Floating matching model for personnel compilation matching degree evaluation and method thereof

Abstract

The invention provides a floating matching model for personnel compilation matching degree evaluation. The floating matching model comprises a multi-index comprehensive evaluation sub-model or a high-dimensional vector evaluation sub-model, a multi-index step-by-step synthesis evaluation sub-model or a high-dimensional vector step-by-step synthesis evaluation sub-model, a personnel compilation matching degree dynamic correction sub-model and a floating matching degree sub-model, the multi-index comprehensive evaluation sub-model is used for solving a D-level unit personnel compilation comprehensive matching degree; the high-dimensional vector evaluation sub-model is based on a feature value set of each factor, and a compilation matching degree is solved through cosine similarity among vectors; according to the personnel compilation matching degree dynamic correction sub-model, the matching degree of the future year is used as one of adjustment factors of the compilation matching degree of the current year, and the original model is corrected; and the floating matching degree sub-model. According to another aspect of the invention, a method of using a floating matching model is provided.

Description

Technical Field

[0001] This invention relates to a data processing system and method for administrative and management purposes, and more specifically, to a floating matching model and method for evaluating the matching degree of personnel staffing. Background Technology

[0002] In modern administration, the precise and efficient allocation of human resources to ensure optimal talent utilization and maximize effectiveness is a crucial component of administrative data processing systems. Personnel staffing and current matching evaluation is a vital tool for all levels of organizations to accurately assess the scientific nature of their human resource allocation, thereby enabling them to develop and revise personnel development paths accordingly. This is of paramount importance for optimizing human resource allocation, maximizing human resource effectiveness, and accelerating productivity generation.

[0003] Staffing levels describe the distribution of personnel positions, numbers, ranks, and specialties within an organization. However, in actual operation, discrepancies often arise between the actual personnel composition and staffing requirements. To measure the staffing-to-position ratio in a given organization, the actual number of personnel is typically divided by the authorized number of positions, resulting in a full-staffing rate as an evaluation indicator. However, this method only reflects the overall surplus or shortage of personnel and fails to reveal structural differences in quantity, rank, and specialties at different levels, thus lacking a comprehensive and accurate measure of staffing-to-position ratio matching. To achieve staffing-to-position ratio matching, promptly and accurately grasp the overall characteristics of personnel, and maximize the efficiency of human resources, this invention takes a hypothetical organization as an example. By analyzing detailed staffing-to-position data, a mathematical model is established to comprehensively evaluate the staffing-to-position ratio matching of the organization's various levels of workforce structure in terms of personnel number, rank, and specialties. It also predicts future development trends and proposes rational measures and suggestions. Summary of the Invention

[0004] Based on the above-mentioned technical problems, this invention proposes a floating matching model and method for evaluating the matching degree of personnel staffing.

[0005] According to one aspect of the present invention, a floating matching model for evaluating personnel staffing matching degree is provided, comprising: a multi-index comprehensive evaluation sub-model or a high-dimensional vector evaluation sub-model, a multi-index hierarchical synthesis evaluation sub-model or a high-dimensional vector hierarchical synthesis evaluation sub-model, a dynamic correction sub-model for personnel staffing matching degree, and a floating matching degree sub-model; wherein, the multi-index comprehensive evaluation sub-model is used to solve the comprehensive matching degree of personnel staffing matching degree in level D units, and includes three secondary sub-objectives: quantity matching degree, professional matching degree, and grade matching degree; the high-dimensional vector evaluation sub-model divides the personnel quantity, grade, professional, and position factors of level D units into... The differences in features are transformed into feature sets of various factors. The feature values ​​of each factor for the current staffing position are combined and represented by a high-dimensional vector. The cosine similarity between the current staffing position vectors is compared to solve for the similarity between the current staffing data vector and the existing data vector. The multi-indicator hierarchical synthesis evaluation sub-model solves for the staffing matching degree of personnel in higher-level units of the D-level unit. It receives the comprehensive staffing matching degree of personnel in the D-level unit solved by the multi-indicator comprehensive evaluation sub-model and performs a weighted average according to the actual number of personnel in each lower-level unit to calculate the staffing matching degree of personnel in higher-level units. The high-dimensional vector hierarchical synthesis evaluation... The sub-model groups the evaluation values ​​of all D-level units according to their corresponding C-level units, and then calculates the average to obtain the evaluation value of each C-level unit. Similarly, it groups the B-level units corresponding to C-level units and calculates the average to obtain the composite evaluation value of each B-level unit, and finally obtains the composite evaluation value of the A-level units. Finally, it synthesizes the evaluation results of all units in A, B, and C levels step by step. The personnel staffing matching degree dynamic correction sub-model considers the changes in the actual number of personnel in future years. Based on the personnel's current level, age, years of service, and resignation probability, it adjusts the multi-indicator comprehensive evaluation sub-model or high-dimensional vector evaluation sub-model step by step. The matching degree of personnel establishment and current status is evaluated using the synthetic evaluation sub-model or the high-dimensional vector stepwise synthetic evaluation sub-model. The matching degree of future years is used as one of the adjustment factors for the matching degree of the current year to correct the sub-model and dynamically correct the matching degree of personnel establishment and current status. The floating matching degree sub-model dynamically adjusts the matching degree of personnel establishment and current status of any sub-model of the multi-index comprehensive evaluation sub-model or the high-dimensional vector evaluation sub-model, the multi-index stepwise synthetic evaluation sub-model or the high-dimensional vector stepwise synthetic evaluation sub-model and / or the dynamic correction sub-model of personnel establishment and current status based on the difference between the number of personnel in the unit and the number of established personnel, and the complementarity of personnel at adjacent levels.

[0006] According to another aspect of the present invention, a method using a floating matching model is provided, comprising: S1: setting initial conditions for the floating matching model, the initial conditions including: the staffing requirement of a given unit will not change; adjustments to the unit structure and personnel transfers are not considered; S2: performing data processing on the multi-index comprehensive evaluation sub-model, including: determining the weights of sub-objectives; processing staffing data; and converting grade to seniority; S3: performing data processing on the high-dimensional vector evaluation sub-model, including: forming a data vector based on the current number of personnel to obtain the staffing data of current members; establishing a staffing data sampling algorithm to sample the existing staffing data to obtain a staffing dataset for a given D-level unit; and calculating the similarity between any vector in the dataset based on the staffing dataset and the existing staffing dataset to obtain the similarity between two personnel feature vectors. S4: Calculate the matching degree of a given D-level unit by averaging the similarity of all feature vectors; S5: Solve the personnel staffing matching degree of the superior unit using a multi-index hierarchical synthesis evaluation sub-model and / or a detailed staffing data model; S6: Perform data analysis and processing on the dynamic personnel staffing matching correction model, including: calculating the minimum and maximum years of service and the maximum age for each level; calculating the overall promotion probability, the probability of exceeding the years of service and the probability of retiring due to age, and the probability of mid-term transfer for each level; completing the age and years of service data of the actual personnel; taking the 1A large unit as the whole as the analysis object, starting from the D-level unit, aggregate level by level to obtain the personnel staffing matching degree of the final top-level unit 1A for each year, and obtain the personnel staffing matching degree of the current year after correction; S7: Based on the fluctuation of the actual number of personnel, recalculate the matching degree using the floating matching degree sub-model, and use the floating matching degree model to meet the needs of the target unit's normal operation. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of a multi-indicator comprehensive evaluation model; Figure 2 Compile a comprehensive matching degree distribution chart for all personnel in Level D units; Figure 3 A distribution chart of the matching degree of all D-level units; Figure 4 A distribution chart of professional matching degree for all D-level units; Figure 5 A distribution map of the matching degree of all D-level units; Figure 6 A comprehensive matching degree distribution map is compiled for all D-level unit personnel based on the high-dimensional vector evaluation model; Figure 7 This is a comparative analysis chart of the two evaluation schemes; Figure 8 A comparative evaluation chart showing the step-by-step composite evaluation and direct evaluation for C-level units; Figure 9A comparative evaluation chart showing the step-by-step composite evaluation and direct evaluation for Class B units; Figure 10 A comparative evaluation chart showing the step-by-step composite evaluation and direct evaluation for Level A units; Figure 11 The process of establishing a dynamic correction model for personnel matching degree; Figure 12 A logic diagram illustrating changes in personnel status; Figure 13 For simulation and prediction models of personnel status changes; Figure 14 To compare the dynamic correction model with the original model for comprehensive evaluation of multiple indicators; Figure 15 To comprehensively evaluate the dynamic development trend of personnel matching over the next 5 years using multiple indicators; Figure 16 To compare the matching degree distribution between the dynamic correction model and the original model for multi-indicator comprehensive evaluation; Figure 17 A comparison of the matching degree distribution between the high-dimensional vector evaluation dynamic correction model and the original model; Figure 18 To evaluate the dynamic development trend of personnel matching degree over the next 5 years using high-dimensional vectors; Figure 19 To evaluate the dynamic development trend of personnel matching degree of all C-level units over the next 5 years using high-dimensional vectors; Figure 20 A comparison of the matching degree distribution between the high-dimensional vector evaluation dynamic correction model and the original model; Figure 21 To comprehensively evaluate the impact of personnel fluctuation ratios using multiple indicators; Figure 22 The impact of personnel fluctuation ratios on high-dimensional vector evaluation; Figure 23 The impact of the fluctuation ratio of the multi-indicator comprehensive evaluation level; Figure 24 The impact of the evaluation level fluctuation ratio on high-dimensional vectors.

[0008] The following are the symbols used in this invention and their descriptions.

[0009] Detailed Implementation

[0010] The problem of evaluating the matching degree of staffing and current status essentially involves establishing an evaluation model that comprehensively reflects the matching degree of staff in terms of quantity, level, and profession through a combination of quantitative and qualitative evaluation methods. Simultaneously, this model can predict the future trend of the matching degree of staffing and current status at various levels within the organization based on empirical data on dynamic changes in personnel. When establishing the evaluation model, the following assumptions must be made: Assumption 1: The organization's staffing requirements will not change. Assumption 2: Adjustments to the organization's structure and personnel transfers are not considered.

[0011] The degree of staffing-position matching in an organization is influenced by multiple factors, including the number of personnel (N), rank (L), and profession (M). These factors are interconnected, and analyzing only one aspect only provides a partial understanding of the organization's staffing-position matching from three different perspectives, failing to accurately describe the overall matching situation of the organization. This invention employs two schemes to evaluate the overall staffing-position matching degree of all D-level units within Organization A. Scheme 1 uses a multi-index comprehensive evaluation model to calculate and analyze the staffing-position matching degree of each D-level unit in terms of personnel number, rank, and profession, and then combines the weights to determine the overall staffing-position matching degree of the organization. Scheme 2 uses a high-dimensional vector evaluation model, transforming the organization's personnel positions, number, rank, and profession factors into high-dimensional vectors, thus transforming the problem of overall staffing-position matching degree into solving the similarity problem between high-dimensional vectors.

[0012] 1. Multi-indicator comprehensive evaluation model

[0013] 1.1 Establishment of a Multi-Indicator Comprehensive Evaluation Model

[0014] The model aims to solve for the overall matching degree of personnel staffing in Level D units. It can be decomposed into quantitative matching degree. Professional matching degree Level matching degree Three secondary sub-objectives. By analyzing and processing the compiled data and actual data, corresponding algorithms were designed to solve each of the three sub-objectives, yielding the degree of quantitative matching. Professional matching degree Level matching degree Then, by combining their respective weights, the overall matching degree of personnel staffing for this Class D unit is calculated. .

[0015] (1) Overall matching degree of personnel staffing in this Class D unit The calculation formula is as follows: (1) In the formula, The weight of the quantity matching degree, Weighting for professional matching degree The weight of the matching degree.

[0016] (2) Quantity matching degree The calculation method is as follows: (2) In the formula, This indicates the number of staff authorized for this Class D unit. This indicates the actual number of personnel in the D-level unit.

[0017] (3) Professional matching degree The calculation method is as follows: (3) In the formula, For the number of majors, , and The first and second data in the compiled and actual data respectively The number of people in each major.

[0018] (4) Level matching degree The calculation method is as follows: (4) In the formula, For the number of levels, and The first and second data are respectively compiled data and actual data. The number of personnel at each level Indicates the first For ease of calculation, the average years of service corresponding to each level are converted to general years of service, as shown in Table 1.

[0019] 1.2 Data Analysis and Processing

[0020] (1) Determination of sub-objective weights

[0021] Want to use quantity matching degree Professional matching degree Level matching degree Three secondary sub-objectives are used to solve for the personnel staffing matching degree of level D units. When determining the weights of the three secondary indicators, it is essential to clearly define these weights. These weights must objectively reflect the contribution of each secondary indicator to the overall indicator and should not change with variations in units or time. This article suggests that these weights be jointly determined by experts in the field of human resources, when assessing the staffing level. Before the evaluation, the expert panel scores the overall staffing matching degree of the unit based on factors such as the number of personnel, their level, and their profession. The weight of each individual matching degree is determined using the analytic hierarchy process (AHP), or alternatively, the entropy weight evaluation method or fuzzy comprehensive evaluation method. This paper will not discuss this aspect in detail. The weight of the number of personnel is determined by consulting the unit's human resources department. 45%, professional weighting 35%, grade weight The percentage is 20%. This setting is also quite consistent with the actual situation: when examining whether the human resource allocation of a unit is reasonable, the first thing to consider is whether the number of personnel is sufficient to ensure that the unit's basic tasks can be completed; secondly, the professional skills are considered to see if the requirements of making the best use of resources and talents can be met, and to maximize the advantages of professional skills; only then is it considered whether the personnel level matches the job requirements.

[0022] (2) Processing of current data

[0023] First, duplicate data is checked for both compiled data and actual personnel data, and the number of personnel with the same position, the same specialty, and the same grade in the same D-level unit is merged.

[0024] Secondly, when calculating the matching degree of quantity, profession, and level, the total number of personnel, the number of professionals, and the number of personnel at each level of each D-level unit are summarized separately.

[0025] (3) Rules for converting grades to years of service

[0026] To facilitate quantitative analysis, this paper performs appropriate transformation on the grade data. Since grades are closely related to years of service, and using years of service is more effective in reflecting the differences between different grades than using grade values ​​directly, the grade indicators are converted into average years of service, as shown in Table 1.

[0027] Table 1 Grade and Seniority Conversion Table

[0028] 1.3 Staffing and Current Matching Degree of Level D Units Calculation and Solution

[0029] Taking unit 1A-1B-1C-1D as an example, its staffing data and actual personnel data are shown in Table 2.

[0030] Table 2. Staffing Data and Actual Personnel Data for Units 1A-1B-1C-1D

[0031] Based on the multi-index comprehensive evaluation model, the matching degree of the unit quantity is obtained respectively. Professional matching degree Level matching degree Matching degree with staffing As shown in Table 3.

[0032] Table 3 Summary of Calculation Results of Staffing Matching Degree for Units 1A-1B-1C-1D

[0033] 1.4 Evaluation Results and Analysis of the Matching Degree of Staffing Level D Personnel

[0034] (1) Overall matching degree of personnel staffing in all D-level units Distribution

[0035] Based on the calculation results, overall, among all D-level units, the personnel staffing and overall matching degree is... Those with a success rate above 90% accounted for 79.5%, those in the 80%-90% range accounted for 16.7%, and those below 80% accounted for only 3.8%. Figure 2 As shown.

[0036] (2) Quantity matching degree Distribution

[0037] From the perspective of quantitative matching results, quantitative matching degree Those with a percentage above 90% accounted for 87.1%, those between 80% and 90% accounted for 9.8%, and those below 80% accounted for 3.1%. Figure 3 As shown.

[0038] (3) Professional matching degree Distribution

[0039] From the perspective of professional matching results, professional matching... Those with a success rate above 90% accounted for 50%, those between 80% and 90% accounted for 34.1%, and those below 80% accounted for 15.9%. Figure 4 As shown.

[0040] (4) Level matching degree Distribution

[0041] From the perspective of the level matching results, the level matching degree Those with a score of 90% or higher accounted for 79.5%, those in the 80%-90% range accounted for 18.2%, and those below 80% accounted for 2.3%. Figure 5 As shown.

[0042] (5) Comprehensive analysis

[0043] Based on a comprehensive comparison of the above four sets of data, the following analysis and judgment can be made: ① The data results from the multi-indicator comprehensive evaluation model in Scheme 1 show that the overall staffing and staffing matching of all D-level units of Unit A is relatively good.

[0044] ② The overall matching degree of personnel quantity is relatively high. Due to the lack of a scientific mathematical model, the most common and simplest way to improve the matching degree of personnel staffing in a unit is to measure the unit's full staffing rate. Therefore, adjusting the number of personnel is the primary choice for many units, resulting in a relatively high overall matching degree of personnel quantity.

[0045] ③ The degree of professional matching is relatively low, and there are significant differences between different units. Firstly, the wide variety of professional fields greatly increases the complexity of matching majors with positions. Secondly, personnel promotion is usually influenced by factors such as years of service, work experience, and training, and professional matching is usually not a primary consideration. Thirdly, personnel mobility is poor; personnel transfers are usually only possible through open selection, and the higher the level of the unit, the more difficult the transfer becomes. It is difficult to optimize the organization for large-scale reassignment based on professional matching, resulting in a large number of personnel working in positions outside their professional field.

[0046] ④ The overall staffing matching degree derived from this model is more meaningful than the full staffing rate. First, this model comprehensively considers factors such as the number of personnel, their rank, and their profession, providing a more objective reflection of the current staffing situation in a given unit. Second, the overall staffing matching degree is influenced by data from three sub-objectives, effectively eliminating the impact of individual exceptions on the overall result. Third, the model algorithm references current human resources regulations, thus the data results have strong guiding and practical significance.

[0047] 2 High-dimensional vector evaluation model

[0048] The factors of personnel quantity, grade, major, and position in Unit A are transformed into feature value sets for each factor according to their different characteristics. The feature values ​​of each factor of the current position are combined and represented by a high-dimensional vector. By comparing the cosine similarity between the current position vectors (the Cos value of the angle between the vectors represents the similarity of the vectors), the problem of matching the current position is transformed into solving the similarity problem between the current position data vector and the existing data vector.

[0049] 2.1 Establishment of High-Dimensional Vector Evaluation Model

[0050] (1) Vectorization of data

[0051] To transform the data of the four factors—current job position, grade, major category, and sub-major category—into vector representations, we performed feature extraction on the data of each factor: For unit numbering, use For example, the unit 1A-1B-1C-1D is represented as .

[0052] For the job factor P, the number of jobs within the evaluated unit is counted and denoted as Pjob. If there are [number], then set [number] 0-1 vectors of dimension For example, in the first position, the first digit is 1, and the rest are 0, denoted as . .

[0053] For the grade factor L, the numbers corresponding to each grade are used as input values. That is, the 11 levels S01-S11 represent the feature vector values ​​of each level, from 1 to 11.

[0054] For the professional factor M, the number of professional categories within the evaluated unit is counted and denoted as M. ,set up 0-1 vectors of dimension For example, in the first major category, the first digit is 1, and the rest are 0, denoted as... Next, count the number S of subcategories within each major category, and select the one with the largest number of subcategories, denoted as [Sum of subcategories]. ,set up 0-1 vectors of dimension For each of their respective professional subcategories, the first one is 1, and the rest are 0, denoted as... .

[0055] The feature vector of each person in the process is: (5) The set of vectors of all compilers .

[0056] The feature vector for each person in the existing data is as follows: (6) The existing set of personnel vectors .

[0057] (2) Vector weighting processing

[0058] In the vector, the importance of the four factors—job title, grade, major category, and subcategory—has varying degrees of impact on the similarity calculation result. Therefore, the weight of each factor must be considered during the vector calculation process. Let's assume the relative importance of job title, grade, major category, and subcategory are respectively... The combined vector is: (7) The weighted vector is:

[0059] in, For the identity matrix, others , , It is the identity matrix of the corresponding dimension. (8) Unless otherwise specified, the personnel feature vectors mentioned below refer to the weighted vectors.

[0060] 2.2 Data Processing

[0061] (1) Tabular data processing

[0062] Our vectors describe attributes at the individual level. To achieve this, we adopted a method of forming data vectors based on the number of people. That is, when there are multiple people with completely identical job titles, levels, and specialties, we extract multiple identical feature vectors based on the number of people. For all vectors in the compiled data, we compare them pairwise one by one to avoid ignoring identical data during the matching process.

[0063] (2) Sampling of current data

[0064] For a given unit, corresponding data is extracted from the total dataset. For example, to calculate the matching degree of personnel in units 1A-2B-3C-4D, entries with values ​​1, 2, 3, and 4 for A, B, C, and D are retrieved from the total dataset and combined into a new dataset. The sampling algorithm is shown in Algorithm 1. This algorithm extracts the feature vectors of all personnel in units 1A-2B-3C-4D from the personnel data. The same sampling algorithm can be used to process the actual personnel dataset Y.

[0065]

[0066] The sampling algorithm iterates through and samples each D-level unit, then performs calculations to obtain the evaluation values ​​for all D-level units.

[0067] 2.3 Vector Similarity Calculation

[0068] After selecting the dataset, it's necessary to calculate the vectors within the dataset. Choosing an appropriate distance calculation method is crucial for this calculation. Distance (difference level) and similarity (similarity level) methods can be viewed as calculating the distance between elements using a certain distance function. These methods are fundamental concepts in machine learning and are widely used in similarity algorithms. Commonly used similarity calculation methods include cosine similarity, covariance, Pearson correlation coefficient, and chi-square test. Cosine similarity uses the cosine value of the angle between vectors as the vector similarity, with a value range of 0 to 1, where 1 indicates perfect correlation and 0 indicates independence. Cosine similarity is independent of vector length, only related to vector direction. Among various distance methods, we choose cosine distance to characterize the similarity between vectors, as its value range of 0-1 meets the requirements and facilitates subsequent calculation of the matching degree.

[0069] For both compiled data and actual personnel data, the vector similarity is calculated by comparing each item. Let the compiled data vector set be denoted as . The actual data vector set is First, based on the positions responsible for compiling the data, for A vector in First of all Search results for related The vector corresponding to the position Calculate separately With all Angle between vectors value.

[0070] According to the vector , dot product formula (9) (10) The similarity between vectors is calculated using the cosine value of the angle between them.

[0071] In this problem, for major categories and subcategories, subcategories are only considered if the major categories are the same; otherwise, they are not considered. Therefore, the similarity between the feature vectors of two individuals is: (11) but The similarity is ,Will , Vector from , Delete it. Then remove it from... Select the next vector Perform the same operation. For the resulting extra vectors, consider them all as mismatched and record their similarity as 0.

[0072] The average of all similarities is: , Matching degree (12)

[0073] 2.4 Model Calculation and Solution

[0074] Based on the aforementioned high-dimensional vector evaluation method, and considering the importance of position, grade, major professional category, and sub-major professional category, the staffing-position matching degree of all 132 D-level units was calculated. The distribution of staffing-position matching degree for each D-level unit is as follows: Figure 6 As shown.

[0075] Unlike multi-indicator comprehensive evaluation methods, high-dimensional vector evaluation methods reflect the correlation between multiple factors such as staffing quantity, position, grade, and profession on the matching degree. In the evaluation process, it transforms the quantitative factor of staffing into factor items based on the characteristic vector of individual positions, increasing its reliability in calculating the overall staffing matching degree.

[0076] Figure 7 This chart compares and analyzes two evaluation schemes for D-level units: the high-dimensional vector evaluation method and the multi-index comprehensive evaluation method. By comparing the comprehensive matching evaluation results of all D-level units using the multi-index comprehensive evaluation method and the high-dimensional vector evaluation method, we can see that the calculation results of the two methods are basically consistent, indicating that both methods are effective and reliable.

[0077] Looking at the overall distribution of the calculation results, the proportion of Grade D units with a staffing-to-current-employment matching rate of less than 80% is relatively small, less than 4%, while the proportion of Grade D units with a matching rate of over 80% reaches 96%. Overall, the staffing-to-current-employment matching rate of Grade D units is relatively high. The staffing-to-current-employment matching rate reflects the general effectiveness of human resource regulation over the years of reform and restructuring. However, it cannot be denied that 4% of units still have insufficient staffing-to-current-employment matching rate, indicating that there is still room for improvement in human resource work.

[0078] Because Level D units have a smaller number of staff, the computational workload for matching staffing levels within a single Level D unit is relatively small. However, if calculations are performed on higher-level units like Level A units, the computational workload of using a multi-indicator comprehensive evaluation model to evaluate detailed staffing data may be too large. Therefore, we adopted two schemes to establish a hierarchical composite evaluation model, enabling higher-level units to directly use various indicator data from lower-level units to evaluate the staffing matching degree of their own level. Scheme 1 is based on a multi-indicator comprehensive evaluation model framework, and Scheme 2 is based on a high-dimensional vector evaluation model framework.

[0079] 3. Multi-index stepwise composite evaluation model

[0080] 3.1 Establishment of a multi-index hierarchical composite evaluation model

[0081] The goal of the model is to solve for the matching degree of personnel staffing at higher-level units within Unit A. This model primarily focuses on the matching degree of personnel allocation in the next-level units of this organization. The staffing matching degree of higher-level units is calculated by weighting the actual number of personnel in each lower-level unit. .

[0082] The personnel staffing matching degree of the superior unit The calculation formula is as follows: (13)

[0083] 3.2 Using a multi-index hierarchical composite evaluation model to solve the personnel staffing matching degree of the superior unit.

[0084] Taking Unit 1A-2B-3C as an example, the personnel staffing matching degree and the actual number of personnel in the unit are calculated using a multi-index comprehensive evaluation model for its four subordinate D-level units, as shown in Table 4.

[0085] Table 4. Data Table of 4 Class D Units under 1A-2B-3C

[0086] The staffing matching degree of units 1A-2B-3C using the hierarchical composite evaluation model can be calculated to be 90.27%. Based on this model algorithm, we calculated the staffing matching degree of all A, B, and C level units of Unit A, as shown in Table 5.

[0087] Table 5. Level A, B, and C Units Summary Table

[0088] 3.3 Using a detailed data model to solve for the matching degree of personnel records in higher-level units

[0089] For ease of comparison, we used a multi-index comprehensive evaluation model to calculate the staffing matching degree of 1 A-level unit, 3 B-level units and 1 C-level unit of Unit A, as shown in Table 6.

[0090] Table 6. Partial List of Units Summary Table

[0091] 3.4 Comparative Analysis of the Step-by-Step Synthesis Evaluation Model and the Model Using Detailed Compilation Data

[0092] We will determine the personnel staffing matching degree of five units: 1A, 1A-1B, 1A-2B, 1A-3B, and 1A-1B-1C. and The comparisons are shown in Table 7.

[0093] Table 7 Partial Units Summary Table

[0094] By comparing the above five sets of data, the following conclusions can be drawn:

[0095] (1) The calculation results of the hierarchical synthesis model are always less than those of the detailed data model. The reason is that when using the hierarchical synthesis model, a weighted average algorithm is introduced, which actually preserves the differences in the data matching degree of each unit; while when using the detailed data model, it is equivalent to treating a certain superior unit as the lowest level D unit, which is equivalent to ignoring the data matching degree of all subordinate units and reshuffling the personnel distribution. Therefore, the calculation results of the hierarchical synthesis model are always less than those of the detailed data model.

[0096] (2) The computational cost of the hierarchical synthesis model is much smaller than that of the detailed data compilation model. By using the hierarchical synthesis model, the various indicator data of the lower-level units can be fully utilized, which greatly reduces the huge amount of computation required to directly use the data compilation model. At the same time, it avoids repetitive calculations and has a high cost-effectiveness ratio.

[0097] (3) The practical significance of the hierarchical synthesis model is superior to that of the detailed data compilation model. The higher the level of the unit, the more necessary it is to solve the data compilation matching problem from a macro perspective. When the unit level is low, calculations based on detailed data compilation can help decision-makers solve specific optimization measures for the personnel of that unit. When the unit level is high, decision-makers are required to grasp the direction and do a good job in top-level design. The hierarchical synthesis evaluation model designed in this paper not only ensures the rationality of the data, but also fully considers the different characteristics of the subordinate units, and the results are more objective.

[0098] 4. High-dimensional vector step-by-step synthesis evaluation model

[0099] 4.1 Evaluation Method Based on Hierarchical Synthesis of High-Dimensional Vectors

[0100] The high-dimensional vector evaluation method employs a step-by-step composite evaluation, with largely the same evaluation steps as the high-dimensional vector evaluation method. For each level of composite evaluation value, the evaluation values ​​of all D-level units are grouped according to their corresponding C-level units, and then the average value is calculated to obtain the evaluation value for each C-level unit. Similarly, the average value is calculated for each B-level unit corresponding to a C-level unit, resulting in the composite evaluation value for the A-level unit. The evaluation results for all units at levels A, B, and C are then synthesized step-by-step.

[0101] 4.2 Direct evaluation methods for units at all levels

[0102] The difference between direct evaluation at each level and the step-by-step composite evaluation method lies in the data sampling. For example, to directly evaluate a level C unit, the direct evaluation method involves extracting the personnel data of all personnel within that level C unit, calculating the similarity of all vectors, and then taking the average as the unit's personnel matching score.

[0103] In this algorithm, to calculate the matching degree of all personnel under a certain C-level unit, simply set the D value to 0 to sample data from all personnel under that C-level unit. For example, for unit 1A-2B-3C, by taking the values ​​ABCD as 1230, the data of all personnel in 1A-2B-3C can be extracted from the total data and combined into a new data unit. After vector calculation, the evaluation value of that C-level unit can be obtained. The sampling algorithm is shown in Algorithm 2.

[0104]

[0105] Similarly, for level B units, if we want to calculate the direct evaluation value of 1A-2B units, then ABCD should be set to 1200, and the sampling algorithm should specify the calculation unit as follows: (14) The evaluation values ​​for all B-level units can be obtained through calculation.

[0106] Similarly, for unit A, taking values ​​of 1000 according to ABCD, the sampling algorithm provides the following calculation unit: (15) This means that the evaluation value of Grade A unit can be obtained by calculating the total data of Unit A.

[0107] In the program, the sampling algorithm iterates through each level of units sequentially, and then performs vector calculations to obtain the evaluation values ​​of all A, B, and C level units.

[0108] 4.3 Comparative Analysis of Calculation Results of Step-by-Step Synthetic Evaluation and Direct Evaluation

[0109] Comprehensive comparison Figure 8 , Figure 9 , Figure 10 The following conclusions can be drawn from the three sets of data shown:

[0110] (1) The calculation results of the two high-dimensional vector algorithms have relatively small differences, which can effectively verify the credibility of the two algorithms.

[0111] (2) The calculation results of the high-dimensional vector stepwise synthesis algorithm are generally slightly smaller than those of the direct evaluation algorithm, which is consistent with the characteristics of the high-dimensional vector algorithm. The reason is that the sampling range of the high-dimensional vector direct evaluation method is larger than that of the stepwise synthesis algorithm. In the process of eigenvector calculation, the range of pairwise comparison of vectors is wider, so the calculated matching degree value is higher.

[0112] (3) The high-dimensional vector direct evaluation method is more operational for calculating the global organizational and functional matching degree of each level of unit. For each level of unit, the direct evaluation method can be used to evaluate the unit without relying on the organizational and functional matching degree data provided by each subordinate unit. It can be directly calculated on the global data, which is more convenient to operate and can better reflect the global organizational and functional matching degree of a unit.

[0113] 5. Dynamic Correction Model for Personnel Matching Degree

[0114] In actual personnel use, personnel ranks, ages, and years of service may change. Therefore, the personnel-to-current-employment matching degree of an organization needs to be dynamically adjusted, necessitating the establishment of a dynamic adjustment model for personnel-to-current-employment matching degree. This model needs to consider changes in the actual number of personnel in future years, influencing factors including current personnel rank, age, years of service, and resignation probability. Then, the personnel-to-current-employment matching degree for each year is evaluated again according to the model established in this invention. Finally, the matching degree for future years is used as one of the adjustment factors for the current year's matching degree, thus correcting the original model and obtaining a dynamic personnel-to-current-employment matching model.

[0115] 5.1 Simulation and Prediction Model for Personnel Status Changes

[0116] The logical model of personnel status changes is as follows: Figure 12 As shown.

[0117] The simulation input includes the minimum and maximum years of service and age for each level, the overall promotion probability, the probability of retiring beyond the required years of service or age, the probability of mid-career transfer, and the number of newly recruited personnel. The simulation completes the data on the age and years of service of existing personnel at each level, reads personnel change probability information line by line, and samples four scenarios: promotion to the next level, remaining at the current level without promotion, retiring beyond the required years of service or age, and mid-career transfer. Personnel changes are recorded, and the status for the next year is generated. The simulation then checks if the termination condition is met. If not, it proceeds to the next loop; if so, it outputs the simulation results and obtains a table of actual personnel change information for each future year. Figure 13 As shown.

[0118] The data required for the simulation prediction model are as follows: (1) Basic information of actual personnel. It needs to be specific to individuals and include the unit name, position, major, grade, age, and years of service at the current grade.

[0119] (2) Upper limit regulations. This includes the maximum age limit and the maximum number of years required for the current level.

[0120] (3) Probability of personnel dynamic changes. It is necessary to give the corresponding probability of promotion, probability of mid-term transfer, and probability of maintaining the current level according to the level and the length of service at the current level.

[0121] (4) Rules for recruiting new employees.

[0122] 5.2 Dynamic Personnel Staffing Matching Correction Model

[0123] To assess the impact of future years' matching rates on the current evaluation, the project team adopted the discounted cash flow method commonly used in economics, using indicators from the next few years as influencing factors to revise the existing model. To improve the matching degree of the corrected personnel allocation, Let R be the staffing-to-current-age matching degree for each year, and R be the discount rate. The dynamic staffing-to-current-age matching correction model is as follows: (16)

[0124] 5.3 Data Analysis and Processing

[0125] Based on the dynamic personnel allocation and matching correction model, existing data needs to be processed and supplemented, mainly including the following aspects:

[0126] 5.3.1 Minimum and maximum years of service and maximum age for each level

[0127] Based on actual work conditions and relevant regulations, this document standardizes the minimum years of service, maximum years of service, and maximum age for each level, as shown in Table 8: Table 8. Correspondence between Grades and Minimum, Maximum, and Maximum Required Years of Service

[0128] 5.3.2 Overall promotion probability, probability of retiring beyond the service limit or age limit, and probability of mid-career transfer for each level

[0129] In practice, each officer and soldier has a different career path, and their age and promotion speed vary, leading to continuous changes in the unit's personnel structure. Therefore, it is necessary to provide reasonable dynamic change conditions based on actual work conditions and relevant regulations, quantitatively analyze the future development trend of the unit's personnel structure through mathematical models, and take targeted measures to make adjustments.

[0130] To dynamically grasp changes in personnel structure, the first consideration should be the impact of promotions and retirements. This includes the probability of promotion to a higher rank after reaching the minimum service requirement for their current rank, and the probability of having to retire after exceeding the maximum service requirement or age limit for their current rank. Furthermore, there is a certain probability of demobilization each year between the minimum and maximum service requirements for each rank. For ease of explanation, this paper refers to the years between the minimum and maximum service requirements for each rank as the promotion interval.

[0131] In practical applications, historical data should be collected or the promotion plan ratio of the military should be strictly followed. However, the data for high-level units is extensive and highly classified, making accurate data difficult to obtain. Therefore, this paper uses proportional sampling, combined with the general situation of unit and individual development, to provide reasonable overall promotion probabilities, probabilities of exceeding the service limit and retirement age, and probabilities of mid-career demobilization for each level. Furthermore, the probability distribution of promotion for each specific year within the promotion range for each level is given. Regarding the situation of mid-career demobilization within the promotion range, considering the extremely low demobilization rate for lower-level units and the relatively low selection rate for higher-level units, this paper sets the demobilization rate for the four mid-level levels (S5-S8) at 3%, and does not consider mid-career demobilization for other levels. Detailed data is shown in Table 9. Table 9. Data on overall promotion probability, probability of retiring beyond the age limit, and probability of mid-career transfer for each level.

[0132] 5.3.4 Complete the data on the age and years of service of the actual personnel.

[0133] In practical applications, this data should use the actual age and years of service of the personnel reported by the unit under study. However, this question does not provide detailed data on the age and years of service of the actual personnel. Therefore, this paper refers to the configuration patterns of actual personnel in actual work, and combines the annual promotion probability distribution ratio within each promotion range given above, to supplement the data through sampling. It provides the common age ranges and years of service for each level under normal circumstances, as well as the distribution patterns of personnel included in each years of service. Sampling is then conducted based on this, as shown in Table 10: Table 10: Age Range and Years of Service for Each Level

[0134] 5.3.5 Analysis of Newly Added Personnel

[0135] Unit A is considered as a system. The replacement is carried out according to the principle of replacing one person when one person leaves the system. In addition, the replacement personnel only consider newly graduated personnel from colleges and universities, that is, the initial level of the newly replaced personnel is S01.

[0136] 5.4 Model Solving and Analysis

[0137] Taking the entire 1A unit as the analysis object, starting from the D-level units, the personnel staffing and current status matching degree of the final top-level unit 1A for each year is obtained by aggregating the data level by level. After correction, the personnel staffing and current status matching degree for that year is obtained. The detailed solution process is as follows:

[0138] (1) Based on the distribution probability of personnel age and years of service at each rank, the initial values ​​of the current personnel age and years of service at each rank were obtained by sampling, as shown in Table 11: Table 11 Sampling data on personnel age and years of service at current level

[0139] (2) Based on the overall promotion probability, the probability of retiring beyond the required service period and age limit, and the probability of mid-career transfer for personnel at each level, the promotion probability and mid-career transfer probability corresponding to each level and the current level's service period are calculated. A typical data processing procedure is shown in Table 12: Table 12 Existing Grade Information Data Table

[0140] This indicates that S04-level personnel are likely to be promoted within 4-7 years, while the promotion probability within 1-3 years and over 7 years is 0. Based on the annual promotion probability distribution within the promotion range, we can assume the promotion probabilities for 4-7 years are X, 2X, 2X, and X, respectively, leading to the following equation: (17) Solving for x, we get: x = 0.1346 Therefore, for S04 level, the promotion probability after 4-7 years is: 0.1346, 0.2692, 0.2692, 0.1346, while the promotion probability for other years is 0.

[0141] (3) Based on the promotion probability, mid-term transfer probability, and retirement limit for personnel of each level and current level, simulate and predict the changes in personnel status in future years to obtain the actual personnel data for the 2nd to 5th years.

[0142] (4) The staffing matching degree for the next 5 years is calculated as follows:

[0143] The matching degree of personnel establishment will continue to be calculated for the next 5 years. The detailed calculation results of the matching degree for each year are shown in Table 13.

[0144] Table 13 Calculation Results for Unit 1A

[0145] (5) Assuming a discount rate R = 0.9, and using this as a weighting factor, the matching factor for the next 5 years is used to adjust the personnel matching degree for the current year as follows: (18) Because calculating the probability of each level and related years, as well as the changes in personnel status in each year, involves a large amount of computation, this invention developed a MATLAB program for processing. Through this program, the personnel statistics data of each level of Unit A can be called up at any time for calculation, and the personnel statistics data for future years can be quickly obtained.

[0146] 5.5 Results Analysis

[0147] 5.5.1 Analysis of Calculation Results of Multi-Indicator Comprehensive Evaluation

[0148] Figure 14 This is a comparison chart of the results of the dynamic correction model and the original model for multi-indicator comprehensive evaluation. Figure 15 To comprehensively evaluate the dynamic development trend of personnel matching over the next 5 years using multiple indicators.

[0149] like Figure 15 As shown, over the next five years, the staffing matching degree of Unit A's 1A level, 1A-1B level, 1A-2B level, and 1A-3B level units will decrease year by year as time goes by.

[0150] Figure 16 This is a comparison chart showing the matching degree distribution between the dynamic correction model and the original model for multi-indicator comprehensive evaluation. (Example) Figure 16 As shown, after adjusting the model to reflect dynamic changes over the next five years, the distribution range of personnel staffing matching degree changes significantly. The number of units with a matching degree of over 90% decreased from 129 to 55, a decrease of 57%. However, the number of units with a matching degree below 80% did not change significantly, increasing from 5 units before the adjustment to 6 units. This indicates that over time, the entry, exit, and retention of personnel have a significant impact on the personnel staffing matching degree of units, and this impact has a certain lag. It will not immediately cause the unit to cease operation, but it poses a significant hidden danger to the unit's development over time and must be given high priority.

[0151] 5.5.2 Analysis of High-Dimensional Vector Evaluation Calculation Results

[0152] (1) Regarding the fluctuation of the matching degree of the modified model

[0153] Figure 17 This diagram illustrates the distribution comparison between the matching degree of the high-dimensional vector evaluation dynamic correction model and the original model. The slight downward trend in the matching degree fluctuation of the correction model aligns with the assumptions of the dynamic model. After correction by the dynamic correction model, the results better reflect the actual situation of personnel turnover.

[0154] (2) Based on the dynamic changes of the assumptions, the matching degree trend of each unit in the next 5 years

[0155] Figure 18To evaluate the dynamic development trend of personnel matching degree over the next 5 years using high-dimensional vectors, Figure 19 This study uses a high-dimensional vector to evaluate the dynamic development trend of personnel matching in all C-level units over the next five years. Looking at the development trend of personnel matching in large units over the next five years, the overall personnel matching in large units shows low sensitivity to personnel dynamics. This aligns with the actual personnel fluctuations in large units; large units have greater flexibility in personnel adjustment and are less likely to experience significant personnel matching pressure or large-scale fluctuations due to personnel promotions or retirements, resulting in overall personnel stability. It is worth noting that the personnel matching shows an upward trend in the fourth and fifth years. Analysis suggests this is mainly due to the promotion pressure faced by personnel at their highest tenure. Typically, the PLA has a hierarchical pyramid structure; higher-level units have more high-ranking personnel, leading to high mobility among high-ranking personnel, either through transfers, adjustments, or retirement. This provides some promotion opportunities for personnel at lower levels.

[0156] Looking at the five-year trend of staffing-to-position matching in various Class C units, they are more sensitive to personnel dynamic adjustments compared to larger units. This situation is consistent with the actual fluctuations in lower-level units. Among them, some units, such as 1A2B1C and 1A2B7C, exhibited significant fluctuations in their annual staffing-to-position matching rates. We can consider these units to be technical units with low differentiation in technical post levels (the job levels are junior, intermediate, and senior, with a wide range of rank connections), resulting in a significant mismatch between staffing and position. Although current regulatory policies allow lower-level staff to occupy higher-level positions, the flexibility in staffing-to-position matching remains considerable. Furthermore, these types of units also experience relatively high personnel turnover (this can be corroborated by examining the original data). High personnel mobility will, in the long run, affect the unit's operational continuity. This requires attention. Therefore, to retain personnel, the staffing-to-position matching rate of these units should be evaluated separately, and the staffing and compensation policies for these personnel should be adjusted accordingly.

[0157] (3) Distribution of current matching degree of each unit and the current matching degree of units after dynamic correction model

[0158] Figure 20 This diagram illustrates the comparison between the matching degree distribution of the high-dimensional vector evaluation dynamic correction model and the original model. Compared to the current matching degree distribution, the matching degree distribution of each unit after dynamic correction converges towards the center of a normal distribution. After dynamic model correction, the number of units with low matching degrees has increased. This indicates that the large unit's intervention and control over the overall personnel prevents individual units from having excessively low matching degrees, adjusting them from high-matching units to matching-degree units to ensure the normal operation of the unit. The overall matching degree distribution shifts towards the median value. This further demonstrates that the dynamic correction model is effective.

[0159] In reality, an organization's staffing situation may not exactly match its established staffing levels. Therefore, many organizations dynamically adjust their staffing based on their specific circumstances to meet their real-time operational needs. Thus, when assessing the staffing-to-actual-staff match of an organization, it is crucial to fully consider these flexible measures to avoid rigidly adhering to staffing levels while ignoring the organization's actual personnel requirements.

[0160] 6. Floating Matching Degree Model and its Solution

[0161] In reality, an organization's personnel situation may not exactly match its established staffing levels. Therefore, many organizations dynamically adjust their staffing based on their specific circumstances to meet their real-time operational requirements. Thus, when assessing the staffing-to-actual-staff match rate of an organization, it is necessary to fully consider these flexible measures to prevent rigidly adhering to staffing levels while ignoring the organization's actual personnel needs. To this end, this invention considers the following measures to recalculate the match rate, using a floating match rate model to meet the normal operation needs of the target organization.

[0162] 6.1 Establishment of the Floating Matching Degree Model

[0163] (1) Fluctuation in the actual number of people

[0164] To meet the daily operational needs of an organization, the actual number of personnel required may be adjusted based on the organization's workload and requirements. Therefore, when examining the staffing-to-current personnel match, we must allow for such reasonable adjustments by the target organization. Consequently, the staffing requirements for this organization need to be relaxed to some extent when calculating the match. Let the required staffing number be... The actual number of people is The matching degree at this point is The allowed personnel fluctuation ratio is defined as follows: Therefore, when the actual number of employees is within the allowed fluctuation range, we consider the number of employees to meet the staffing requirements; when the actual number of employees is outside the allowed fluctuation range, we consider the staffing level to be at the critical value closest to the actual number of employees. Specifically, regarding the staffing-to-actual-employment matching degree... We define the matching degree of the current program as allowing for fluctuations in the number of participants as: (19)

[0165] (2) Complementary personnel of adjacent levels

[0166] In actual organizational operations, a certain degree of error is usually permissible in personnel ranks. For example, personnel with a rank one level higher or lower can work in positions requiring a certain rank, commonly referred to as "lower-ranking staff with higher-ranking positions" or "higher-ranking staff with lower-ranking positions." This satisfies the actual needs of the organization while ensuring that personnel dynamics remain within a reasonable range, as shown in the research results of Question 3. Therefore, to fully consider the matching of personnel rank fluctuations, what should be done regarding the required rank? And the level of actual personnel Let the matching degree of the current encoding be . We define the matching degree that allows for fluctuating levels as: (20)

[0167] in For randomly generated values, The percentage of personnel allowed to fluctuate for a given level, i.e., the allowed percentage among all personnel. The people are matched based on their ranking; the calculation of the matching is as follows:

[0168] (twenty one) in This is a permissible difference in grade level. For example... This means that 10% of the personnel in the unit are allowed to be at a level different from the required staffing level, and the allowed level difference is one level above or below.

[0169] 6.2 Solution and Analysis of the Floating Matching Model

[0170] In this section, we recalculate the personnel matching degree of several typical units using the constructed floating matching degree model. For each unit, we recalculate using both the multi-index comprehensive evaluation method and the high-dimensional vector evaluation method. We further analyze the impact of the floating model parameters on the calculation results and provide corresponding rationalization suggestions.

[0171] 6.2.1 Selection of Typical Units

[0172] In this invention, we selected five typical units for analysis. These five units have the characteristics of high matching degree, low matching degree, and large number of personnel, respectively. The basic information of the five units is shown in Table 14.

[0173] Table 14 Data Table of Five Typical Units

[0174] 6.2.2 Impact of Staff Floating Rate

[0175] In this section, for the five selected typical units, the personnel fluctuation ratio is set as follows: The matching degree was recalculated using the multi-index comprehensive evaluation method and the high-dimensional vector evaluation method, respectively.

[0176] The results obtained using the multi-index comprehensive evaluation method are as follows: Figure 21 As shown in the figure, the staffing-to-work matching degree of each unit gradually increases with the increase in the personnel fluctuation ratio. This indicates that relaxing the personnel matching requirements can improve the staffing-to-work matching degree of most units, which also reflects that these units actually have a problem with poor personnel matching. However, the matching degree of units 1A-3B-7C-5D and 1A-3B-8C-7D did not increase significantly, indicating that their personnel matching is actually better, and therefore they are not very sensitive to the personnel fluctuation ratio.

[0177] The results obtained using the high-dimensional vector evaluation method are as follows: Figure 22 As shown. From Figure 22 As can be seen, the staffing-to-work matching degree of each unit gradually improves as the staffing-to-work ratio increases. This indicates that relaxing the staffing-to-work matching requirements can improve the staffing-to-work matching degree of most units, which also reflects that these units actually have a problem with poor staffing matching. However, the matching degree of units 1A-3B-7C-5D and 1A-3B-8C-7D did not improve significantly, indicating that their staffing-to-work matching is actually better, and therefore they are not very sensitive to the staffing-to-work ratio.

[0178] 6.2.3 Impact of Grade Fluctuations

[0179] In this section, for the three selected typical units, we set different level fluctuation ratios and level fluctuation differences, and recalculated the compilation matching degree.

[0180] To investigate the impact of grade fluctuation ratio on the matching degree of compilation, we fixed the grade fluctuation difference at 1. Set the level fluctuation ratios as follows: The matching degree was recalculated using the multi-index comprehensive evaluation method and the high-dimensional vector evaluation method, respectively.

[0181] The results obtained by calculating the impact of grade fluctuation ratio using the multi-index comprehensive evaluation method are as follows: Figure 23 As shown. From Figure 23 As can be seen, the staffing and personnel matching degree of each unit gradually increases with the increase in the grade fluctuation ratio. This shows that relaxing the requirements for grade and personnel matching can improve the staffing and personnel matching degree of most units, which also reflects that these units actually have a problem with poor grade matching. However, the matching degree of a certain unit does not improve much, indicating that its grade matching is actually better, and therefore it is not very sensitive to the grade and personnel fluctuation ratio.

[0182] The results obtained by calculating the impact of grade fluctuation ratio using the high-dimensional vector evaluation method are as follows: Figure 24As shown. From Figure 24 As can be seen, the matching degree of staffing for each unit gradually increases with the increase in the percentage of grade fluctuations. This indicates that relaxing the requirements for matching the number of staff at each grade can improve the matching degree of staffing for most units, which also reflects that these units actually have a problem with poor grade matching. The matching degree of units 1A-3B-7C-5D and 1A-3B-8C-7D did not improve much, indicating that their grade matching was actually better, and therefore they were not very sensitive to the percentage of grade fluctuations.

[0183] In summary, the advantages of the floating matching model are:

[0184] (1) In the multi-indicator comprehensive evaluation model, the past practice of simply evaluating the staffing level based on the full staffing rate has been optimized, taking into account the influence of staff quantity, grade and profession. At the same time, a scientific method for accurately determining the weight of the three indicators is proposed, which can comprehensively examine the overall staffing matching of the unit.

[0185] (2) The design and application of batch processing programs have significantly improved data processing efficiency. Batch processing programs not only meet the functional requirements of the model for both Problem 1 and Problem 2, but also allow modification of input data such as weights, staffing, and actual personnel, making it easier to dynamically predict the changing trend of staffing matching degree.

[0186] (3) Some of the assumptions of the multi-indicator comprehensive evaluation model refer to the current human resources regulations, and the data results have strong reference value.

[0187] (4) The high-dimensional vector evaluation model has strong scalability, can add other influencing factors at will, and can evaluate the comprehensive influence of all factors on the matching degree at the same time.

[0188] (5) The hierarchical synthesis model can better reflect the authenticity of the matching of sub-units at all levels, and avoids the problem of false complementarity between over-complementary and under-complementary sub-units causing the overall matching degree to be artificially high.

Claims

1. A floating matching model for evaluating the matching degree of personnel staffing, comprising: Multi-indicator comprehensive evaluation sub-model or high-dimensional vector evaluation sub-model, multi-indicator step-by-step synthesis evaluation sub-model or high-dimensional vector step-by-step synthesis evaluation sub-model, personnel staffing matching degree dynamic correction sub-model, floating matching degree sub-model; among them; The multi-indicator comprehensive evaluation sub-model is used to solve the comprehensive matching degree of personnel staffing in level D units, which includes three secondary sub-objectives: quantity matching degree, professional matching degree, and grade matching degree. The high-dimensional vector evaluation sub-model transforms the personnel quantity, grade, profession, and position of the D-level unit into a set of feature values ​​for each factor according to different characteristics. It combines the feature values ​​of each factor of the current position and represents them with a high-dimensional vector. By comparing the cosine similarity between the current position vectors, it transforms the calculation into solving the similarity between the compilation data vector and the existing data vector. The multi-indicator hierarchical composite evaluation sub-model is used to solve the staffing and actual matching degree of personnel in higher-level units of the D-level unit. It receives the comprehensive staffing and actual matching degree of personnel in the D-level unit solved by the multi-indicator comprehensive evaluation sub-model, and performs a weighted average according to the actual number of personnel in each lower-level unit to calculate the staffing and actual matching degree of personnel in higher-level units. The high-dimensional vector hierarchical synthesis evaluation sub-model obtains the evaluation value of each C-level unit by grouping the evaluation values ​​of all D-level units according to their corresponding C-level units and then calculating the average value. Similarly, it obtains the composite evaluation value of each B-level unit by grouping the B-level units corresponding to the C-level units and calculating the average value. Finally, it obtains the composite evaluation value of the A-level units and synthesizes the evaluation results of all A, B, and C-level units hierarchically. The dynamic correction sub-model for personnel staffing matching degree considers the changes in actual personnel in future years. Based on the personnel's current level, age, years of service, and resignation probability, it evaluates the staffing matching degree of the multi-indicator comprehensive evaluation sub-model, high-dimensional vector evaluation sub-model, multi-indicator stepwise synthesis evaluation sub-model, or high-dimensional vector stepwise synthesis evaluation sub-model. The matching degree of future years is used as one of the adjustment factors for the staffing matching degree of the current year to correct the sub-model and dynamically correct the staffing matching degree. The floating matching degree sub-model dynamically adjusts the personnel-to-position matching degree of any sub-model based on the differences in the number of personnel in the unit and the complementarity of personnel at adjacent levels. This adjustment is made based on the multi-indicator comprehensive evaluation sub-model, the high-dimensional vector evaluation sub-model, the multi-indicator step-by-step synthesis evaluation sub-model, the high-dimensional vector step-by-step synthesis evaluation sub-model, and / or the personnel-to-position matching degree dynamic correction sub-model.

2. The floating matching model for evaluating personnel staffing matching degree according to claim 1, characterized in that, The overall matching degree of personnel establishment in the D-level unit The calculation is as follows: In the formula, The weight of the quantity matching degree, Weighting for professional matching degree The weight of the level matching degree, the quantity matching degree The calculation method is as follows: In the formula, This indicates the number of staff authorized for this Class D unit. This indicates the actual number of personnel in the D-level unit.

3. The floating matching model for evaluating personnel staffing matching degree according to claim 2, characterized in that, The professional matching degree The calculation method is as follows: In the formula, For the number of majors, , and The first and second data in the compiled and actual data respectively The number of students in each major; the matching degree of the level The calculation method is as follows: In the formula, For the number of levels, and The first and second data are respectively compiled data and actual data. The number of personnel at each level Indicates the first The average years of service converted to each level.

4. The floating matching model for evaluating personnel staffing matching degree according to claim 1, characterized in that, The high-dimensional vector evaluation sub-model includes: extracting features from four factors—position, grade, major category, and sub-category—to obtain a dataset of feature vectors; and weighting the feature vectors according to the relative importance of the four factors. The feature extraction process includes: For unit numbering, use This indicates that the unit 1A-1B-1C-1D is represented as ; For the job factor P, the number of jobs within the evaluated unit is counted and denoted as Pjob. If there are [number], then set [number] 0-1 vectors of dimension For the first position, the first digit is 1, and the rest are 0, denoted as . ; For the grade factor L, the numbers corresponding to each grade are used as input values. The 11 levels S01-S11 represent the feature vector values ​​of each level, from 1 to 11. For the professional factor M, the number of professional categories within the evaluated unit is counted and denoted as M. ,set up 0-1 vectors of dimension For the first major category, the first digit is 1, and the rest are 0, denoted as . Next, count the number S of professional subcategories included in each major category, and select the one with the largest number of professional subcategories, denoted as . ,set up 0-1 vectors of dimension For each of their respective professional subcategories, the first one is 1, and the rest are 0, denoted as... ; The feature vector of each person in the process is: The set of vectors of all compilers ; The feature vector for each person in the existing data is as follows: The existing set of personnel vectors .

5. The floating matching model for evaluating personnel staffing matching degree according to claim 4, characterized in that, The weighting process for the feature vectors includes: Assuming the relative importance of job position, grade, major category, and subcategory are respectively... The combined vector is: The weighted vector is: in, For the identity matrix, others It is the identity matrix of the corresponding dimension. 。 6. The floating matching model for evaluating personnel staffing matching degree according to claim 1, characterized in that, The multi-index hierarchical synthesis evaluation sub-model is used to solve the personnel staffing matching degree of the higher-level units of the D unit. It calculates the personnel staffing matching degree of the next-level units of the D unit by weighting the average of the actual number of personnel in each next-level unit. The formula for calculating the personnel matching degree of the superior unit is as follows: 。 7. The floating matching model for evaluating personnel staffing matching degree according to any one of claims 1-6, characterized in that, The dynamic correction model for personnel matching degree further includes a simulation prediction model for personnel status changes. The data required for the simulation prediction model are as follows: Basic information of actual personnel. This needs to be specific to each individual, including the organization name, position, major, grade, age, and length of service at the current grade. The upper limit regulations include the maximum age limit and the maximum number of years required to reach the current level; The probability of personnel dynamic changes needs to be given according to the level and the number of years at the current level, including the corresponding probability of promotion, the probability of mid-term transfer, and the probability of maintaining the current level. Newcomer supplementary rules.

8. The floating matching model for evaluating personnel staffing matching degree according to claim 7, characterized in that, The dynamic correction model for personnel staffing matching degree uses evaluation indicators for the next few years as influencing factors, and its correction method is as follows: ； in, To improve the matching degree of the corrected personnel allocation, R represents the staffing matching degree for each year, and R is the discount rate.

9. The floating matching model for evaluating personnel staffing matching degree according to claim 7, characterized in that, The floating matching degree sub-model calculates the matching degree of each level of unit based on the up and down fluctuation of the actual number of people; let the required number of people be... The actual number of people is The matching degree at this point is The allowable personnel fluctuation ratio is: If the actual number of employees is within the allowed fluctuation range, the number is considered to meet the staffing requirements; if the actual number of employees is outside the allowed fluctuation range, the staffing level is considered to be at the critical value closest to the actual number. Regarding the staffing-to-actual-employment matching degree... Define the matching degree of the number of people allowed to fluctuate as: 。 10. A method using a floating matching model as described in any one of claims 1-9, comprising: S1: Set the initial conditions for the floating matching model. The initial conditions include: the staffing requirement of the given unit will not change; adjustments to the unit structure and personnel transfers are not considered. S2: Data processing and handling of the multi-index comprehensive evaluation sub-model, including: determining the weights of sub-objectives; processing current data; and converting grades to years of service. S3: Data processing is performed on the high-dimensional vector evaluation sub-model, including: forming a data vector according to the current number of people to obtain the current staffing data; establishing a staffing data sampling algorithm to sample the current staffing data to obtain a staffing dataset for a given D-level unit; based on the staffing dataset and the current staffing dataset, calculating the similarity between any vector in the dataset to obtain the similarity between two personnel feature vectors; and averaging the similarities of all feature vectors to obtain the matching degree for a given D-level unit. S4: Use a multi-indicator hierarchical synthetic evaluation sub-model and / or use a detailed compilation data model to solve the personnel compilation matching degree of the superior unit; S5: Perform data analysis and processing on the dynamic personnel staffing matching correction model, including: calculating the minimum and maximum years of service and the maximum age for each level; calculating the overall promotion probability, the probability of exceeding the years of service and the probability of retiring due to age, and the probability of mid-career transfer for each level; completing the age and years of service data for the actual personnel; taking the 1A large unit as the whole as the analysis object, starting from the D level unit, aggregating level by level to obtain the personnel staffing matching degree for each year of the final top-level unit 1A, and obtaining the personnel staffing matching degree for the current year after correction; S6: Based on the fluctuation of the actual number of people, the matching degree is recalculated using the floating matching degree sub-model, and the floating matching degree model is used to meet the needs of the target unit for normal operation.

Images