A scheduling iteration adjustment method, device, equipment and medium

Abstract

This invention relates to the field of artificial intelligence technology, and discloses a method, apparatus, equipment, and medium for iterative adjustment of shift scheduling. This method, through shift type matching and adjustability screening, jointly measures and iteratively optimizes the distribution of manpower across time periods to obtain balanced manpower deployment data. Addressing the problem that existing scheduling systems lack means to jointly measure and iteratively optimize the degree of manpower fit and shift dispersion in each time period after automatic scheduling, leading to uneven manpower deployment, this invention uses a joint measurement of manpower fit and shift dispersion to comprehensively evaluate the distribution of manpower across time periods. Then, through shift iterative optimization and convergence determination, it can be applied to the financial and medical fields, effectively improving the degree of fit between actual manpower and target demand in each time period, as well as the dispersion of shift events along the time axis, thereby ensuring the balance of service capacity throughout the day.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a scheduling iteration adjustment method, apparatus, equipment, and medium. Background Technology

[0002] In financial services and healthcare, the scientific scheduling and rational allocation of human resources are crucial for ensuring service quality. In the financial sector, customer service centers handle various customer service needs, including loan approval consultation, policy inquiries, claims processing, and account management. The number of calls and work orders fluctuates significantly across different business hours, demanding highly refined management of agent scheduling. Similarly, in healthcare, hospital appointment scheduling systems, health management service hotlines, and remote consultation platforms face the challenge of fluctuating user demand throughout the day. The balance of staff allocation across different time periods directly impacts waiting times and the overall patient experience. Therefore, achieving a balanced deployment of staff across all service hours is a common technical challenge that urgently needs to be addressed in these business sectors.

[0003] Currently, mainstream scheduling systems typically employ automatic scheduling technology based on call volume or business volume forecasting. This involves predicting business volume for each time period within the scheduling cycle based on historical data and automatically generating staffing plans for each shift. However, in practical applications, existing scheduling systems often exhibit significant imbalances in staffing deployment across different time periods after automatic scheduling. Specifically, the actual number of seats allocated in some time periods far exceeds the target staffing needs, resulting in redundant and wasted human resources. Conversely, other time periods face substantial staffing shortages, leading to decreased service connection rates and longer customer wait times. The root cause of these problems lies in the lack of a technical means for joint measurement and iterative optimization of the fit between staffing distribution and target demand across different time periods, as well as the dispersion of shift events along the timeline. This prevents refined and balanced adjustments to staffing deployment after scheduling, making it difficult to guarantee the balance and stability of service capacity throughout the day. Summary of the Invention

[0004] This invention provides a scheduling iteration adjustment method, device, equipment, and medium to solve the technical problem that existing scheduling systems lack means to jointly measure and iteratively optimize the degree of manpower fit and the degree of shift dispersion in each time period after automatic scheduling, resulting in uneven manpower deployment in different time periods.

[0005] Firstly, a scheduling iteration adjustment method is provided, including: Obtain the business configuration data of the shift scheduling cycle, and perform shift type personnel matching processing on the business configuration data to establish the association mapping of time period-person-shift type to obtain personnel configuration data; The personnel configuration data is filtered for available shifts and personnel adjustability to remove single-shift data and unscheduled data and mark personnel who can be finely adjusted, thus obtaining adjustable shift data. Based on the adjustable shift data, the manpower distribution of the shift scheduling period is initialized to perform time slice discretization and shift time window mapping to obtain the manpower distribution data of the time period. The manpower distribution data for the time period is fitted, and the number of shifts starting is counted based on the fitting results. The shift dispersion joint measurement is performed based on the number of shifts starting to obtain comprehensive and balanced measurement data. Based on the comprehensive and balanced measurement data, iterative deployment data is obtained by changing the currently scheduled shifts and performing shift iteration processing. The iterative deployment data is subjected to iterative convergence condition determination and termination output processing to obtain balanced manpower deployment data.

[0006] Secondly, a scheduling iteration adjustment device is provided, comprising: The data configuration module is used to obtain business configuration data for the shift scheduling cycle, perform shift type personnel matching processing on the business configuration data, so as to establish a correlation mapping between time period, personnel and shift type to obtain personnel configuration data; The shift adjustment module is used to filter the personnel configuration data for available shifts and personnel adjustability, so as to remove single-shift data and unscheduled data and mark personnel that can be finely adjusted, so as to obtain adjustable shift data. The manpower calculation module is used to perform initial processing of manpower distribution during shift scheduling based on the adjustable shift data, so as to perform time slice discretization and shift time window mapping to obtain manpower distribution data for the time period. The joint measurement module is used to fit the manpower distribution data of the time period, and to count the number of shifts starting based on the fitting results. Based on the number of shifts starting, the module performs joint measurement processing on shift dispersion to obtain comprehensive and balanced measurement data. The iterative deployment module is used to obtain iterative deployment data by changing the currently scheduled shifts and performing shift iteration processing based on the comprehensive balance measurement data; The data output module is used to determine the iterative convergence condition and terminate the output of the iterative deployment data to obtain balanced manpower deployment data.

[0007] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described scheduling iterative adjustment method.

[0008] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described scheduling iterative adjustment method.

[0009] In the above-mentioned scheduling iteration adjustment method, device, equipment, and medium, business configuration data of the scheduling cycle can be obtained. The business configuration data is then processed for shift type and personnel matching to establish a time period-personnel-shift type association mapping, resulting in personnel configuration data. The personnel configuration data is then filtered for available shifts and personnel adjustability to remove single-shift data and unscheduled data and mark personnel who can be finely adjusted, resulting in adjustable shift data. Based on the adjustable shift data, the manpower distribution of the scheduling period is initialized by performing time slice discretization and shift time window mapping, resulting in time period manpower distribution data. The time period manpower distribution data is then fitted, and the number of starting shifts is counted based on the fitting results. Based on the number of starting shifts, a combined shift dispersion metric is performed to obtain comprehensive equilibrium metric data. Based on the comprehensive equilibrium metric data, shift iteration processing is performed by changing the currently scheduled shifts to obtain iterative deployment data. Finally, the iterative deployment data is processed for iterative convergence condition determination and termination output to obtain balanced manpower deployment data. In this invention, the existing scheduling system lacks a means to jointly measure and iteratively optimize the degree of manpower fit and shift dispersion in each time period after automatic scheduling, resulting in uneven manpower deployment in each time period. This problem can be solved by matching shift types and filtering adjustability in business configuration data, initializing the manpower distribution in each time period, and using a joint measurement of manpower fit and shift dispersion to comprehensively evaluate the manpower distribution in each time period. Then, through shift iterative optimization and convergence judgment, the fine-grained and balanced adjustment of manpower deployment in each time period after automatic scheduling can be achieved. This effectively improves the degree of fit between actual manpower and target demand in each time period and the dispersion of shift events on the time axis, thereby ensuring the balance and stability of service capacity in each time period throughout the day. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of an application environment for a scheduling iteration adjustment method according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating a scheduling iteration adjustment method according to an embodiment of the present invention; Figure 3 yes Figure 2 A detailed implementation flow diagram of step S10 Figure 1 ; Figure 4 yes Figure 2 A detailed implementation flow diagram of step S20 Figure 2 ; Figure 5 yes Figure 2 A detailed implementation flow diagram of step S30 Figure 3 ; Figure 6 yes Figure 2 A detailed implementation flow diagram of step S40 Figure 4 ; Figure 7 yes Figure 2 A schematic flowchart of a specific implementation method for step S50 Figure 5 ; Figure 8 yes Figure 2 A detailed implementation flow diagram of step S60 Figure 6 ; Figure 9 This is a schematic diagram of a scheduling iteration adjustment device according to an embodiment of the present invention; Figure 10 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention; Figure 11 This is another structural schematic diagram of a computer device according to one embodiment of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] The scheduling iteration adjustment method provided in this embodiment of the invention can be applied to, for example, Figure 1In this application environment, the client communicates with the server via a network. The server can obtain input conditions from the client, namely, the business configuration data of the scheduling cycle. It performs shift type and personnel matching processing on the business configuration data to establish a time period-personnel-shift type association mapping, obtaining personnel configuration data. The server then performs available shifts and personnel adjustability filtering processing on the personnel configuration data to remove single-shift data and unscheduled data and mark personnel who can be finely adjusted, obtaining adjustable shift data. Based on the adjustable shift data, it performs initialization processing of the manpower distribution during the scheduling period, performing time slice discretization and shift time window mapping, obtaining time period manpower distribution data. It performs fitting processing on the time period manpower distribution data, and based on the fitting results, it counts the number of starting shifts. Based on the number of starting shifts, it performs shift dispersion joint measurement processing to obtain comprehensive balanced measurement data. Based on the comprehensive balanced measurement data, it performs shift iteration processing by changing the currently scheduled shifts, obtaining iterative deployment data. Finally, it performs iterative convergence condition determination and termination output processing on the iterative deployment data, obtaining balanced manpower deployment data. This invention addresses the problem of uneven manpower deployment in existing scheduling systems after automatic scheduling, which lacks a means to jointly measure and iteratively optimize the fit between manpower and shift distribution for different time periods. This is achieved by matching shift types and filtering adjustability in business configuration data to initialize manpower distribution for each time period. A comprehensive evaluation of the manpower distribution is then conducted using a joint measurement of manpower fit and shift distribution. Through shift iterative optimization and convergence determination, refined and balanced adjustment of manpower deployment for each time period after automatic scheduling is achieved. This effectively improves the fit between actual manpower and target demand for each time period, as well as the dispersion of shift events along the time axis, thereby ensuring the balance and stability of service capacity throughout the day. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention is described in detail below through specific embodiments.

[0014] Please see Figure 2 As shown, Figure 2 A flowchart illustrating a scheduling iteration adjustment method provided in an embodiment of the present invention includes the following steps: S10: Obtain the business configuration data of the shift scheduling cycle, perform shift type personnel matching processing on the business configuration data to establish the association mapping of time period-person-shift type, and obtain personnel configuration data.

[0015] The system retrieves the scheduling cycle corresponding to the current business activity and the configuration parameters of each time period it includes, thereby establishing a correlation mapping between time period, personnel, and shift type. This information is then integrated into business configuration data. In the financial customer service scheduling scenario, this business configuration data can include the scheduling cycles and time period divisions for different business lines such as credit consultation, policy inquiry, claims processing, and account management. In the healthcare service scheduling scenario, this business configuration data can include the scheduling cycles and shift type configurations for different service types such as outpatient appointment reception, remote consultation, and health management consultation. Based on the shift type settings in the business configuration data, the system retrieves the agents configured with the corresponding shift types and establishes a two-way association between personnel and shift types. After deduplication and verification, the system summarizes the personnel configuration data for that scheduling cycle. This personnel configuration data reflects the shift type information associated with each agent within the scheduling cycle.

[0016] Combination Figure 3 As shown, step S10 specifically includes: S101: Obtain the scheduling cycle and time period configuration parameters corresponding to the business activities, and integrate them to obtain business configuration data.

[0017] Retrieve the scheduling cycle corresponding to the current business activity and the time period configuration parameters included within that scheduling cycle. The scheduling cycle refers to the scheduling management period corresponding to the business activity, such as a natural week, natural month, or a custom time interval as a complete scheduling cycle. Time period configuration parameters refer to the various time periods within that scheduling cycle, divided according to business operation needs, and their corresponding scheduling configuration information, including the start and end times of each time period, whether scheduling is enabled, and the applicable shift types within each time period. In the financial customer service field, the above-mentioned business activity could be the daily call handling service of a customer service center, the scheduling operation of an insurance company's claims processing hotline, or the duty arrangement of a financial consultation hotline. The time period configuration parameters typically need to be differentiated based on factors such as financial market trading hours, peak customer call times, and meal times. In the healthcare service field, the above-mentioned business activity could be the scheduling management of a hospital outpatient appointment dispatch center, the seat scheduling of a health management service hotline, or the staffing scheduling of a remote consultation platform. The time period configuration parameters need to be configured based on actual conditions such as peak user visit times, lunch break times, and nighttime emergency duty times. The obtained shift schedule and all time period configuration parameters are integrated to obtain business configuration data, which constitutes the basic data source for subsequent shift type personnel matching processing.

[0018] S102: Based on the business configuration data, read the page settings information of each class type to obtain class type settings data.

[0019] After obtaining the business configuration data, the system reads the page settings information for each shift type in the scheduling system one by one. A shift type refers to a predefined set of work time arrangement patterns in the scheduling system; different shift types correspond to different start times, work durations, and rest arrangements. Page settings information refers to the detailed parameters configured for each shift type on the scheduling system page, including but not limited to the shift type's name, applicable time period range, corresponding work start and end times, and whether the shift type is enabled. The system iterates through all shift types involved in the business configuration data, reading their page settings information sequentially to obtain shift type setting data. This shift type setting data contains complete setting parameters for all configured shift types within the current scheduling cycle.

[0020] S103: The class type setting data is categorized and divided according to time period intervals to obtain time period class type mapping data.

[0021] After obtaining the shift type settings data, the system needs to categorize and classify the data according to time period intervals, since different shift types may be applicable to different time periods within the scheduling cycle. Each shift type is grouped into its corresponding time period group based on the applicable time period interval configured in its page settings, thus establishing a mapping relationship between time periods and shift types. For example, in a financial customer service scheduling scenario, the morning shift may be applicable to the morning time period, the afternoon shift to the afternoon time period, and the evening shift to the evening time period. The system categorizes each shift type according to these time period intervals. After this categorization and classification process, time period shift mapping data is obtained, which clearly reflects the set of available shift types included in each time period interval.

[0022] S104: Based on the time period and class type mapping data, retrieve the corresponding class type of agent personnel to obtain agent personnel retrieval data.

[0023] After obtaining the time-slot shift mapping data, the system retrieves the corresponding agent information from the personnel database within each time slot for each shift type. Agents refer to frontline service personnel registered and assigned specific shift types in the scheduling system. In financial customer service scenarios, this could be bank loan consultants, insurance underwriting and claims processing agents, or account management agents; in healthcare service scenarios, it could be outpatient appointment operators, health management consultants, or remote consultation dispatchers. The system iterates through each shift type in the time-slot shift mapping data, retrieving all agent information configured with that shift type's qualifications, and then compiles the agent retrieval data.

[0024] S105: Perform bidirectional association binding processing on the seat personnel retrieval data to obtain personnel and shift type association data.

[0025] After obtaining the agent retrieval data, a two-way association binding process is performed on this data, linking agents to shift types. This two-way association binding refers to simultaneously establishing index relationships in two directions at the data level: from agent to shift type and from shift type to agent. This allows the system to quickly query all available shift types associated with a specified agent, and also to quickly query all agents configured for a specified shift type. This two-way association binding process ensures efficient cross-referencing and matching operations between agents and shift types during subsequent shift adjustments and iterative optimizations. After the above processing, agent-shift type association data is obtained.

[0026] S106: Perform deduplication verification on the personnel shift type association data according to the shift scheduling cycle, and summarize to obtain personnel configuration data.

[0027] After obtaining the personnel shift association data, since the same agent may be searched repeatedly in different time periods, or the same agent may be associated with multiple overlapping shift configurations within the same scheduling cycle, it is necessary to perform deduplication and verification processing on the personnel shift association data according to the scheduling cycle to eliminate redundant data records caused by multiple searches or multiple associations. After deduplication and verification, the cleaned data is summarized and integrated according to the scheduling cycle to obtain personnel configuration data. This personnel configuration data fully reflects the shift information associated with each agent in the current scheduling cycle and its corresponding time period, forming the input data foundation for subsequent available shifts and personnel adjustability filtering processing.

[0028] S20: Perform available shift and personnel adjustability filtering on the personnel configuration data to remove single-shift data and unscheduled data and mark personnel that can be finely adjusted, thereby obtaining adjustable shift data.

[0029] After obtaining the personnel allocation data, not all personnel are eligible for shift adjustments. Therefore, it is necessary to filter the personnel allocation data based on available shifts and personnel adjustability to remove single-shift data and unscheduled data, and mark personnel who can be fine-tuned, thus obtaining adjustable shift data. This adjustable shift data includes information on personnel who can be fine-tuned and their adjustable shifts, as well as shift distribution information for personnel with fixed shifts.

[0030] Combination Figure 4 As shown, step S20 specifically includes: S201: Initialize all available shifts for each person in the corresponding time period based on the personnel configuration data to obtain the available shift set data.

[0031] After obtaining the personnel configuration data, based on the shift type information associated with each agent in the personnel configuration data, the total number of available shifts that can be scheduled for that agent within the corresponding time period is initialized. Here, available shifts refer to the various candidate shifts that the agent can theoretically be assigned to within the corresponding time period of the current scheduling date, based on the association between the agent and the shift type and the time period configuration parameters. The process iterates through each agent in the personnel configuration data, determining all available shifts for each agent within each corresponding time period, and then aggregates and organizes this information to obtain the available shift set data. This available shift set data records the complete list of candidate shifts for each agent within the corresponding time period of each scheduling date, providing a complete candidate space for subsequent availability filtering processing.

[0032] S202: Mark and exclude personnel in the available shift set data who have only one available shift to obtain single shift exclusion data.

[0033] In the available shift data set, some agents may have only one available shift during a given time period. This means they can only be assigned to a single, specific shift during that time period, with no other candidate shifts available for adjustment. For such agents, due to the lack of shift adjustment options, the system marks them as non-adjustable and excludes them from subsequent shift adjustment candidates. For example, during peak insurance claims periods, some agents with specific claims review qualifications may only be assigned one dedicated shift; these agents are considered single-shift agents. In healthcare scenarios, some dedicated nursing dispatchers undertaking night emergency shifts may also only have one shift. After the above marking and exclusion process, single-shift exclusion data is obtained. This data removes agents with only one available shift from the shift adjustment candidate range. However, it should be noted that the fixed shift information of excluded agents will still be retained and included in subsequent shift distribution calculations.

[0034] S203: Mark and exclude personnel who are not scheduled for work on the same day from the single-shift exclusion data to obtain the on-duty screening data.

[0035] After obtaining the single-shift exclusion data, the system further marks and excludes personnel who were not scheduled for that day. Personnel not scheduled for that day are those who, on the current scheduling date, have not been assigned any shifts by the scheduling system and are in a rest or non-working state. Since these personnel are not on duty that day and there is no need or possibility for shift adjustments, the system also excludes them from the shift adjustment candidate pool. After the above processing, the on-duty screening data is obtained. This data only retains agents who have been scheduled for that day and have more than one available shift, ensuring that subsequent adjustability determinations are only made for on-duty personnel who have the actual conditions for shift adjustments.

[0036] S204: Determine whether the shift event time period of the personnel in the shift screening data falls within the fine-tuning time period, and integrate them to obtain the adjustability determination data.

[0037] After obtaining the on-duty screening data, the system further analyzes the event time period of each employee's currently scheduled shift to confirm whether it falls within a preset fine-tuning time period. The event time period refers to the actual working hours covered by the employee's currently scheduled shift on the timeline, while the fine-tuning time period refers to the system's pre-defined valid time range for allowing shift adjustments. An employee is only eligible for fine-tuning if their entire or partial shift event time period falls within the fine-tuning time period. The system performs this analysis on each employee in the on-duty screening data and integrates the results to obtain adjustability assessment data.

[0038] S205: Based on the adjustability determination data, mark the personnel who meet the conditions as fine-tunable personnel to obtain fine-tunable personnel data.

[0039] After obtaining the adjustability assessment data, the system marks personnel who simultaneously meet the following conditions as fine-tunable personnel: First, the personnel have already been scheduled for shifts that day, and the number of available shifts is greater than one; second, the time segment of the personnel's currently scheduled shift falls within the fine-tuning time period. Personnel meeting these conditions are candidates whose shifts can be adjusted during subsequent shift iterations and optimizations. All personnel meeting the conditions are marked and aggregated to obtain fine-tunable personnel data. This data includes the identity information of each fine-tunable personnel, their currently scheduled shift, and a list of all available candidate shifts.

[0040] S206: Incorporate the fixed shifts in the adjustable personnel data into the shift distribution calculation and summarize them to obtain adjustable shift data.

[0041] After obtaining the data on adjustable personnel, the fixed shift information of the non-adjustable personnel, who were previously excluded in steps S202 and S203, needs to be included in the overall shift distribution calculation. Although these personnel do not participate in subsequent shift adjustment operations, the distribution of their fixed shifts on the time axis still constitutes part of the overall manpower layout of the current scheduling scheme, affecting subsequent manpower fitting and shift dispersion metrics; therefore, they must be included in the statistical scope. The adjustable shift information of adjustable personnel and the fixed shift information of non-adjustable personnel are summarized and integrated to obtain adjustable shift data. This adjustable shift data includes complete information on adjustable personnel and their candidate shifts, as well as shift distribution information for fixed shift personnel, providing complete data support for the initial processing of manpower distribution during subsequent scheduling periods.

[0042] S30: Based on the adjustable shift data, perform initialization processing of manpower distribution during shift periods to discretize time slices and map shift time windows to obtain manpower distribution data for each shift period.

[0043] After obtaining the available shift data, it is necessary to perform initial statistics on the manpower distribution in each time period within the current scheduling cycle, in order to discretize the time slices and map the shift time windows to obtain the manpower distribution data for each time period. This manpower distribution data for each time period includes the actual seat distribution, target demand distribution, and the distribution of the starting number of shifts at each shift adjustment time point for each time slice.

[0044] Combination Figure 5 As shown, step S30 specifically includes: S301: Load the adjustable shift data and initialize the header configuration information of the corresponding shift schedule to obtain the shift schedule header data.

[0045] After obtaining the adjustable shift data, the data is loaded and the header configuration information for the corresponding shift scheduling cycle is initialized. The header configuration information refers to the configuration parameters used in the scheduling system to define the time structure of the shift scheduling cycle, including the total duration of the shift scheduling cycle, the start and end time boundaries, and the division method of each time period. The header configuration information for the current shift scheduling cycle is read and initialized to obtain the shift scheduling cycle header data, which provides the time structure framework for subsequent time discretization processing.

[0046] S302: Discretize the schedule period header data according to a preset time granularity to divide it into time slice sequences.

[0047] After obtaining the schedule header data, it is discretized according to a preset time granularity, dividing the continuous schedule time axis into several equally spaced discrete time slices, thus forming a time slice sequence. Here, a time slice refers to the smallest unit of time after time discretization, and the preset time granularity refers to the actual time length corresponding to each time slice. For example, using 15 minutes as a time slice granularity, a complete full-day schedule cycle can be divided into... A time slice, of which This represents the total number of time slices. In financial customer service scheduling scenarios, this time discretization allows the system to accurately calculate the agent staffing allocation within each time slot, using 15-minute units as the smallest unit. This supports refined manpower management for different business periods, such as peak credit consultation periods, concentrated insurance claims periods, and active securities trading periods. In healthcare service scheduling scenarios, this time discretization is also applicable to manpower distribution statistics for time intervals such as peak outpatient registration periods, lunch break periods, and afternoon follow-up visits.

[0048] S303: Traverse the time slice sequence and count the actual number of seats arranged in each time slice to obtain the actual seat distribution data.

[0049] After obtaining the time-slice sequence, iterate through each time slice in the sequence. Based on the current scheduled shift information of all agents in the adjustable shift data, count the number of agents actually on duty in each time slice to obtain the actual agent distribution data. This actual agent distribution data can be used... To indicate, among which For the first The actual number of seats arranged within a given time slot. This data reflects the actual manpower allocation for each time slot under the current scheduling scheme.

[0050] S304: Based on the actual seat distribution data, load the target manpower demand for each time slot to obtain the target demand distribution data.

[0051] After obtaining the actual agent distribution data, the time-slot framework of this data is used to load the pre-set target manpower requirements for each time slot in the scheduling system. The target manpower requirement refers to the expected number of agents to be allocated in each time slot, determined based on historical business volume forecasts or FTE (Full-Time Equivalent) manpower curves. To indicate, among which For the first The target manpower requirement is set for each time slice. In financial customer service scenarios, this target manpower requirement is typically determined based on call volume prediction models and average case processing time. Differences in call volume at different times will cause the target manpower requirement for each time slice to exhibit a fluctuating distribution. In healthcare service scenarios, this target manpower requirement is determined based on user appointment volume prediction and average consultation time. The system loads the target manpower requirement values ​​for all time slices to obtain the target demand distribution data.

[0052] S305: Map each shift in the target demand distribution data to a time axis with a fixed time window length and mark the start time to obtain shift event distribution data.

[0053] After obtaining the target demand distribution data, the system performs time-axis mapping on each shift in the adjustable shift data. Each shift can be considered as a fixed length. The time window in which This indicates the number of time slots covered by this service on the timeline. Each service has a start time. This corresponds to forming a sequence of length in the time slice sequence. The continuous coverage area, that is, the mapping representation of this shift's event on the timeline. The system will include all... Each shift is mapped to a timeline according to its fixed time window length, and the start time of each shift is marked. This represents the total number of shifts in the current scheduling plan, with the set of start times denoted as . Each This corresponds to a time slice index. After the above mapping and marking process, the shift event distribution data is obtained, which records the starting position and coverage area of ​​all shifts on the time axis.

[0054] S306: Based on the shift event distribution data, the number of shifts starting at each shift adjustment time point is counted to obtain the manpower distribution data for the time period.

[0055] After obtaining the shift event distribution data, statistics are compiled based on the shift start time to calculate the number of shifts with that start time at each shift adjustment time point. Here, the shift adjustment time point refers to the set of discrete time points on the timeline that can serve as the shift start time; assuming there are a total of... The shift change time points, among which This indicates the total number of shift change times. The system counts the number of shifts starting at each shift change time, obtaining the shift start count for each shift change time. Indicates the first The number of shifts starting at each shift adjustment time point, of which The actual seat distribution, target demand distribution, and shift start quantity distribution for each time slot are summarized and integrated to obtain the manpower distribution data for each time slot. This manpower distribution data comprehensively reflects the current manpower allocation status of the scheduling plan in the time dimension, providing a complete computational foundation for subsequent manpower fitting and shift dispersion joint measurement processing.

[0056] S40: Fit the manpower distribution data for the time period, and count the number of shifts starting based on the fitting results. Perform a combined measurement of shift dispersion based on the number of shifts starting to obtain comprehensive balanced measurement data.

[0057] After obtaining the manpower distribution data for each time period, in order to quantitatively assess the balance of the current manpower deployment, the system performs a joint fitting of the manpower distribution data from two dimensions: manpower fit and shift dispersion. Based on the fitting results, it counts the number of shift starts and performs a joint measurement of shift dispersion based on the number of shift starts, finally obtaining comprehensive balance measurement data. This comprehensive balance measurement data reflects the overall balance level of the current manpower deployment in both manpower fit and shift dispersion dimensions in a unified numerical form.

[0058] Combination Figure 6 As shown, step S40 specifically includes: S401: Calculate the relative deviation of manpower based on the time slices in the manpower distribution data of the time period to obtain the relative deviation data.

[0059] After obtaining the manpower distribution data for each time slot, the actual number of seats arranged for each time slot in the data is determined accordingly. With target human resource needs The relative error of manpower is calculated for each time slice. The relative error of manpower is an indicator that measures the degree of difference between the actual manpower allocation and the target demand within a given time slice. Specifically, for the... The relative deviation of manpower in each time slice The calculation formula is: in, For the first The actual number of seats allocated for each time slot. For the first The target manpower requirement for each time slot. This is the minimum demand threshold, used to avoid situations where the target demand... When the value is zero, a division by zero occurs; its default value is set to zero. This is because the smallest human unit in all agent scheduling systems is one person. When, it indicates that there is manpower redundancy in that time slot, and the actual number of seats arranged exceeds the target demand; when This indicates a manpower shortage in that time slot. For example, in credit consultation agent scheduling, if the target manpower requirement for a 15-minute time slot is 5 people, but 6 people are actually scheduled, then the relative manpower deviation for that time slot is [missing information]. This translates to a 20% redundancy in manpower. The system iterates through all... The relative deviation of manpower for each time slice is calculated, and the relative deviation data is obtained by summing them up.

[0060] S402: The relative deviation data is partitioned according to the tolerance threshold and subjected to in-threshold square measurement and out-of-threshold linear extension processing to obtain robust deviation data.

[0061] After obtaining the relative deviation data, the relative deviation of manpower in each time slice is partitioned according to a tolerance threshold, and different measurement methods are used in different partitions. This partitioned measurement strategy is adopted because the following conflicting requirements exist in actual scheduling scenarios: On the one hand, when the scheduling plan is close to the target demand, such as when 6 people are actually arranged in a medical clinic when 5 people are needed, sensitivity to such small errors is required to support fine-grained optimization; on the other hand, when there is a serious over-scheduling or under-scheduling, such as when 20 people are actually arranged in a clinic when 5 people are needed, the business considers such large errors to be serious deviations, and there is no need to strictly distinguish between the difference between 20 and 25 people. Furthermore, from the perspective of optimization stability, if the original SSE (Sum of Squared Errors) is used directly for measurement, a large deviation will lead to an excessively large gradient value (for example, the gradient value can reach 30 when the deviation is 20 people), thus causing oscillations in the optimization process. To simultaneously meet the requirements of sensitivity to small errors, robustness to large errors, and continuity of the first derivative, the system adopts the following piecewise robust measurement function. : in, The tolerance threshold represents the maximum relative deviation that is acceptable in business operations; its default value is set to [value missing]. That is, 150% error is the business tolerance limit; when the relative deviation exceeds this threshold, it is considered a serious deviation. Within the threshold range ( ), using square metric To retain sensitivity to small errors, the system can fine-tune scheduling schemes that closely approximate target requirements. Outside the threshold range ( ), using linear extension To avoid the destructive impact of large errors on the overall metric, and to ensure the continuity of the first derivative of the piecewise function at the threshold connection points, thereby ensuring gradient smoothing and computational stability in subsequent iterative optimization processes, the system applies the aforementioned piecewise robust metric function to the relative deviations of all time slices to obtain robust deviation data.

[0062] S403: Aggregate the mean of the robust deviation data over the entire time period and truncate it to the upper limit according to a preset empirical threshold to obtain human fitting measurement data.

[0063] After obtaining the robust bias data, the robust bias values ​​for all time slices are averaged and aggregated, then truncated to an upper limit according to a preset empirical threshold to obtain the final human fit metric. Human Fit Metric ( The formula for calculating ) is: in, The total number of time slices. The first one calculated in step S402 Robustness deviation value for each time slice, This is a preset upper limit of the experience threshold, used to truncate the upper limit of the human fitting metric. Its default value is set to [value missing]. The rationale for setting this upper limit cutoff is that when the manpower fit metric exceeds 2.0, it indicates that the current scheduling plan's manpower allocation has significantly deviated from the target requirements. From a business perspective, the scheduling plan is unusable, and further increasing the metric value is meaningless. The upper limit cutoff mechanism ensures that the manpower fit metric remains within a reasonable range, avoiding extreme deviations from unduly dominating subsequent joint measurement and iterative optimization processes. After the above-mentioned all-time average aggregation and upper limit cutoff processing, the manpower fit metric data is obtained.

[0064] S404: Based on the time slice framework of the aforementioned human resource fitting measurement data, the number of shifts starting at each shift adjustment time point is counted to obtain shift count data.

[0065] After obtaining the manpower fitting measurement data, based on the time slice framework established by the aforementioned time discretization, the number of shifts starting at each shift adjustment time point is counted. Using the shift start number information for each shift adjustment time point already counted in step S306, i.e., the [number of shifts starting at each time point]... The starting number of shifts at each shift adjustment time point ,in This represents the total number of shift adjustment times. In a financial customer service scheduling scenario, if the entire day's schedule can be fine-tuned, then adjustments are made in 15-minute increments. The system extracts and organizes the number of shift starts at each shift change time point to obtain shift count data. This data reflects the distribution of the number of shift start events at each shift change time point and serves as a direct input for subsequent assessment of shift dispersion.

[0066] S405: Perform normalized chi-square statistical measurement on the shift count data to obtain shift dispersion measurement data.

[0067] After obtaining the shift count data, a normalized chi-square statistical measure is applied to quantify the dispersion of shift start times along the time axis. The core of this application lies in transforming the dispersion of shift start times into a measure of the difference between an empirical distribution and a uniform distribution, and implementing this using a statistically rigorous and normalizable chi-square statistic. Shift Dispersion Measurement ( The formula for calculating ) is: in, This represents the total number of shift change times. For the first The initial number of shifts at each shift adjustment time point The expected uniform density of events per shift is calculated using the following formula: , This refers to the total number of shifts in the current scheduling plan, i.e., how many people are scheduled within this time period. That is, how much. If all... The start time of each train is at If the shift changes are perfectly uniformly distributed across all shift change times, then the expected number of shifts at each shift change time is: At this time, the actual number of shifts at each shift adjustment point. All equal to , Conversely, if the start times of all shifts are concentrated in a few shift rescheduling times, then... It approaches 1. It should be noted that the chi-square statistic after the above normalization process has clear business interpretability: when... When this occurs, it indicates a significant concentration of shift start times, suggesting poor scheduling rationality; when... When the start times are highly evenly distributed, it indicates that the scheduling is relatively reasonable. In financial customer service scheduling, excessively concentrated start times mean that a large number of agents are leaving and starting get off work at the same time, which may lead to temporary interruptions in service capacity during handover periods. In healthcare scheduling, overly concentrated start times for outpatient shifts can also affect the continuity of outpatient services. Therefore, pursuing an even distribution of shift start times is an important goal for improving the overall rationality of the scheduling plan. After the above normalized chi-square statistical measurement, the shift distribution measurement data is obtained.

[0068] S406: The shift dispersion measurement data and the manpower fitting measurement data are weighted and fused according to a preset weighting coefficient to obtain comprehensive balanced measurement data.

[0069] After obtaining the manpower fitting measurement data and shift dispersion measurement data respectively, the system weights and merges the measurement results of the two dimensions according to preset weight coefficients to obtain a unified comprehensive balanced measurement value. Comprehensive Balanced Measurement ( The formula for calculating ) is: in, The human resource fitting metric value calculated in step S403. The shift dispersion metric value calculated in step S405. These are preset weighting coefficients used to control the relative importance of shift dispersion metrics in the overall balanced metrics. Because... and Both calculations have undergone normalization, making their numerical scales comparable; therefore, the weighting coefficients... The default value can be set to This is achieved automatically through a normalization mechanism, balancing the two metric dimensions. In practical applications, business users can adjust the weighting coefficients according to their specific scheduling management priorities. Adjustments can be made, for example, when the business side is more concerned about the service connection rate at different times, it can be appropriately reduced. The value should primarily focus on manpower fit; however, it can be appropriately increased when business stakeholders prioritize the rationality and fairness of the scheduling plan. The value emphasizes shift frequency distribution. After the above weighted fusion process, a comprehensive and balanced metric is obtained, which is presented as a single numerical value. It comprehensively reflects the overall balance level of the current scheduling scheme in terms of both manpower fit and shift dispersion. The smaller the value, the higher the balance of the current scheduling scheme.

[0070] S50: Based on the comprehensive balance measurement data, iterative deployment data is obtained by changing the currently scheduled shifts and performing shift iteration processing.

[0071] After obtaining the comprehensive equilibrium measurement data, the system uses the current comprehensive equilibrium measurement value as the iterative comparison benchmark, and sequentially performs shift adjustment operations on the adjustable personnel to obtain iterative deployment data. Through the above iterative optimization process, person by person and shift by shift, the manpower deployment plan gradually approaches a better equilibrium state in each iteration.

[0072] Combination Figure 7 As shown, step S50 specifically includes: S501: Use the current total loss value of the comprehensive equilibrium metric data as the iterative comparison benchmark to obtain the benchmark loss data.

[0073] After obtaining the comprehensive balance measurement data, the comprehensive balance measurement value corresponding to the current scheduling plan will be... This serves as the initial benchmark for shift iteration. The benchmark loss data records the overall balance metric value before the iteration begins or after the last successful shift change. Each subsequent trial shift change operation will use this benchmark loss data as a comparison reference to determine whether the shift change helps improve the overall balance of the scheduling scheme.

[0074] S502: Change the currently scheduled shifts to obtain trial shift data.

[0075] After determining the baseline loss data, one adjustable staff member is selected from the pool of available personnel, and their currently scheduled shift is changed to another available candidate shift. This change is a tentative adjustment, temporarily modifying the staff member's scheduled shift at the data level for subsequent calculation and evaluation. The changed shift arrangement is recorded as trial shift data. For example, in an insurance customer service scheduling scenario, a customer service representative is currently scheduled for the morning shift (08:00-16:00), and their available candidate shifts include the afternoon shift (09:00-17:00). The system changes the staff member's scheduled shift from the morning shift to the afternoon shift, generating the corresponding trial shift data. In a healthcare scheduling scenario, an outpatient dispatcher is currently scheduled for the morning shift. The system can tentatively adjust their shift to a full-day shift or an afternoon shift to observe the impact of this adjustment on the overall manpower balance.

[0076] S503: Calculate the comprehensive balance measurement value after the shift change using the trial shift data, and compare the comprehensive balance measurement value with the baseline loss data to obtain shift change judgment data.

[0077] After obtaining the trial shift data, the joint measurement calculation process of manpower fitting and shift dispersion in step S40 is re-executed based on the trial shift data to obtain the new comprehensive equilibrium measurement value after the shift change. Then change the new metric. The current metric recorded with the baseline loss data determined in step S501 Numerical comparisons are performed to determine whether the shift change has improved or at least prevented the overall balance of the scheduling plan from deteriorating. The comparison results are then recorded as shift adjustment data.

[0078] S504: Based on the shift adjustment judgment data, the shift adjustment is confirmed when the loss value does not increase, and shift adjustment confirmation data is obtained.

[0079] After obtaining the shift change determination data, the shift change confirmation logic is executed based on the determination result. The overall balance metric value after the shift change... Less than or equal to the current metric in the baseline loss data At that time, that is This indicates that the shift change did not worsen the overall balance of the scheduling plan, the system confirmed the shift adjustment was valid, and generated shift adjustment confirmation data. Conversely, if the metric value increases after the shift change, i.e. If the change fails, it indicates that the adjustment has led to a decrease in the overall balance. The trial change will be revoked, the employee's previous shift will be restored, no shift adjustment confirmation data will be generated, and the trial adjustment operation will continue for the next candidate shift or the next available employee. This shift adjustment confirmation logic ensures that each confirmed shift adjustment will not worsen the overall balance metric of the scheduling scheme, thus guaranteeing the monotonically non-increasing convergence characteristic of the iterative optimization process.

[0080] S505: Update the current manpower deployment based on the shift adjustment confirmation data, and refresh the iterative comparison benchmark with the confirmed metric value to obtain iterative deployment data.

[0081] After receiving confirmation of shift changes, the current manpower deployment plan is updated based on this data, and the scheduled shifts of the personnel whose shift changes are confirmed are officially updated to the new shift arrangements. Simultaneously, the system uses the corresponding comprehensive balance metric value after the shift change confirmation. The iterative comparison benchmark is refreshed and used as the new comparison reference for the next shift change operation. The updated manpower deployment plan and its corresponding latest comprehensive equilibrium metric value are recorded as iterative deployment data. The above process S501 to S505 constitutes a complete shift iteration operation. The system sequentially executes the above iterative process for all adjustable personnel, gradually optimizing the manpower deployment plan through strategies, so that the comprehensive equilibrium metric value monotonically decreases or remains unchanged in each effective adjustment, thereby driving the shift scheduling plan to gradually converge towards a better equilibrium state.

[0082] S60: Perform iterative convergence condition determination and termination output processing on the iterative deployment data to obtain balanced manpower deployment data.

[0083] After each iteration, the system determines the convergence condition of the iterative deployment data to decide whether to terminate the iteration process, ultimately obtaining balanced manpower deployment data. This balanced manpower deployment data is the balanced manpower allocation scheme for each time period after joint measurement and iterative optimization. It can effectively improve the fit between the actual manpower and target demand in each time period and the dispersion of shift events on the time axis, thereby ensuring the balance and stability of service capacity throughout the day.

[0084] Combination Figure 8 As shown, step S60 specifically includes: S601: Extract the currently executed cumulative iteration rounds from the iterative deployment data to obtain iteration round data.

[0085] After each iteration, the system extracts the cumulative number of iterations executed from the iteration deployment data. An iteration round refers to the number of times the system completes a full round of person-by-person shift traversal and adjustment operations for all adjustable personnel. The system records this cumulative number of iteration rounds, obtaining the iteration round data.

[0086] S602: Compare the iteration round data with the preset maximum number of cycles to obtain round determination data.

[0087] After obtaining the iteration round data, the current cumulative iteration rounds are compared with the preset maximum number of iterations. The maximum number of iterations refers to the upper limit of the maximum number of iterations allowed in the iterative optimization process, which can be determined through experimental experience in actual scheduling scenarios. The system records the comparison result as round determination data to indicate whether the current iteration round has reached the preset maximum number of iterations.

[0088] S603: Based on the round determination data, extract the comprehensive equilibrium metric change between two adjacent rounds of iteration to obtain metric change data.

[0089] After obtaining the round determination data, if the current iteration round has not yet reached the maximum number of iterations, the system further extracts the change in the comprehensive balance metric value between two adjacent iterations. Specifically, it calculates the absolute value of the difference between the comprehensive balance metric value after the current iteration and the comprehensive balance metric value after the previous iteration, obtaining the metric change data. This metric change data reflects the extent to which the most recent iteration has improved the overall balance of the scheduling scheme.

[0090] S604: Compare the measured change data with a preset convergence threshold to obtain convergence determination data.

[0091] After obtaining the measurement change data, the amount of measurement change is compared with a preset convergence threshold. The convergence threshold is a minimum change threshold set by the system to determine whether the iterative process has converged. This parameter value can also be determined through experimental experience in actual scheduling scenarios. When the measurement change between two adjacent iterations is lower than this convergence threshold, it indicates that further iterations have little effect on improving the balance of the scheduling scheme, and the iterative process can be considered to have reached a convergence state. The system records the comparison result as convergence judgment data.

[0092] S605: When the number of rounds in the convergence determination data reaches the upper limit or the metric change is lower than the threshold, the iteration is terminated and termination signal data is obtained.

[0093] After obtaining the round determination data and convergence determination data, the system combines the above two determination results and triggers iteration termination when either of the following conditions is met: Condition 1 is that the current cumulative number of iterations has reached the preset maximum number of iterations; Condition 2 is that the change in the comprehensive equilibrium metric between two adjacent iterations is lower than the preset convergence threshold. Of these two termination conditions, the former is a hard time constraint on the iteration process, ensuring that the algorithm completes the calculation within a finite time; the latter is a convergence constraint on the iteration process, ensuring that the algorithm stops in time when the balance of the scheduling scheme no longer significantly improves, avoiding invalid and redundant calculations. When either condition is met, the system generates termination signal data.

[0094] S606: Based on the termination signal data, output the final manpower deployment scheme for the current iteration to obtain balanced manpower deployment data.

[0095] Upon receiving the termination signal data, the iterative optimization process terminates, and the final manpower deployment plan obtained in the current iteration is output, resulting in balanced manpower deployment data. This balanced manpower deployment data is the balanced manpower configuration plan for each time period after joint measurement of manpower fitting and shift dispersion, and multiple rounds of iterative optimization. It includes the final confirmed shift for each agent and its corresponding time arrangement. In financial customer service scenarios, this balanced manpower deployment data can be directly distributed to the scheduling management system of the financial institution's customer service center to guide various agents such as credit consultation agents, claims processing agents, and account management agents to execute their work plans according to the optimized shift arrangements. This effectively improves the fit between actual manpower and target demand in each time period and the dispersion of shift events on the time axis, thereby ensuring the service connection rate and customer response quality throughout the day. In healthcare service scenarios, this balanced manpower deployment data can be applied to the manpower scheduling management of hospital outpatient appointment scheduling systems, health management service hotlines, or remote consultation platforms to ensure a balanced allocation of reception manpower in each consultation period, shorten user waiting time, and improve the medical service experience.

[0096] As can be seen, this application initializes the manpower distribution for each time period by matching shift types and screening adjustability in business configuration data. It then uses a joint metric of manpower fit and shift dispersion to comprehensively evaluate the manpower distribution for each time period. Finally, through shift iterative optimization and convergence determination, balanced manpower deployment data is obtained. This application addresses the technical problem of unbalanced manpower deployment in existing scheduling systems, which lack methods for jointly measuring and iteratively optimizing the degree of manpower fit and shift dispersion for each time period after automatic scheduling. By weighted fusion of manpower fit and shift dispersion metrics to form a comprehensive equilibrium metric, and iteratively optimizing adjustable personnel person-by-person and shift-by-shift based on this comprehensive equilibrium metric, the scheduling scheme gradually approaches a better equilibrium state in each effective iteration. This effectively improves the fit between actual manpower and target demand for each time period and the dispersion of shift events on the time axis, thereby ensuring the balance and stability of service capabilities throughout the day. Furthermore, through progressive configuration data reading, time period classification, personnel retrieval and bidirectional association binding, and deduplication verification, the integrity and accuracy of personnel configuration data are ensured. A multi-level, progressive screening strategy accurately identifies personnel eligible for fine-tuning while ensuring that fixed shifts are included in the overall distribution statistics, balancing precise definition of shift adjustment scope with complete distribution statistics. Time discretization is used to simultaneously establish the actual seat distribution, target demand distribution, and shift start event distribution within a unified time-slice framework, providing a structured computational foundation for joint metrics. A piecewise robust metric is employed for the manpower fitting dimension, balancing sensitivity to small errors with robustness to large errors. A normalized chi-square statistical metric is used for the shift dispersion dimension to achieve a rigorous quantitative representation of dispersion. The two are weighted and fused to achieve coordinated optimization of the dual objectives. An iterative strategy ensures the monotonically non-increasing convergence of the comprehensive equilibrium metric, and dual termination conditions—maximum number of iterations and convergence threshold—ensure the algorithm's timeliness and usability. This method can be applied to scheduling scenarios in financial customer service and healthcare services.

[0097] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0098] In one embodiment, a scheduling iteration adjustment device is provided, which corresponds one-to-one with the scheduling iteration adjustment method in the above embodiments. For example... Figure 9 As shown, the shift scheduling iterative adjustment device includes: a data configuration module 100, a shift adjustment module 200, a manpower calculation module 300, a joint measurement module 400, an iterative deployment module 500, and a data output module 600.

[0099] Detailed descriptions of each functional module are as follows: The data configuration module 100 is used to obtain business configuration data for the shift scheduling cycle, perform shift type personnel matching processing on the business configuration data, and establish a correlation mapping between time period, personnel and shift type to obtain personnel configuration data. The shift adjustment module 200 is used to perform available shift and personnel adjustability filtering on the personnel configuration data, so as to remove single shift data and unscheduled data and mark personnel that can be finely adjusted, so as to obtain adjustable shift data. The manpower calculation module 300 is used to perform initial processing of manpower distribution during shift scheduling based on the adjustable shift data, so as to perform time slice discretization and shift time window mapping to obtain manpower distribution data for the time period. The joint measurement module 400 is used to fit the manpower distribution data of the time period, and to count the number of shifts starting based on the fitting results. Based on the number of shifts starting, the module performs joint measurement processing of shift dispersion to obtain comprehensive and balanced measurement data. The iterative deployment module 500 is used to obtain iterative deployment data by changing the currently scheduled shifts and performing shift iteration processing based on the comprehensive balance measurement data. The data output module 600 is used to determine the iterative convergence condition and terminate the output of the iterative deployment data to obtain balanced manpower deployment data.

[0100] In one embodiment, the data configuration module 100 is specifically used for: Obtain the scheduling cycle and time period configuration parameters corresponding to business activities, and integrate them to obtain business configuration data; Based on the business configuration data, the page settings information for each class type is read to obtain the class type settings data; The class type setting data is categorized and divided according to time period intervals to obtain time period class type mapping data; Based on the time period and class type mapping data, the corresponding agent personnel are retrieved and configured to obtain agent personnel retrieval data; The retrieval data of the agents is processed by two-way association binding between agents and shift types to obtain agent shift type association data; The personnel shift type association data is deduplicated and verified according to the shift scheduling cycle, and the personnel configuration data is obtained by summarizing.

[0101] In one embodiment, the shift adjustment module 200 is specifically used for: Initialize all available shifts for each person in the corresponding time period based on the personnel configuration data to obtain the available shift set data; Personnel with only one available shift in the available shift set data are marked and excluded to obtain single-shift exclusion data; The personnel who were not scheduled for work on that day in the single-shift exclusion data were marked and excluded to obtain the on-duty screening data; Determine whether the time period of the shift event of the personnel in the shift screening data falls within the fine-tuning time period, and integrate them to obtain the adjustability determination data; Based on the adjustability determination data, personnel who meet the conditions are marked as fine-tunable personnel, thus obtaining fine-tunable personnel data; The fixed shifts in the adjustable personnel data are included in the shift distribution calculation and summarized to obtain the adjustable shift data.

[0102] In one embodiment, the human resource calculation module 300 is specifically used for: Load the adjustable shift data and initialize the header configuration information of the corresponding shift schedule to obtain the shift schedule header data; The scheduling cycle header data is discretized according to a preset time granularity to divide it into time slice sequences; Traverse the time slice sequence and count the actual number of seats arranged in each time slice to obtain the actual seat distribution data; Based on the actual seat distribution data, the target manpower demand for each time slot is loaded to obtain the target demand distribution data; Each shift in the target demand distribution data is mapped to a time axis with a fixed time window length and the start time is marked to obtain shift event distribution data; Based on the shift event distribution data, the number of shifts starting at each shift adjustment time point is counted to obtain the manpower distribution data for the time period.

[0103] In one embodiment, the joint metric module 400 is specifically used for: The relative deviation of manpower is calculated based on the time slices in the manpower distribution data for the aforementioned time period, and the relative deviation data is obtained. The relative deviation data is partitioned according to the tolerance threshold and subjected to in-threshold squared measurement and out-of-threshold linear extension processing to obtain robust deviation data; The robust deviation data is aggregated with the mean over all time periods and truncated to the upper limit according to a preset empirical threshold to obtain human fitting measurement data. Based on the time slice framework of the aforementioned human resource fitting measurement data, the number of shifts starting at each shift adjustment time point is counted to obtain shift count data. The shift count data is normalized to a chi-square statistical measure to obtain shift dispersion data. The shift dispersion measurement data and the manpower fitting measurement data are weighted and fused according to preset weight coefficients to obtain comprehensive and balanced measurement data.

[0104] In one embodiment, the iterative deployment module 500 is specifically used for: The current total loss value of the comprehensive equilibrium metric data is used as the iterative comparison benchmark to obtain the benchmark loss data; Modify the currently scheduled shifts to obtain trial shift data; The comprehensive balance metric after the shift change is calculated using the trial shift data, and the comprehensive balance metric is compared with the baseline loss data to obtain shift change judgment data. Based on the shift adjustment judgment data, shift adjustment is confirmed when the loss value does not increase, thus obtaining shift adjustment confirmation data; The current manpower deployment is updated based on the shift adjustment confirmation data, and the iterative comparison benchmark is refreshed with the confirmed metric value to obtain iterative deployment data.

[0105] In one embodiment, the data output module 600 is specifically used for: Extract the cumulative number of iterations that have been executed so far from the iterative deployment data to obtain iteration round data; The iteration round data is compared with the preset maximum number of iterations to determine the round determination data. Based on the round determination data, the comprehensive equilibrium metric change between two adjacent rounds of iteration is extracted to obtain the metric change data; The measured change data is compared with a preset convergence threshold to determine convergence determination data. When the number of rounds in the convergence determination data reaches the upper limit or the metric change is lower than the threshold, the iteration is terminated, and termination signal data is obtained. Based on the termination signal data, the final manpower deployment scheme for the current iteration is output, and balanced manpower deployment data is obtained.

[0106] Specific limitations regarding the scheduling iteration adjustment device can be found in the limitations of the scheduling iteration adjustment method described above, and will not be repeated here. Each module in the aforementioned scheduling iteration adjustment device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0107] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a scheduling iterative adjustment method on the server side.

[0108] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 11 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When executed by the processor, the computer program implements the functions or steps of a scheduling iterative adjustment method on the client side.

[0109] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed, can perform the steps provided in the above embodiments.

[0110] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0111] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0112] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0113] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0114] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A scheduling iterative adjustment method, characterized in that, include: Obtain the business configuration data of the shift scheduling cycle, and perform shift type personnel matching processing on the business configuration data to establish the association mapping of time period-person-shift type to obtain personnel configuration data; The personnel configuration data is filtered for available shifts and personnel adjustability to remove single-shift data and unscheduled data and mark personnel who can be finely adjusted, thus obtaining adjustable shift data. Based on the adjustable shift data, the manpower distribution of the shift scheduling period is initialized to perform time slice discretization and shift time window mapping to obtain the manpower distribution data of the time period. The manpower distribution data for the time period is fitted, and the number of shifts starting is counted based on the fitting results. The shift dispersion joint measurement is performed based on the number of shifts starting to obtain comprehensive and balanced measurement data. Based on the comprehensive and balanced measurement data, iterative deployment data is obtained by changing the currently scheduled shifts and performing shift iteration processing. The iterative deployment data is subjected to iterative convergence condition determination and termination output processing to obtain balanced manpower deployment data.

2. The scheduling iteration adjustment method according to claim 1, characterized in that, The process of obtaining the business configuration data for the shift scheduling cycle, and performing shift type personnel matching processing on the business configuration data to establish a time period-person-shift type association mapping, yields personnel configuration data, including: Obtain the scheduling cycle and time period configuration parameters corresponding to business activities, and integrate them to obtain business configuration data; Based on the business configuration data, the page settings information for each class type is read to obtain the class type settings data; The class type setting data is categorized and divided according to time period intervals to obtain time period class type mapping data; Based on the time period and class type mapping data, the corresponding agent personnel are retrieved and configured to obtain agent personnel retrieval data; The retrieval data of the agents is processed by two-way association binding between agents and shift types to obtain agent shift type association data; The personnel shift type association data is deduplicated and verified according to the shift scheduling cycle, and the personnel configuration data is obtained by summarizing.

3. The scheduling iteration adjustment method according to claim 1, characterized in that, The process of filtering the personnel configuration data based on available shifts and personnel adjustability, to remove single-shift data and unscheduled data and mark personnel who can be fine-tuned, yields adjustable shift data, including: Initialize all available shifts for each person in the corresponding time period based on the personnel configuration data to obtain the available shift set data; Personnel with only one available shift in the available shift set data are marked and excluded to obtain single-shift exclusion data; The personnel who were not scheduled for work on that day in the single-shift exclusion data were marked and excluded to obtain the on-duty screening data; Determine whether the time period of the shift event of the personnel in the shift screening data falls within the fine-tuning time period, and integrate them to obtain the adjustability determination data; Based on the adjustability determination data, personnel who meet the conditions are marked as fine-tunable personnel, thus obtaining fine-tunable personnel data; The fixed shifts in the adjustable personnel data are included in the shift distribution calculation and summarized to obtain the adjustable shift data.

4. The scheduling iteration adjustment method according to claim 1, characterized in that, The initialization process for manpower distribution during shift scheduling based on the adjustable shift data, including time slice discretization and shift time window mapping, yields manpower distribution data for each shift, including: Load the adjustable shift data and initialize the header configuration information of the corresponding shift schedule to obtain the shift schedule header data; The scheduling cycle header data is discretized according to a preset time granularity to divide it into time slice sequences; Traverse the time slice sequence and count the actual number of seats arranged in each time slice to obtain the actual seat distribution data; Based on the actual seat distribution data, the target manpower demand for each time slot is loaded to obtain the target demand distribution data; Each shift in the target demand distribution data is mapped to a time axis with a fixed time window length and the start time is marked to obtain shift event distribution data; Based on the shift event distribution data, the number of shifts starting at each shift adjustment time point is counted to obtain the manpower distribution data for the time period.

5. The scheduling iteration adjustment method according to claim 1, characterized in that, The process of fitting the manpower distribution data for the specified time period, calculating the number of shift starts based on the fitting results, and performing a combined shift dispersion measurement based on the number of shift starts to obtain comprehensive and balanced measurement data includes: The relative deviation of manpower is calculated based on the time slices in the manpower distribution data for the aforementioned time period, and the relative deviation data is obtained. The relative deviation data is partitioned according to the tolerance threshold and subjected to in-threshold squared measurement and out-of-threshold linear extension processing to obtain robust deviation data; The robust deviation data is aggregated with the mean over all time periods and truncated to the upper limit according to a preset empirical threshold to obtain human fitting measurement data. Based on the time slice framework of the aforementioned human resource fitting measurement data, the number of shifts starting at each shift adjustment time point is counted to obtain shift count data. The shift count data is normalized to a chi-square statistical measure to obtain shift dispersion data. The shift dispersion measurement data and the manpower fitting measurement data are weighted and fused according to preset weight coefficients to obtain comprehensive and balanced measurement data.

6. The scheduling iteration adjustment method according to claim 1, characterized in that, Based on the comprehensive and balanced measurement data, iterative deployment data is obtained by changing the currently scheduled shifts and performing shift iteration processing, including: The current total loss value of the comprehensive equilibrium metric data is used as the iterative comparison benchmark to obtain the benchmark loss data; Modify the currently scheduled shifts to obtain trial shift data; The comprehensive balance metric after the shift change is calculated using the trial shift data, and the comprehensive balance metric is compared with the baseline loss data to obtain shift change judgment data. Based on the shift adjustment judgment data, shift adjustment is confirmed when the loss value does not increase, thus obtaining shift adjustment confirmation data; The current manpower deployment is updated based on the shift adjustment confirmation data, and the iterative comparison benchmark is refreshed with the confirmed metric value to obtain iterative deployment data.

7. The scheduling iteration adjustment method according to claim 1, characterized in that, The process of determining the convergence condition and terminating the output of the iterative deployment data to obtain balanced manpower deployment data includes: Extract the cumulative number of iterations that have been executed so far from the iterative deployment data to obtain iteration round data; The iteration round data is compared with the preset maximum number of iterations to determine the round determination data. Based on the round determination data, the comprehensive equilibrium metric change between two adjacent rounds of iteration is extracted to obtain the metric change data; The measured change data is compared with a preset convergence threshold to determine convergence determination data. When the number of rounds in the convergence determination data reaches the upper limit or the metric change is lower than the threshold, the iteration is terminated, and termination signal data is obtained. Based on the termination signal data, the final manpower deployment scheme for the current iteration is output, and balanced manpower deployment data is obtained.

8. A scheduling iteration adjustment device, characterized in that, include: The data configuration module is used to obtain business configuration data for the shift scheduling cycle, perform shift type personnel matching processing on the business configuration data, so as to establish a correlation mapping between time period, personnel and shift type to obtain personnel configuration data; The shift adjustment module is used to filter the personnel configuration data for available shifts and personnel adjustability, so as to remove single-shift data and unscheduled data and mark personnel that can be finely adjusted, so as to obtain adjustable shift data. The manpower calculation module is used to perform initial processing of manpower distribution during shift scheduling based on the adjustable shift data, so as to perform time slice discretization and shift time window mapping to obtain manpower distribution data for the time period. The joint measurement module is used to fit the manpower distribution data of the time period, and to count the number of shifts starting based on the fitting results. Based on the number of shifts starting, the module performs joint measurement processing on shift dispersion to obtain comprehensive and balanced measurement data. The iterative deployment module is used to obtain iterative deployment data by changing the currently scheduled shifts and performing shift iteration processing based on the comprehensive balance measurement data; The data output module is used to determine the iterative convergence condition and terminate the output of the iterative deployment data to obtain balanced manpower deployment data.

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

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the scheduling iterative adjustment method as described in any one of claims 1 to 7.

Images