The estimated demand intensity sequence is multiplied by the corresponding duration in the time period length sequence, and then written into the resource status data in the order of the registered resource items to obtain the reservation demand estimation result.

4. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The process of reading the patient reception capacity and adjusting the limits, calculating the number of available appointment slots, comparing the estimated appointment demand with the number of available appointment slots, and performing supply and demand status marking is as follows: Extract the segmented appointment slot configuration and appointment demand estimation results from the resource status data, collect the patient reception capacity and adjustment limits, and arrange them according to the order of registration resource items to obtain the resource constraint sequence; Extract the patient capacity from the resource constraint sequence item by item, compare the number of appointment slots in the segmented appointment slot configuration with the corresponding patient capacity, determine the number of available appointment slots, and obtain the sequence of available appointment slots. Subtract the corresponding number in the available appointment slot sequence from the estimated appointment demand, and determine the number of appointment slots to be transferred in and the number of appointment slots to be transferred out. Select the registration resource items with a quantity of available appointment slots and a quantity of available appointment slots that are greater than 0, and sort them in descending order of the quantity of available appointment slots and the quantity of available appointment slots, respectively, to obtain the sequence of available appointment resource items and the sequence of available appointment resource items. According to the adjustment restrictions in the resource constraint sequence, the sequence of resource items to be transferred in and the sequence of resource items to be transferred out are matched, and the number of source numbers for this adjustment is determined; Based on the adjusted number of available appointment slots, the corresponding number of available appointment slots to be transferred in and the number of available appointment slots to be transferred out are deducted respectively. For registration resource items with values ​​greater than 0, pairing continues until there are no more pairings that meet the adjustment restrictions, and the first candidate resource allocation scheme is obtained. The sequence of resources to be transferred in and the sequence of resources to be transferred out are rearranged according to the start time of the effective time period to obtain the second candidate resource allocation scheme, which is then merged with the first candidate resource allocation scheme to obtain the candidate resource allocation scheme.

5. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The specific steps for obtaining the target resource allocation scheme and expected usage results are as follows: According to the order of department, doctor and effective time period, extract the available number of appointment slots, the estimated appointment demand, and the adjusted number of appointment slots in each candidate resource allocation scheme to obtain the scheme time sequence record; By recording the estimated reservation demand and the number of available appointment slots item by item along the time sequence of the plan, the cumulative demand estimation sequence and the cumulative available slot sequence are obtained. The initial demand gap is determined based on the cumulative demand estimation sequence and the cumulative allocable sequence, thus obtaining the initial demand gap sequence. For each candidate resource allocation scheme, the number of available slots is accumulated and adjusted along the time sequence of the scheme, and compared with the cumulative demand estimation sequence to obtain the candidate demand gap sequence and the candidate remaining slot sequence; Calculate the demand fulfillment and utilization of license plates based on the initial demand gap sequence, the candidate demand gap sequence, and the adjusted number of available license plates. Summarize the number of available slots adjusted in each candidate resource allocation scheme to obtain the resource adjustment amount, and write the demand fulfillment amount, slot utilization amount, and resource adjustment amount into the corresponding candidate resource allocation scheme, and determine the allocation scheme to be evaluated; The evaluation values ​​of each allocation scheme to be evaluated are calculated by using scheme evaluation parameters and then filtered to obtain the target resource allocation scheme. The candidate demand gap sequence and candidate remaining number source sequence corresponding to the target resource allocation scheme are then determined as the expected usage results.

6. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The specific details of obtaining the demand feedback results and evaluation feedback results are as follows: According to the target resource allocation plan, the adjusted number of available appointment slots is written into the appointment slot configuration of the corresponding registration resource item. The appointment requests are processed and the actual number of requests, actual number of appointments, and actual number of remaining appointment slots are counted to obtain the actual appointment results. The actual appointment results are grouped according to department and doctor, and then arranged according to the start time of the valid time period within each group to obtain the actual usage time sequence record; By sequentially adding up the actual request volume, actual reservation volume, and actual remaining number of slots along the actual usage timeline, we can obtain the cumulative actual request sequence, the cumulative actual reservation sequence, and the cumulative actual remaining sequence. The cumulative actual request sequence corresponding to the actual reservation results is compared with the cumulative demand estimation sequence corresponding to the reservation demand estimation results to obtain the demand feedback results; The actual demand gap is determined based on the cumulative actual request sequence and the cumulative actual reservation sequence, thus obtaining the actual demand gap sequence. Candidate demand gap sequences and candidate remaining number source sequences are extracted from the expected usage results, and compared with the actual demand gap sequence and the cumulative actual remaining sequence respectively to obtain the gap deviation sequence and the remaining number source deviation sequence; The evaluation feedback results are obtained by combining the gap deviation sequence and the remaining number source deviation sequence according to the order of the registration resource items.

7. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The improved SPSA algorithm is as follows: Extract the demand deviation direction and demand deviation amount from the demand feedback results, and extract the gap deviation sequence and remaining number source deviation sequence from the evaluation feedback results to obtain the feedback sequence; The absolute value of the deviation is calculated based on the feedback sequence and then standardized to obtain the standardized feedback sequence. The CRITIC weighting method is used to calculate the feedback weights based on the dispersion of each deviation and the degree of inter-item correlation in the standardized feedback sequence; Set up parameter indexes according to the sequence numbers of demand estimation parameters and scheme evaluation parameters, record the number of parameter corrections, and establish an initial disturbance symbol sequence; Summarize the signed values ​​of demand deviation, gap deviation sequence, and remaining source deviation sequence, and determine the feedback sign; Set a base perturbation amplitude and a lower limit for the perturbation amplitude that are greater than 0, and calculate the perturbation amplitude sequence based on the feedback weights; The corrected perturbation symbol sequence is obtained by multiplying each initial perturbation symbol in the initial perturbation symbol sequence with its corresponding feedback symbol. According to the corrected disturbance symbol sequence and disturbance amplitude sequence, positive and negative disturbances are applied to the demand estimation parameters and the scheme evaluation parameters respectively to obtain the first disturbance parameter set and the second disturbance parameter set; The first and second disturbance parameter sets are used to back-calculate the current allocation period, and the obtained reservation demand estimation results and expected usage results are compared with the actual reservation results and summarized according to the feedback weight to obtain the first disturbance feedback value and the second disturbance feedback value. The difference between the first disturbance feedback value and the second disturbance feedback value is divided by the difference between the corresponding parameters in the first disturbance parameter group and the second disturbance parameter group to obtain the parameter correction amount; According to the parameter index, subtract the corresponding parameter correction amount from each parameter in the demand estimation parameters and the scheme evaluation parameters, and perform normalization respectively to obtain the updated demand estimation parameters and the updated scheme evaluation parameters.

8. The method for dynamic allocation of hospital appointment registration resources according to claim 1, characterized in that, The specific steps for obtaining the target resource allocation scheme for the next allocation cycle are as follows: The resource status data is updated based on the actual reservation results, the target resource allocation plan, the updated demand estimation parameters, and the updated plan evaluation parameters to obtain the resource status data for the next cycle. Collect the current reservation request volume for the next allocation cycle, and perform demand estimation based on the resource status data for the next cycle to obtain the reservation demand estimation result for the next cycle; Based on the estimated appointment demand for the next cycle, the segmented appointment allocation, the patient capacity, and the adjustment restrictions, the supply and demand status is marked to obtain the expected resource items to be transferred in the next week and the resource items that can be transferred out in the next cycle. Pair the sequence of resource items expected to be added in the next week with the sequence of resource items that can be added out in the next cycle to obtain the candidate resource allocation scheme for the next cycle. Based on the estimated reservation demand for the next cycle, calculate the demand satisfaction, appointment utilization, and resource adjustment for each candidate resource allocation scheme for the next cycle to obtain the expected evaluation allocation scheme for the next week. By updating the evaluation parameters of the scheme, calculating the evaluation value of each expected evaluation allocation scheme for the next week, and filtering them, the target resource allocation scheme for the next allocation cycle is obtained.

9. A dynamic allocation system for hospital appointment registration resources, used to execute the dynamic allocation method for hospital appointment registration resources as described in any one of claims 1 to 8, characterized in that, include: The resource status construction module is used to establish registration resource items and generate resource status data based on the schedule, appointment configuration, historical reservation data, demand estimation parameters and scheme evaluation parameters. The demand estimation module is used to generate reservation demand estimation results based on the current reservation request volume, resource status data, and demand estimation parameters. The resource allocation module is used to perform supply and demand status marking based on the patient capacity, adjustment restrictions, and appointment demand estimation results, and generate candidate resource allocation plans. The scheme evaluation and screening module is used to calculate the demand satisfaction, number of slots utilized, and resource adjustment of each candidate resource allocation scheme, and to generate the target resource allocation scheme and expected usage results through scheme evaluation parameters. The appointment execution feedback module is used to update the appointment resource configuration according to the target resource allocation plan and process appointment requests, and generate demand feedback results and evaluation feedback results based on the actual appointment results; The parameter correction module is used to perturb the demand estimation parameters and scheme evaluation parameters using the improved SPSA algorithm, calculate the feedback weights, and generate updated demand estimation parameters and updated scheme evaluation parameters. The periodic allocation module is used to update resource status data based on actual reservation results, updated demand estimation parameters, and updated scheme evaluation parameters, and to generate the target resource allocation scheme for the next allocation cycle.

10. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed by a processor, implements the method of any one of claims 1-8.

Images