A multi-dimensional hierarchical work assessment management method and system

Abstract

This invention discloses a multi-dimensional, hierarchical performance evaluation management method and system, relating to the field of enterprise management. The method includes: obtaining an organizational structure tree diagram; converting the hierarchical relationships between the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database; extracting annual strategic goals based on the hierarchical database, breaking them down into quarterly OKRs for departmental management, and further subdividing them into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table; and using the Analytic Hierarchy Process (AHP) to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor. This invention achieves cross-level collaborative adjustment by converting the organizational structure tree diagram into a hierarchical database and constructing a three-level indicator mapping table, using the AHP method for weight allocation, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor, and further constructing a three-body entangled state through qubit encoding.

Description

Technical Field

[0001] This invention relates to the field of enterprise management, and in particular to a multi-dimensional, hierarchical performance evaluation management method and system. Background Technology

[0002] In modern enterprise management, the design and implementation of performance appraisal systems directly affect the efficiency of achieving organizational strategic goals. Traditional performance appraisal methods mainly adopt a hierarchical decomposition model based on the Balanced Scorecard or Key Performance Indicators. This model breaks down corporate strategic goals into departmental tasks, which are then further allocated to individual job responsibilities, forming a three-tiered goal management system. This technical solution calculates the weights of each level through linear weighting and uses a regular manual review mechanism to ensure the rationality of the indicators. Such methods have formed standardized processes in terms of goal alignment and data traceability, which can meet basic management needs. However, as enterprises expand and business complexity increases, traditional methods face significant challenges in terms of dynamic adaptability. When the organizational structure is adjusted or the strategic direction changes, the existing system is unable to respond quickly and recalibrate the cross-level indicator correlations, resulting in deviations between the appraisal results and the actual situation.

[0003] The existing technology system has room for optimization in achieving cross-level collaboration, especially when dealing with non-linear relationships between the strategic, departmental, and execution levels. Traditional linear weight allocation methods are difficult to accurately quantify the transmission efficiency of cross-level influences. When a department needs to temporarily adjust its OKRs due to changes in the external environment, the existing system cannot automatically trigger the synchronous correction of related strategic goals and individual KPIs, and still requires manual intervention. This limitation is essentially due to the insufficient modeling of the coupling relationship between levels by traditional methods and the failure to establish a dynamic weight feedback mechanism. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a multi-dimensional hierarchical work performance evaluation and management method that solves the problems of insufficient cross-level dynamic collaboration and difficulty in multi-source data fusion in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] Firstly, the present invention provides a multi-dimensional, hierarchical performance evaluation and management method, which includes,

[0008] Obtain the organizational structure tree diagram and convert the hierarchical relationships of the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database;

[0009] Based on the hierarchical database, the annual strategic goals are extracted, broken down into quarterly OKRs for departmental management, and then further divided into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The AHP (Analytic Hierarchy Process) is used to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor.

[0010] The weight vector of the strategic decision-making level, the weight matrix of the department management level, and the weight tensor of the grassroots execution level are encoded into qubits. A three-body entangled state of the strategic decision-making level indicators, the department management level indicators, and the grassroots execution level indicators is constructed. Cross-level collaborative adjustment is triggered by quantum measurement to obtain the adjusted weight data.

[0011] Based on the adjusted weighted data, multi-source assessment data from the strategic decision-making level, department management level, and grassroots execution level were collected from the ERP module, CRM module, and code repository and normalized to generate a standardized data table.

[0012] By embedding the topological insulator algorithm into a standardized data table, a clean data stream is obtained.

[0013] The clean data stream is input into the hierarchical evaluation model. DEA data envelopment analysis is used for the strategic decision-making level, the entropy weight-TOPSIS algorithm is used for the department management level, and XGBoost is used for the grassroots execution level to generate an assessment score table with hierarchical labels.

[0014] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation and management method described in this invention, the following steps are included: obtaining an organizational structure tree diagram and converting the hierarchical relationships between the strategic decision-making layer, departmental management layer, and grassroots execution layer into a hierarchical database.

[0015] Export the organizational structure tree diagram through the enterprise OA module, analyze the hierarchical depth of the organizational structure tree diagram, and obtain the affiliation relationship between the strategic decision-making level, departmental management level, and grassroots execution level;

[0016] The strength of hierarchical associations is quantified by a dynamic weighted adjacency matrix algorithm, and the strategic decision-making level, departmental management level, and grassroots execution level are transformed into a hierarchical database by combining the Cypher statement of the Neo4j graph database.

[0017] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation and management method described in this invention, the following steps are included: extracting annual strategic goals from a hierarchical database, breaking them down into quarterly OKRs for departmental management, and further subdividing them into monthly KPIs for the grassroots execution layer, forming a three-level indicator mapping table; using the Analytic Hierarchy Process (AHP) to assign weights to the three-level indicator mapping table, generating a strategic decision-making layer weight vector, a departmental management weight matrix, and a grassroots execution layer weight tensor; and including the following steps.

[0018] The annual targets of strategic decision-making level nodes are queried from the hierarchical database to obtain the original strategic targets. Quantitative indicators are extracted from the original strategic targets using regular expressions to obtain structured indicators.

[0019] Based on the actual completion rate data of historical strategic indicators in the CRM module, calculate the historical completion rate of each department for each item in the structured indicators, and obtain the departmental capability score;

[0020] The structured OKR indicators are broken down to obtain a departmental OKR list;

[0021] Based on the weighting ratio of the department's OKR list target values ​​and employee contribution data, an individual KPI list is calculated.

[0022] By linking structured indicators, departmental OKR lists, and individual KPI lists, a three-level indicator mapping table is obtained;

[0023] A quantum state comparison matrix is ​​generated based on a three-level index mapping table. A phase estimation algorithm is then performed on the quantum comparison matrix on an IBM quantum processor to obtain quantum state eigenvalues.

[0024] The quantum state eigenvalues ​​are converted into strategic decision-making layer weight vectors, departmental layer weight matrices, and execution layer weight tensors.

[0025] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation and management method described in this invention, the following steps are included: encoding the strategic decision-making layer weight vector, the departmental management layer weight matrix, and the grassroots execution layer weight tensor into qubits; constructing a three-body entangled state of strategic decision-making layer indicators, departmental management layer indicators, and grassroots execution layer indicators; triggering cross-level collaborative adjustment through quantum measurement to obtain adjusted weight data; and including the following steps.

[0026] The strategic decision-making layer weight vector is encoded as the amplitude state of a qubit to obtain a quantum state. The weight matrix of the department management layer is loaded onto two qubits through a controlled rotation gate to obtain an entangled state. The weight tensor of the grassroots execution layer is used as a third-level conditional probability and loaded onto three qubits through a multi-control gate to obtain a three-body entangled state.

[0027] Based on the reduced density matrix obtained from the three-body entangled state, the entanglement entropy is verified, the verification flag is obtained, the three-body entangled state is measured, and the measurement result statistics table is obtained.

[0028] Based on the statistical table of measurement results, the strategy-department covariance and the department-execution covariance are obtained, and the covariance matrix is ​​obtained.

[0029] Based on the median in the covariance matrix, the critical value for the strength of cross-level management associations is obtained;

[0030] When the strategy-department covariance exceeds the critical value of cross-level management correlation strength, cross-level collaborative adjustment is triggered, resulting in adjusted weight data.

[0031] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation management method described in this invention, the method includes the following steps: Based on adjusted weighted data, multi-source performance evaluation data from the ERP module, CRM module, and code repository is collected from the strategic decision-making level, departmental management level, and grassroots execution level, and normalized to generate a standardized data table.

[0032] Based on the adjusted weighted data, a structured dictionary is obtained. Then, the performance indicator acquisition interface of the ERP module is used to obtain market share performance data from the list of strategic indicators in the structured dictionary, thus obtaining the strategic decision-making level assessment data.

[0033] Using the list of department IDs in the structured dictionary, the quarterly targets and achievement rates of each department are queried through the CRM module interface to obtain the performance evaluation data of the department management.

[0034] Based on the execution unit mapping table in the weight dictionary, the code version control unit commit logs are scanned to form the code contribution density of each execution unit, and the grassroots execution layer assessment data is obtained.

[0035] The assessment data of the strategic decision-making level, the department management level, and the grassroots execution level are normalized to generate standardized data tables.

[0036] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation and management method described in this invention, the method involves: embedding a topological insulator algorithm into a standardized data table to obtain a clean data stream, including the following steps.

[0037] Add hierarchical labels to the standardized data table to obtain a labeled graph structure;

[0038] Calculate the invariants of the boundary edges connecting different levels on a labeled graph structure to obtain a list of anomalous edges;

[0039] Abnormal edges are obtained by analyzing the three-level index mapping table. Abnormal edges are removed according to the abnormal edge list to obtain the purified adjacency matrix. Then, spectral clustering is performed to obtain the three-level node ID list.

[0040] The list of third-level node IDs is filtered to obtain a clean data stream.

[0041] As a preferred embodiment of the multi-dimensional hierarchical performance evaluation and management method described in this invention, the method includes: inputting a clean data stream into a hierarchical evaluation model; employing DEA (Data Envelopment Analysis) for the strategic decision-making level; using the entropy-weighted TOPSIS algorithm for the departmental management level; and using XGBoost for the grassroots execution level to generate an evaluation score table with hierarchical identifiers. This includes the following steps:

[0042] Based on the pure data stream, the data is split into strategic decision-making layer data, departmental layer data and execution layer data according to the hierarchical label field. The resource input and performance indicator achievement values ​​of each strategic indicator are extracted from the strategic decision-making layer data to construct a decision-making unit matrix and obtain the strategic decision-making layer DEA.

[0043] Solve the linear programming problem of the DEA (Decision Analysis) layer of the strategic decision-making level on a quantum processor to form the relative efficiency value of the strategic indicators and obtain the strategic efficiency score;

[0044] Based on departmental data, information entropy and dynamic weights are derived to form an entropy-weight matrix and a departmental weight matrix.

[0045] Based on the department weight matrix, a weighted standardization process is performed to obtain the distance between the positive and negative ideal solutions, and a department ranking is generated.

[0046] From the raw data of code submission records and task response times in the execution layer data, the average number of submissions and the centrality of the code submission network are extracted to obtain the feature matrix execution layer matrix. The feature matrix execution layer matrix is ​​used to train the employee performance ranking model to obtain the XGBoost model.

[0047] The performance scores of all employees are predicted based on the XGBoost model to obtain the performance scores of the execution level.

[0048] By integrating strategic efficiency scores, departmental rankings, and performance scores at the execution level, an assessment score table with hierarchical labels is obtained.

[0049] Secondly, the present invention provides a multi-dimensional hierarchical work performance evaluation and management system, including a data acquisition module, which acquires an organizational structure tree diagram and converts the affiliation relationships of the strategic decision-making layer, departmental management layer and grassroots execution layer into a hierarchical database.

[0050] The decomposition module extracts annual strategic goals from a hierarchical database, decomposes them into quarterly OKRs for departmental management, and further subdivides them into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The AHP (Analytic Hierarchy Process) is used to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor.

[0051] The weight allocation module encodes the weight vector of the strategic decision-making layer, the weight matrix of the department management layer, and the weight tensor of the grassroots execution layer into qubits, constructs a three-body entangled state of the strategic decision-making layer indicators, the department management layer indicators, and the grassroots execution layer indicators, and triggers cross-level collaborative adjustment through quantum measurement to obtain the adjusted weight data;

[0052] Cross-level modules collect multi-source assessment data from the ERP module, CRM module, and code repository, including strategic decision-making level, department management level, and grassroots execution level, based on the adjusted weighted data, and perform normalization processing to generate standardized data tables.

[0053] The data cleaning module embeds a topological insulator algorithm into a standardized data table to obtain a clean data stream.

[0054] The tiered evaluation module inputs clean data streams into the tiered evaluation model. It uses DEA data envelopment analysis for the strategic decision-making level, the entropy weight-TOPSIS algorithm for the department management level, and XGBoost for the grassroots execution level, generating an assessment score table with tiered labels.

[0055] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the multi-dimensional hierarchical work assessment and management method as described in the first aspect of the present invention.

[0056] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the multi-dimensional hierarchical work assessment management method as described in the first aspect of the present invention.

[0057] The beneficial effects of this invention are as follows: By converting the organizational structure tree diagram into a hierarchical database and constructing a three-level indicator mapping table, the Analytic Hierarchy Process (AHP) is used for weight allocation, generating a weight vector for the strategic decision-making layer, a weight matrix for the departmental management layer, and a weight tensor for the grassroots execution layer. Furthermore, cross-level collaborative adjustment is achieved by constructing a three-body entangled state through qubit encoding. Based on the adjusted weight data, assessment data is collected from a multi-source system, purified by the topological insulator algorithm, and input into the hierarchical evaluation model, ultimately generating an assessment score table with hierarchical identifiers. This method innovatively applies quantum computing and topological principles to the field of performance appraisal, solving the technical difficulties of traditional methods in cross-level dynamic collaboration and multi-source data fusion. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0059] Figure 1 A flowchart for a multi-dimensional, tiered performance evaluation and management method.

[0060] Figure 2 This is a schematic diagram of a file encryption system.

[0061] Figure 3 A diagram illustrating the breakdown of annual strategic goals.

[0062] Figure 4 To obtain an organizational structure tree diagram and convert it into a hierarchical database diagram. Detailed Implementation

[0063] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0064] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0065] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0066] Reference Figures 1-4 This is one embodiment of the present invention, which provides a multi-dimensional hierarchical performance evaluation management method, including the following steps:

[0067] S1. Obtain the organizational structure tree diagram and convert the hierarchical relationships of the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database.

[0068] S1.1 Export the organizational structure tree diagram through the enterprise OA module, analyze the hierarchical depth of the organizational structure tree diagram, and obtain the affiliation relationship between the strategic decision-making level, departmental management level, and grassroots execution level.

[0069] Furthermore, the organizational structure tree diagram data containing node ID, parent node ID, and node type is obtained from the enterprise OA interface. A breadth-first search algorithm is used to traverse the tree diagram. Based on the node type field, strategic decision-making level nodes, department management level nodes, and grassroots execution level nodes are labeled. The path length from the root node to each leaf node is calculated. Nodes with a path length of 1 are classified as strategic decision-making level, nodes with a path length of 2 are classified as department management level, and nodes with a path length of 3 or more are classified as grassroots execution level. A hierarchical relationship table containing node ID, hierarchical label, and affiliation is output. In the hierarchical relationship table, strategic decision-making level nodes establish management relationships with their direct subordinate department management level nodes, and department management level nodes establish execution relationships with their subordinate grassroots execution level nodes, forming a complete three-level affiliation relationship of strategic decision-making level - department management level - grassroots execution level.

[0070] S1.2. The hierarchical association strength is quantified by the dynamic weighted adjacency matrix algorithm, and the strategic decision-making layer, departmental management layer and grassroots execution layer are transformed into a hierarchical database by combining the Cypher statement of Neo4j graph database.

[0071] Specifically, the expression is,

[0072]

[0073] Among them, W ij d represents the hierarchical association weight between the current node i and the target node j, β is the hierarchical decay coefficient, and d ij Let i be the management span between the current node i and the target node j, where i is the current node and j is the target node.

[0074] S2. Based on the hierarchical database, extract the annual strategic goals, break them down into quarterly OKRs for departmental management, and then further break them down into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. Use the AHP (Analytic Hierarchy Process) to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management weight matrix, and a grassroots execution level weight tensor.

[0075] S2.1 Query the annual targets of strategic decision-making layer nodes from the hierarchical database to obtain the original strategic targets. Use regular expressions to extract quantitative indicators from the original strategic targets to obtain structured indicators.

[0076] Furthermore, Cypher queries are executed from the hierarchical database to retrieve the annual target text of the strategic decision-making layer node records, obtaining the original set of strategic targets containing natural language descriptions. The re module of Python is used to construct regular expression patterns to match percentage values ​​and key indicator names, match market share values, extract quantitative indicator pairs from the original strategic target text, and generate structured indicators.

[0077] S2.2 Based on the actual completion rate data of historical strategic indicators in the CRM module, calculate the historical completion rate of each department for each item in the structured indicators, and obtain the departmental capability score.

[0078] Specifically, the expression is,

[0079]

[0080] Among them, R k Let A be the historical completion rate score of the target department k for the current strategic indicators, T be the total number of statistical periods, t be the index of the statistical period, and A be the index of the current period. kt Let G be the actual completion value of target department k in the t-th period. kt Let λ be the target value of target department k in the t-th period, λ be the time decay coefficient, and k be the target department.

[0081] S2.3 Decompose the structured indicators OKR to obtain the departmental OKR list.

[0082] Specifically, the expression is,

[0083]

[0084] OKR k Let Sv be the final OKR value for target department k, Sv be the target value for strategic indicator v, n be the total number of departments participating in the allocation, α be the historical performance adjustment factor, and R be the final OKR value for target department k. k R scores the historical completion rate of target department k. m The sum of historical scores for each department is traversed, where v is a strategic indicator and m is the index of the department being traversed.

[0085] S2.4 Calculate the individual KPI list based on the target values ​​of the department's OKR list and the weighting ratio of employee contribution data.

[0086] Specifically, the expression is,

[0087]

[0088] Among them, KPI h Let F be the KPI target value for employee h. h C represents the job level weight for employee h. h Let h be the historical contribution rate of employee, P be the total number of employees in the target department, p be the index of the total number of employees in the target department, and C be the index of the total number of employees in the target department. p Value of personal contribution;

[0089] S2.5. Link the structured indicators, departmental OKR list, and individual KPI list to obtain a three-level indicator mapping table.

[0090] Furthermore, the strategic objectives in the structured indicators are matched with the departmental objectives in the departmental OKR list through the department ID field, establishing a one-to-many mapping relationship from the strategic decision-making level to the departmental management level. Each departmental objective in the departmental OKR list is matched with the employee performance indicators in the individual KPI list through the employee ID field, establishing a one-to-many mapping relationship from the departmental management level to the grassroots execution level. The pandas library in Python is used to vertically merge the associated data of the three levels, generating a three-level indicator mapping table containing the indicator names of the strategic decision-making level, the OKR descriptions of the departmental management level, and the KPI values ​​of the grassroots execution level.

[0091] S2.6. Generate a quantum state comparison matrix based on a three-level index mapping table, and perform a phase estimation algorithm on the quantum comparison matrix on an IBM quantum processor to obtain quantum state eigenvalues.

[0092] Specifically, the expression is,

[0093]

[0094] Where A is the quantum state comparison matrix, A set of indicators for strategic decision-making. Let ε be the set of indicators at the departmental level, and ε be the set of indicators at the execution level. Here, represents the three-level correlation weights, max(Y) is the weight normalization factor, and ι is the index of the strategic decision-making level indicator. An index for departmental management indicators. This serves as an index for indicators at the grassroots execution level.

[0095] Specifically, the expression is,

[0096]

[0097] Where U(A) is the linear transformation of the quantum state comparison matrix A, B is the number of quantum state eigenvalues, b is the quantum state eigenvalue index, and η b Let θ be the probability amplitude of the characteristic state of the quantum state. b Let θ be the characteristic phase of the quantum state with the b-th quantum state eigenvalue.

[0098] S2.7. Convert the quantum state eigenvalues ​​into strategic decision-making layer weight vectors, departmental layer weight matrices, and execution layer weight tensors.

[0099] Furthermore, the real-valued parts of the quantum state eigenvalues ​​output by the quantum processor are extracted to obtain a set of eigenvectors containing phase information. Based on the predefined hierarchical indexing rules in the quantum state comparison matrix, the eigenvector set is divided into three sub-vectors: strategic decision-making layer eigenvectors, departmental layer eigenvectors, and execution layer eigenvectors. The Gram-Schmidt orthogonalization method is used to normalize each sub-vector to ensure that the sum of the weights at each level is 1. The normalized strategic decision-making layer eigenvectors are converted into weight vectors, the departmental layer eigenvectors are rearranged into a two-dimensional weight matrix according to the department ID, and the execution layer eigenvectors are constructed into a three-dimensional weight tensor according to the employee ID and task ID. During the conversion process, double-precision floating-point numbers are used to retain 15 significant digits to avoid the accumulation of measurement errors inherent in quantum computing. The final output strategic decision-making layer weight vector, departmental layer weight matrix, and execution layer weight tensor are all stored in HDF5 format to retain the complete quantum state eigenvalue mapping relationship.

[0100] S3. Encode the weight vector of the strategic decision-making layer, the weight matrix of the department management layer, and the weight tensor of the grassroots execution layer into qubits, construct a three-body entangled state of the strategic decision-making layer indicators, the department management layer indicators, and the grassroots execution layer indicators, and trigger cross-level collaborative adjustment through quantum measurement to obtain the adjusted weight data.

[0101] S3.1 Encode the strategic decision-making layer weight vector into the amplitude state of a qubit to obtain a quantum state. Load the department management weight matrix onto two qubits through a controlled rotation gate to obtain an entangled state. Load the grassroots execution layer weight tensor as a third-level conditional probability onto three qubits through a multi-control gate to obtain a three-body entangled state.

[0102] Furthermore, the strategic decision-making layer weight vector is mapped onto the state vector of a single qubit using quantum amplitude encoding technology to generate a strategic decision-making layer quantum state. Each ground state probability amplitude corresponds to a component of the original weight vector. Singular value decomposition is performed on the weight matrix of the department management layer, and the left singular vector is used as a control signal to act on two qubits through a controlled rotation gate operation to prepare the entangled state of the department management layer. This makes the joint state probability distribution of the two qubits match the row and column correlation of the weight matrix. After the weight tensor of the grassroots execution layer is expanded, a conditional phase rotation is achieved using a three-qubit multi-control gate circuit. This constructs a three-body entangled state network between the strategic decision-making layer quantum state, the entangled state of the department management layer, and the quantum state of the grassroots execution layer. The quantum circuit design is implemented using the IBM Qiskit framework. The angle parameter of the controlled rotation gate is determined by the singular value of the weight matrix, and the phase parameter of the multi-control gate is calculated from the magnitude of the weight tensor, resulting in the three-body entangled state.

[0103] S3.2. Based on the three-body entangled state, the reduced density matrix is ​​obtained to verify the entanglement entropy, and the verification flag is obtained. The three-body entangled state is measured, and the measurement result statistics table is obtained.

[0104] Furthermore, the density operator of the reduced density matrix is ​​reconstructed using quantum state tomography. The entanglement entropy of each subsystem is obtained using von Neumann entropy, as well as the entanglement entropy of the departmental management-grassroots execution layer. Based on the entanglement entropy, a verification flag is derived. The verified three-body entangled state is then subjected to projection measurement under the obtained basis. The quantum measurement experiment is repeated 1000 times. The joint measurement results of the quantum state of the strategic decision-making layer, the entangled state of the departmental management layer, and the quantum state of the grassroots execution layer are statistically analyzed. A measurement result statistics table containing the frequency of occurrence of each state is generated. During the measurement process, quantum error mitigation technology is used to correct the decoherence effect. The measurement result statistics table records the ratio of the number of measurements of each ground state to the total number of measurements.

[0105] S3.3. Based on the measurement results statistics table, obtain the strategy-department covariance and the department-execution covariance, and obtain the covariance matrix.

[0106] Furthermore, a joint probability distribution analysis was performed on the quantum state measurement values ​​of the strategic decision-making level, the entangled state measurement values ​​of the departmental management level, and the quantum state measurement values ​​of the grassroots execution level in the measurement results statistics table. The covariance of the quantum state measurement values ​​of the strategic decision-making level and the entangled state measurement values ​​of the departmental management level was calculated to obtain the strategic-department covariance. The covariance of the entangled state measurement values ​​of the departmental management level and the quantum state measurement values ​​of the grassroots execution level was then calculated to obtain the department-execution covariance. The strategic-department covariance and the department-execution covariance were arranged in hierarchical order to construct a 2×2 symmetric covariance matrix. The diagonal elements of the matrix are the variances of the measurement values ​​at each level, and the off-diagonal elements are the cross-level covariances. The construction of the covariance matrix was implemented using the Python NumPy library to ensure that the numerical calculation accuracy reaches the double floating-point standard.

[0107] S3.4. Based on the median in the covariance matrix, obtain the critical value of cross-level management association strength.

[0108] Furthermore, the off-diagonal elements of the covariance matrix (strategy-department covariance and department-execution covariance) are sorted, and the median of the sorted numerical sequence is extracted as the critical value for the cross-level management correlation strength. The two covariance values ​​are arranged in ascending order to generate an ordered sequence. When the sequence length is even, the average of the two middle numbers is taken; when the sequence length is odd, the median value is taken directly. The critical value for the cross-level management correlation strength is used to determine whether to trigger weight adjustment. When any covariance value is lower than the critical value, it is determined that the cross-level correlation is insufficient.

[0109] S3.5 When the strategy-department covariance is greater than the critical value of cross-level management correlation strength, cross-level collaborative adjustment is triggered to obtain the adjusted weight data.

[0110] Furthermore, an adjustment coefficient is calculated based on the ratio of the difference between the strategic-department covariance and the critical value. The formula for the adjustment coefficient is (covariance - critical value) / critical value. This adjustment coefficient is used to synchronously scale the strategic decision-making layer weight vector and the department management layer weight matrix. The scaling ratio of the strategic decision-making layer weight vector is 1 + 0.5 adjustment coefficient, and the scaling ratio of the department management layer weight matrix is ​​1 + 0.3 adjustment coefficient. The weight tensor of the grassroots execution layer is not directly adjusted, but is indirectly affected by the change in the weight matrix of the department management layer. The adjusted strategic decision-making layer weight vector and the department management layer weight matrix need to be normalized again to ensure that the sum of the weights of each level remains 1. The normalized weight data and the original grassroots execution layer weight tensor together constitute the adjusted weight dataset.

[0111] S4. Based on the adjusted weighted data, collect multi-source assessment data from the strategic decision-making level, departmental management level, and grassroots execution level from the ERP module, CRM module, and code repository, and perform normalization processing to generate a standardized data table.

[0112] S4.1. Based on the adjusted weight data, a structured dictionary is obtained. For the list of strategic indicators in the structured dictionary, the performance indicator acquisition interface of the ERP module is used to obtain market share performance data and thus strategic decision-making level assessment data.

[0113] Furthermore, by calling the BAPI_KPI_GETLIST interface provided by the ERP module, and passing in the list of strategic indicator names from the structured dictionary as query parameters, the market share performance data of each strategic indicator is obtained. The performance data returned by the interface is matched and associated with the strategic decision-making layer node ID in the structured dictionary to form a strategic decision-making layer assessment data set containing node ID, indicator name, and market share value. Integrity verification is performed during the data matching process. When any strategic indicator is found to have no returned performance data, a default value is automatically filled and an abnormal state is marked. The strategic decision-making layer assessment data set is stored in JSON format, retaining the original timestamp and numerical unit information returned by the ERP interface to ensure that the data can be correctly parsed during subsequent normalization processing. The performance data acquisition process records detailed request parameters and response logs for tracing data sources and handling abnormal situations. The final generated strategic decision-making layer assessment data set serves as an input component of the standardized data table.

[0114] S4.2 Using the department ID list in the structured dictionary, query the quarterly goals and achievement completion rates of each department through the CRM module interface to obtain the performance evaluation data of the department management.

[0115] Furthermore, using the list of department IDs recorded in the structured dictionary as input parameters, the Salesforce API interface SOQL query service provided by the CRM module is called. A query request containing the fields of department ID, quarterly time range, and target type is sent. The API interface returns the list of target values ​​and the list of actual completed values ​​for each department in the specified quarter, obtaining the quarterly target completion rate for each department. This generates a set of performance evaluation data for department management. Each record in the set contains five fields: department ID, target name, target value, actual value, and completion rate, and is sorted in ascending order by department ID, thus obtaining the performance evaluation data for department management.

[0116] S4.3 Based on the execution unit mapping table in the weight dictionary, scan the code version control unit commit logs to form the code contribution density of each execution unit and obtain the basic execution layer assessment data.

[0117] Furthermore, based on the correspondence between employee IDs and code repository accounts recorded in the execution unit mapping table in the weight dictionary, the Git command-line tool is used to execute log analysis commands to scan code commit records within a specified time range. For each employee account in the execution unit, the number of code commits, the number of newly added lines of code, and the number of deleted lines of code are counted to obtain the code contribution density index. The three indices of employee ID, number of code commits, and code contribution density are correlated to obtain the performance evaluation data of the grassroots execution layer.

[0118] S4.4 Normalize the assessment data of the strategic decision-making level, the assessment data of the department management level, and the assessment data of the grassroots execution level to generate a standardized data table.

[0119] Furthermore, the three heterogeneous indicators—market share from the strategic decision-making level assessment data, quarterly target completion rate from the department management level assessment data, and code contribution density from the grassroots execution level assessment data—are converted into standard values ​​within the [0,1] interval using the Min-Max normalization method. The normalization formula for the department management level indicator is (completion rate - minimum completion rate) / (maximum completion rate - minimum completion rate), and the normalization formula for the grassroots execution level indicator is (contribution density - minimum density) / (maximum density - minimum density). The three types of normalized data are then vertically merged according to hierarchical relationships to generate a standardized data table containing three sets of key fields: strategic decision-making level node ID, department management level department ID, and grassroots execution level employee ID.

[0120] S5. Implant the topological insulator algorithm into the standardized data table to obtain a clean data stream.

[0121] S5.1 Add hierarchical labels to the standardized data table to obtain a labeled graph structure.

[0122] Furthermore, based on the hierarchical relationships recorded in the standardized data table, management edges are established between strategic decision-making level nodes and directly subordinate department management level nodes, and execution edges are established between department management level nodes and subordinate grassroots execution level nodes, constructing a labeled directed graph structure. The weight attributes of the edges come from the normalized indicator values ​​recorded in the standardized data table. The weight of the strategic decision-making level-department management level edge is the market share standard value, and the weight of the department management level-grassroots execution level edge is the target completion rate standard value. The graph structure is stored using the DiGraph class of the NetworkX library. Node attributes include node ID, hierarchical label, and original indicator value, while edge attributes include weight value and data source marker. The labeled graph structure is stored serialized in JSON format, preserving complete topological connections and attribute data.

[0123] S5.2 Calculate the invariants of the boundary edges connecting different levels on the labeled graph structure to obtain a list of abnormal edges.

[0124] Specifically, the expression is,

[0125]

[0126] Where Z2 is the list of abnormal edges, and φ is the cross-level management association edge. Λ represents the parent level node, and Λ represents the child level node. This represents the edge node weight.

[0127] S5.3. Abnormal edges are obtained by analyzing the three-level index mapping table. Abnormal edges are removed according to the abnormal edge list to obtain the purified adjacency matrix. Spectral clustering is then performed to obtain the three-level node ID list.

[0128] Furthermore, by analyzing the correlation between indicators recorded in the three-level indicator mapping table, edges with weight values ​​lower than 30% of the median weight of edges in the same level are identified as abnormal edges, generating an abnormal edge list. Based on the edge identifiers in the abnormal edge list, the corresponding row and column elements are removed from the adjacency matrix of the labeled graph structure to obtain a purified adjacency matrix. A normalized Laplace matrix transformation is performed on the purified adjacency matrix to obtain the first three eigenvectors as the spectral feature space. The K-means algorithm is used to cluster the nodes, with the number of clusters set to 3, corresponding to the three levels of strategic decision-making, departmental management, and grassroots execution. The node IDs of each cluster in the clustering results constitute a three-level node ID list. The node IDs in the list are stored according to the original level labels. During the spectral clustering process, a random initialization seed and a 10-iteration convergence strategy are used to ensure the stability of the clustering results, resulting in the three-level node ID list.

[0129] S5.4 Filter the list of third-level node IDs to obtain a clean data stream.

[0130] Furthermore, based on the node identifiers recorded in the three-level node ID list, the complete data records of the corresponding nodes are extracted from the standardized data table. These records include the market share standard value of strategic decision-making level nodes, the target completion rate standard value of department management level nodes, and the code contribution density standard value of grassroots execution level nodes. The data records of the three types of nodes are then reorganized according to their hierarchical relationship, with strategic decision-making level node data as the root node, department management level node data as intermediate nodes, and grassroots execution level node data as leaf nodes. A tree-structured data stream is constructed, in which each node contains four fields: node ID, hierarchical label, indicator value, and parent node ID. The data is stored using Protocol Buffers serialization format. During the data stream construction process, hierarchical consistency checks are implemented to ensure that each department management level node has one and only one strategic decision-making level parent node, and each grassroots execution level node has one and only one department management level parent node, ultimately generating a clean data stream.

[0131] S6. Input the clean data stream into the hierarchical evaluation model, use DEA data envelopment analysis for the strategic decision-making level, use the entropy weight-TOPSIS algorithm for the department management level, and use XGBoost evaluation for the grassroots execution level to generate an assessment score table with hierarchical labels.

[0132] S6.1 Based on the pure data stream, the data is split into strategic decision-making layer data, departmental layer data and execution layer data according to the hierarchical label field. The resource input and performance indicator achievement values ​​of each strategic indicator are extracted from the strategic decision-making layer data to construct the decision-making unit matrix and obtain the strategic decision-making layer DEA.

[0133] Furthermore, based on the node data marked as "strategy" in the clean data stream, the resource input and performance indicator achievement fields are extracted for each strategic decision-making level node. Resource input includes three dimensions: capital input, human resource input, and technology input. Performance indicator achievement includes two dimensions: market share achievement and profit margin achievement. Each strategic decision-making level node is treated as a decision-making unit, with resource input as the input variable and performance indicator achievement as the output variable. A decision-making unit matrix is ​​constructed, where rows correspond to different strategic decision-making level nodes, columns correspond to input and output variables, and matrix element values ​​are standardized values ​​for each node. The number of rows and columns of the indicator value and the decision unit matrix are the sum of the number of nodes in the strategic decision layer and the number of input and output variables, respectively. Missing values ​​in the matrix are filled with the median of the same indicator. The completed decision unit matrix serves as the input data for the strategic decision layer DEA and is used for subsequent efficiency value calculation. The decision unit matrix is ​​stored in CSR sparse matrix format to optimize storage and computation efficiency under large-scale data. Data range verification is implemented during matrix construction to ensure that all input and output variable values ​​are within the range of [0,1]. The final generated strategic decision layer DEA input matrix contains a complete mapping relationship between node IDs and indicator values, and retains the hierarchical identification information in the original pure data stream.

[0134] S6.2 Solve the linear programming problem of the strategic decision layer DEA on a quantum processor to form the relative efficiency value of strategic indicators and obtain the strategic efficiency score.

[0135] Furthermore, the linear programming problem of the strategic decision-making layer DEA is transformed into a quadratic unconstrained binary optimization form. Quantum annealing computation is performed on the D-Wave quantum processor. The quantum annealing process solves the objective function minimization problem. The objective function includes the sum of squares of input and output constraint terms and an efficiency maximization term. Hamiltonian encoding adopts a one-hot encoding strategy. The low-energy state solution returned by the quantum annealer is converted into an efficiency value through classical post-processing. The efficiency value θ∈[0,1] of each strategic decision-making layer node represents its relative efficiency level. The efficiency value calculation results are stored in association with the node ID to generate a list of strategic efficiency scores. The efficiency scores in the list retain 6 decimal places of precision. Decoherence time monitoring is implemented during the quantum computing process to obtain the strategic efficiency score.

[0136] S6.3. Based on departmental data, information entropy and dynamic weights are derived to form an entropy-weight matrix and a departmental weight matrix.

[0137] Specifically, the expression is,

[0138]

[0139] in, For the first Information entropy of the indicator χ represents the proportion of target department k in indicator σ, where χ is the number of departments. For indexing indicators,

[0140]

[0141] in, For the first The weight of each indicator.

[0142] S6.4. Perform weighted standardization based on the department weight matrix to obtain the distance between the positive and negative ideal solutions and generate department rankings.

[0143] Furthermore, the entropy weight matrix, department weight matrix, and standardized value matrix of departmental data are subjected to Hadamard product operation to obtain a weighted standardized decision matrix. In the weighted standardized decision matrix, the optimal and worst values ​​of each indicator are determined as positive and negative ideal solutions. The positive ideal solution consists of the maximum value of the benefit-type indicator and the minimum value of the cost-type indicator, and the negative ideal solution consists of the minimum value of the benefit-type indicator and the maximum value of the cost-type indicator. The Euclidean distance from each department to the positive ideal solution and the Euclidean distance to the negative ideal solution are obtained. The distance formula is the square root of the sum of squared differences. Departments are sorted according to relative proximity. The larger the proximity value, the better the overall performance of the department. The department ranking results are arranged in descending order of proximity, generating a department management ranking list containing department ID, proximity value, and ranking, thus generating the department ranking.

[0144] S6.5. Extract the average number of submissions and the centrality of the code submission network from the raw data of code submission records and task response times in the execution layer data to obtain the feature matrix execution layer matrix. Use the feature matrix execution layer matrix to train the employee performance ranking model to obtain the XGBoost model.

[0145] Furthermore, the code submission count and task response time for each employee are extracted from the execution layer data to obtain the average submission volume within a 30-day sliding window. An employee collaboration network graph is constructed based on the code version control logs, and the feature vector centrality of each employee node is calculated. The average submission volume and centrality are used as feature vectors, and combined with the employee's historical performance score to form a feature matrix execution layer matrix. The feature matrix execution layer matrix is ​​used to train an XGBoost ranking model, with the objective function defined as pairwise ranking loss to optimize the relative ranking of employee performance. The XGBoost model training adopts an early stopping strategy, terminating training when the validation set NDCG index does not improve for three consecutive rounds. The trained XGBoost model can predict the performance score of new employee samples and generate an execution layer performance score list containing employee ID, predicted score, and ranking. During model training, feature importance scores are recorded to ensure a balance between the contribution of code contribution features and task response features. The final generated XGBoost model serves as an execution layer evaluation tool, which, together with the strategic efficiency score and department management ranking, constitutes a hierarchical assessment score table, resulting in the XGBoost model.

[0146] S6.6. Based on the XGBoost model, predict the performance scores of all employees to obtain the performance scores of the execution level.

[0147] Furthermore, the trained XGBoost model is loaded into the prediction environment. The input features include the average number of submissions and the code submission network centrality feature of each employee in the execution layer matrix. The XGBoost model outputs the raw predicted score for each employee. The score is linearly mapped to the [0,100] interval through Min-Max normalization to obtain the execution layer performance score. The execution layer performance score list contains three fields: employee ID, normalized score, and percentile ranking, and is arranged in descending order of score. During the prediction process, score consistency verification is implemented. When the difference between multiple prediction results for the same employee exceeds 5 points, the review mechanism is automatically triggered. The final generated execution layer performance score list is merged with the strategic efficiency score and the department management ranking to form a complete assessment score table with hierarchical identification. The prediction process records the complete mapping relationship between the model input features and the output scores.

[0148] S6.7. Integrate strategic efficiency scores, departmental rankings, and execution-level performance scores to obtain an assessment score table with hierarchical labels.

[0149] Furthermore, the efficiency values ​​of strategic decision-making level nodes in the strategic efficiency score list, the department proximity values ​​in the department management ranking list, and the normalized employee scores in the execution level performance score list are matched according to hierarchical relationships. The efficiency values ​​of strategic decision-making level nodes are weighted and summed with the proximity values ​​of their direct department management nodes, with the weights using the corresponding components in the strategic decision-making level weight vector after quantum collaboration adjustment. The comprehensive scores of department management nodes are weighted and summed with the performance scores of their subordinate grassroots execution level employees, with the weights using the corresponding values ​​in the department management weight matrix. The fused score table contains three sets of identifier fields: strategic decision-making level node ID, department management department ID, and grassroots execution level employee ID, as well as the original scores, weighted scores, and final assessment scores for each level. The score table is organized in a tree structure of strategic decision-making level - department management level - grassroots execution level, using Parquet columnar storage format, retaining a complete hierarchical association index. During the score fusion process, score range verification is implemented to ensure that the weighted assessment scores fall within the range of [0,100], resulting in an assessment score table with hierarchical identifiers.

[0150] This embodiment also provides a multi-dimensional, hierarchical performance evaluation management system, including:

[0151] The data acquisition module obtains an organizational structure tree diagram and converts the hierarchical relationships between the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database.

[0152] Based on the hierarchical database, the annual strategic goals are extracted, broken down into quarterly OKRs for departmental management, and then further divided into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The AHP (Analytic Hierarchy Process) is used to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor.

[0153] The weight allocation module encodes the weight vector of the strategic decision-making layer, the weight matrix of the department management layer, and the weight tensor of the grassroots execution layer into qubits, constructs a three-body entangled state of the strategic decision-making layer indicators, the department management layer indicators, and the grassroots execution layer indicators, and triggers cross-level collaborative adjustment through quantum measurement to obtain the adjusted weight data;

[0154] Cross-level modules collect multi-source assessment data from the ERP module, CRM module, and code repository, including strategic decision-making level, department management level, and grassroots execution level, based on the adjusted weighted data, and perform normalization processing to generate standardized data tables.

[0155] The data cleaning module embeds a topological insulator algorithm into a standardized data table to obtain a clean data stream.

[0156] The tiered evaluation module inputs clean data streams into the tiered evaluation model. It uses DEA data envelopment analysis for the strategic decision-making level, the entropy weight-TOPSIS algorithm for the department management level, and XGBoost for the grassroots execution level, generating an assessment score table with tiered labels.

[0157] This embodiment also provides a computer device applicable to the multi-dimensional hierarchical work performance evaluation and management method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the multi-dimensional hierarchical work performance evaluation and management method proposed in the above embodiment.

[0158] The computer device can be a terminal, comprising a processor, memory, communication 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 communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0159] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the multi-dimensional hierarchical work assessment management method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0160] In summary, this invention transforms the organizational structure tree diagram into a hierarchical database and constructs a three-level indicator mapping table. It employs the Analytic Hierarchy Process (AHP) for weight allocation, generating a weight vector for the strategic decision-making layer, a weight matrix for departmental management, and a weight tensor for the grassroots execution layer. Furthermore, it constructs a three-body entangled state using qubit encoding to achieve cross-level collaborative adjustment. Based on the adjusted weight data, it collects assessment data from a multi-source system, purifies it using a topological insulator algorithm, and inputs it into a hierarchical evaluation model. Finally, it generates an assessment score table with hierarchical identifiers. This method innovatively applies quantum computing and topological principles to the field of performance evaluation, solving the technical challenges of traditional methods in cross-level dynamic collaboration and multi-source data fusion.

[0161] It should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multi-dimensional, hierarchical performance evaluation management method, characterized in that: include, Obtain the organizational structure tree diagram and convert the hierarchical relationships between the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database; Based on the hierarchical database, the annual strategic goals are extracted, broken down into quarterly OKRs for departmental management, and then further divided into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The AHP (Analytic Hierarchy Process) is used to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor. The weight vector of the strategic decision-making level, the weight matrix of the department management level, and the weight tensor of the grassroots execution level are encoded into qubits. A three-body entangled state of the strategic decision-making level indicators, the department management level indicators, and the grassroots execution level indicators is constructed. Cross-level collaborative adjustment is triggered by quantum measurement to obtain the adjusted weight data. Based on the adjusted weighted data, multi-source assessment data from the strategic decision-making level, department management level, and grassroots execution level were collected from the ERP module, CRM module, and code repository and normalized to generate a standardized data table. By embedding the topological insulator algorithm into a standardized data table, a clean data stream is obtained. The clean data stream is input into the hierarchical evaluation model. DEA data envelopment analysis is used for the strategic decision-making level, the entropy weight-TOPSIS algorithm is used for the department management level, and XGBoost is used for the grassroots execution level to generate an assessment score table with hierarchical labels.

2. The multi-dimensional hierarchical work assessment management method of claim 1, wherein: Obtaining an organizational structure tree diagram and converting the hierarchical relationships between the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database includes the following steps: Export the organizational structure tree diagram through the enterprise OA module, analyze the hierarchical depth of the organizational structure tree diagram, and obtain the affiliation relationship between the strategic decision-making level, departmental management level, and grassroots execution level; The strength of hierarchical associations is quantified by a dynamic weighted adjacency matrix algorithm, and the strategic decision-making level, departmental management level, and grassroots execution level are transformed into a hierarchical database by combining the Cypher statement of the Neo4j graph database.

3. The multi-dimensional hierarchical work assessment management method of claim 2, wherein: Based on a hierarchical database, annual strategic goals are extracted and broken down into quarterly OKRs for departmental management, and further subdivided into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The Analytic Hierarchy Process (AHP) is then used to assign weights to this three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor. This process includes the following steps: The annual targets of strategic decision-making level nodes are queried from the hierarchical database to obtain the original strategic targets. Quantitative indicators are extracted from the original strategic targets using regular expressions to obtain structured indicators. Based on the actual completion rate data of historical strategic indicators in the CRM module, calculate the historical completion rate of each department for each item in the structured indicators, and obtain the departmental capability score; The structured OKR indicators are broken down to obtain a departmental OKR list; Based on the weighting ratio of the department's OKR list target values ​​and employee contribution data, an individual KPI list is calculated. By linking structured indicators, departmental OKR lists, and individual KPI lists, a three-level indicator mapping table is obtained; A quantum state comparison matrix is ​​generated based on a three-level index mapping table. A phase estimation algorithm is then performed on the quantum comparison matrix on an IBM quantum processor to obtain quantum state eigenvalues. The quantum state eigenvalues ​​are converted into a weight vector for the strategic decision-making layer, a weight matrix for the departmental management layer, and a weight tensor for the grassroots execution layer.

4. The multi-dimensional, hierarchical performance evaluation management method as described in claim 3, characterized in that: The weight vectors of the strategic decision-making level, the weight matrix of the departmental management level, and the weight tensor of the grassroots execution level are encoded into qubits. A three-body entangled state is constructed between the strategic decision-making level indicators, the departmental management level indicators, and the grassroots execution level indicators. Cross-level collaborative adjustment is triggered by quantum measurement to obtain the adjusted weight data, including the following steps. The strategic decision-making layer weight vector is encoded as the amplitude state of a qubit to obtain a quantum state. The weight matrix of the department management layer is loaded onto two qubits through a controlled rotation gate to obtain an entangled state. The weight tensor of the grassroots execution layer is used as a third-level conditional probability and loaded onto three qubits through a multi-control gate to obtain a three-body entangled state. Based on the reduced density matrix obtained from the three-body entangled state, the entanglement entropy is verified, the verification flag is obtained, the three-body entangled state is measured, and the measurement result statistics table is obtained. Based on the statistical table of measurement results, the strategy-department covariance and the department-execution covariance are obtained, and the covariance matrix is ​​obtained. Based on the median in the covariance matrix, the critical value for the strength of cross-level management associations is obtained; When the strategy-department covariance exceeds the critical value of cross-level management correlation strength, cross-level collaborative adjustment is triggered, resulting in adjusted weight data.

5. The multi-dimensional, hierarchical performance evaluation management method as described in claim 4, characterized in that: Based on the adjusted weighted data, multi-source performance evaluation data from the strategic decision-making level, departmental management level, and grassroots execution level are collected from the ERP module, CRM module, and code repository, and normalized to generate standardized data tables. This includes the following steps: Based on the adjusted weighted data, a structured dictionary is obtained. Then, the performance indicator acquisition interface of the ERP module is used to obtain market share performance data from the list of strategic indicators in the structured dictionary, thus obtaining the strategic decision-making level assessment data. Using the list of department IDs in the structured dictionary, the quarterly targets and achievement rates of each department are queried through the CRM module interface to obtain the performance evaluation data of the department management. Based on the execution unit mapping table in the structured dictionary, the code version control unit commit logs are scanned to form the code contribution density of each execution unit, and the basic execution layer assessment data is obtained. The assessment data of the strategic decision-making level, the department management level, and the grassroots execution level are normalized to generate standardized data tables.

6. The multi-dimensional hierarchical performance evaluation management method as described in claim 5, characterized in that: By embedding the topological insulator algorithm into a standardized data table, a clean data stream is obtained. Includes the following steps, Add hierarchical labels to the standardized data table to obtain a labeled graph structure; Calculate the invariants of the boundary edges connecting different levels on a labeled graph structure to obtain a list of anomalous edges; Abnormal edges are obtained by analyzing the three-level index mapping table. Abnormal edges are removed according to the abnormal edge list to obtain the purified adjacency matrix. Then, spectral clustering is performed to obtain the three-level node ID list. The list of third-level node IDs is filtered to obtain a clean data stream.

7. The multi-dimensional hierarchical performance evaluation management method as described in claim 6, characterized in that: The clean data stream is input into the hierarchical evaluation model. DEA (Data Envelopment Analysis) is used for the strategic decision-making level, the Entropy Weight-TOPSIS algorithm is used for the departmental management level, and XGBoost is used for the grassroots execution level. This generates a performance score table with hierarchical labels, including the following steps. Based on the pure data stream, the data is split into strategic decision-making layer data, departmental layer data and execution layer data according to the hierarchical label field. The resource input and performance indicator achievement values ​​of each strategic indicator are extracted from the strategic decision-making layer data to construct a decision-making unit matrix and obtain the strategic decision-making layer DEA. Solve the linear programming problem of the DEA (Decision Analysis) layer of the strategic decision-making level on a quantum processor to form the relative efficiency value of the strategic indicators and obtain the strategic efficiency score; Based on departmental data, information entropy and dynamic weights are derived to form an entropy-weight matrix and a departmental weight matrix. Based on the department weight matrix, a weighted standardization process is performed to obtain the distance between the positive and negative ideal solutions, and a department ranking is generated. From the raw data of code submission records and task response times in the execution layer data, the average number of submissions and the centrality of the code submission network are extracted to obtain the feature matrix execution layer matrix. The feature matrix execution layer matrix is ​​used to train the employee performance ranking model to obtain the XGBoost model. The performance scores of all employees are predicted based on the XGBoost model to obtain the performance scores of the execution level. By integrating strategic efficiency scores, departmental rankings, and performance scores at the execution level, an assessment score table with hierarchical labels is obtained.

8. A multi-dimensional hierarchical performance evaluation management system, based on the multi-dimensional hierarchical performance evaluation management method according to any one of claims 1 to 7, characterized in that: include, The data acquisition module obtains an organizational structure tree diagram and converts the hierarchical relationships between the strategic decision-making level, departmental management level, and grassroots execution level into a hierarchical database. The decomposition module extracts annual strategic goals from a hierarchical database, decomposes them into quarterly OKRs for departmental management, and further subdivides them into monthly KPIs for the grassroots execution level, forming a three-level indicator mapping table. The AHP (Analytic Hierarchy Process) is used to assign weights to the three-level indicator mapping table, generating a strategic decision-making level weight vector, a departmental management level weight matrix, and a grassroots execution level weight tensor. The weight allocation module encodes the weight vector of the strategic decision-making layer, the weight matrix of the department management layer, and the weight tensor of the grassroots execution layer into qubits, constructs a three-body entangled state of the strategic decision-making layer indicators, the department management layer indicators, and the grassroots execution layer indicators, and triggers cross-level collaborative adjustment through quantum measurement to obtain the adjusted weight data; Cross-level modules collect multi-source assessment data from the ERP module, CRM module, and code repository, including strategic decision-making level, department management level, and grassroots execution level, based on the adjusted weighted data, and perform normalization processing to generate standardized data tables. The data cleaning module embeds a topological insulator algorithm into a standardized data table to obtain a clean data stream. The tiered evaluation module inputs clean data streams into the tiered evaluation model. It uses DEA data envelopment analysis for the strategic decision-making level, the entropy weight-TOPSIS algorithm for the department management level, and XGBoost for the grassroots execution level, generating an assessment score table with tiered labels.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the multi-dimensional hierarchical work assessment and management method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the multi-dimensional hierarchical work assessment and management method according to any one of claims 1 to 7.

Images