Human resource planning and decision support method and system for group-oriented enterprises

Abstract

This invention discloses a method and system for human resource planning and decision support for group enterprises, relating to the field of human resource management. Based on the group enterprise's organizational structure information, human resource allocation information, and job competency requirements information, this method constructs a digital twin model of the group's human resources. It then simulates and extrapolates organizational adjustments or human resource planning schemes, generating planning impact assessment results. Building upon this, it constructs local human resource allocation objective functions for each subsidiary and, combined with global constraints at the group level, collaboratively solves these functions through a distributed game optimization approach, resulting in group-level human resource planning decision results. Simultaneously, it reliably stores the planning decision results and their formation process, and updates the model and objective functions based on execution feedback after decision implementation, forming a closed-loop optimization mechanism. This method improves the dynamic adaptability, collaboration, and traceability of decision results in group human resource planning.

Description

Technical Field

[0001] This specification relates to the field of human resource management; more specifically, this application relates to a method and system for human resource planning and decision support for group enterprises. Background Technology

[0002] As large corporations expand and their organizational structures become increasingly complex, human resource management is gradually shifting from internal management within a single enterprise to cross-level and cross-subsidiary planning and collaborative decision-making. However, existing technologies largely focus on static personnel information management or post-event statistical analysis, making them ill-suited to the frequent adjustments in group organizational structures. This leads to a disconnect between human resource planning and the actual organizational structure, resulting in poor implementation of planned solutions. Furthermore, during the unified planning process across the group, the human resource data of each subsidiary involves sensitive information, creating a conflict between cross-entity data sharing and privacy protection, making it difficult to form complete and accurate decision-making basis at the group level. In addition, existing decision-making processes generally rely on lagging data compiled manually, lacking the ability to assess the impact of changes in permissions and business collaborations in real time, resulting in insufficient scientific rigor and timeliness in decision-making. After implementation, there is also a lack of reliable traceability and closed-loop verification mechanisms, making it difficult to quantitatively evaluate and continuously optimize the decision-making process and its effectiveness.

[0003] Therefore, it is necessary to propose a human resource planning and decision support method for group enterprises to solve some of the above-mentioned problems. Summary of the Invention

[0004] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. This summary section is not intended to limit the key and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] Firstly, this application proposes a human resource planning and decision support method for group enterprises, including: Obtain information on the group company's organizational structure, human resource allocation, and job competency requirements. Based on the organizational structure information, human resource allocation information, and job competency requirements of the aforementioned group companies, a digital twin model of the group's human resources is constructed. This digital twin model is used to depict the human resource mapping relationship between the group level, subsidiary level, and job level. Based on the aforementioned digital twin model of the group's human resources, it receives organizational restructuring events or human resources planning requests, and simulates and extrapolates the corresponding organizational restructuring plan or human resources allocation plan in the aforementioned digital twin model of the group's human resources to generate planning impact assessment results. Based on the above planning impact assessment results, construct the local human resource allocation objective function for each subsidiary; By combining the overall constraints at the group level, the objective function of local human resource allocation is solved collaboratively through distributed game optimization, resulting in a group-level human resource planning decision. The above-mentioned group-level human resource planning decision-making results and their corresponding inference parameters and solution process information shall be reliably preserved; After the aforementioned group-level human resources planning and decision-making results are implemented, the aforementioned group human resources digital twin model and the aforementioned local human resources allocation objective function are updated based on the implementation feedback information.

[0006] In one feasible implementation, the above-mentioned construction of a digital twin model of the group's human resources based on the group's organizational structure information, human resource allocation information, and job competency requirements information includes: The above organizational structure information is parsed hierarchically to generate group-level organizational nodes, subsidiary-level organizational nodes, and job-level organizational nodes; Associate personnel size constraint parameters, job competency matching constraint parameters, and permission boundary constraint parameters with each of the above organizational nodes to form an organizational node constraint set; Based on the hierarchical and business collaboration relationships among the aforementioned organizational nodes, human resource mapping rules are established to constrain the above simulation and deduction. Based on the aforementioned organizational nodes, the aforementioned set of organizational node constraints, and the aforementioned human resource mapping rules, the aforementioned digital twin model of the group's human resources is generated.

[0007] In one feasible implementation, the specific construction steps of the above objective function and the above global constraints include: Based on the above planning impact assessment results, objective functions for each subsidiary are constructed to characterize the rationality of personnel size, job matching degree, or human resource cost constraints. Based on the group's overall strategic goals and resource boundary conditions, a group-level global constraint function is constructed to limit the feasible solution space of the objective functions of the aforementioned subsidiaries; among them, the objective functions of the aforementioned subsidiaries and the aforementioned global constraint function together constitute the basis for solving the distributed game optimization problem.

[0008] In one feasible implementation, the aforementioned local human resource allocation objective function is collaboratively solved using a distributed game optimization method to form a group-level human resource planning decision result, including: Under the premise of satisfying the above global constraint function, an iterative update mechanism is used to solve the strategy variables corresponding to the objective function of each of the above subsidiaries in a distributed manner. In each iteration, the pre-set equilibrium convergence condition is determined based on the changes in the above-mentioned strategy variables. When the above equilibrium convergence condition is met, the iterative solution is stopped, and the corresponding strategy combination is taken as the decision result of the above group-level human resource planning.

[0009] In one feasible implementation, the aforementioned reliable notarization of the group-level human resource planning decision results and their corresponding derivation parameters and solution process information includes: The above-mentioned group-level human resource planning decision-making results, the above-mentioned inference parameters, and the above-mentioned solution process information are summarized and extracted to generate evidence data. The aforementioned evidence data will be written into an immutable evidence storage medium to form a decision evidence record corresponding to the aforementioned group-level human resources planning decision results. By recording the aforementioned decision-making evidence, the formation process of the above-mentioned group-level human resource planning decision-making results can be traced and verified.

[0010] In one feasible implementation, after the execution of the aforementioned group-level human resource planning and decision-making results, the aforementioned group human resource digital twin model and the aforementioned local human resource allocation objective function are updated based on the execution feedback information, including: Collect information on actual organizational structure changes and human resource allocation results generated during the implementation of the aforementioned group-level human resource planning and decision-making results, and use this information as the aforementioned implementation feedback information; By comparing and analyzing the above implementation feedback information with the above planning impact assessment results, a planning deviation index is generated. When the above-mentioned planning deviation index exceeds the preset threshold, the organizational node constraint parameters in the above-mentioned group human resources digital twin model or the above-mentioned local human resources configuration objective function are corrected, and the above-mentioned distributed game optimization is re-executed based on the corrected model.

[0011] In one feasible implementation, before modifying the constraint parameters in the aforementioned group human resources digital twin model or the aforementioned local human resources allocation objective function, the method further includes: The above planning deviation indicators are judged by deviation type to distinguish whether the above planning deviation indicators are caused by mismatch of organizational structure constraints or by deviation of human resource allocation optimization goals; When the above planning deviation indicators are determined to be of the organizational structure constraint mismatch type, priority will be given to correcting the personnel size constraint parameters, job competency matching constraint parameters or authority boundary constraint parameters in the above group human resources digital twin model. When the above planning deviation indicators are determined to be of the type of human resource allocation optimization target deviation, the objective function parameters in the above local human resource allocation objective function will be corrected first. Based on the above deviation type determination results, the execution order or combination of the above constraint parameter correction and the above objective function parameter correction is determined.

[0012] In one feasible implementation, the above-mentioned deviation type determination of the planning deviation index includes: Based on the differences in the execution feedback information and the planning impact assessment results, a structural deviation analysis was conducted on the planning deviation indicators. When the above-mentioned differences in distribution characteristics are mainly concentrated in the range of organizational level changes, the range of deviation in job staffing, or the range of authority coverage, the above-mentioned planning deviation indicators are judged as organizational structure constraint mismatch types. When the above-mentioned differences in distribution characteristics are mainly concentrated in the range of changes in personnel allocation ratio, job matching degree, or human resource cost weight, the above-mentioned planning deviation indicators are judged as the type of deviation from the human resource allocation optimization target.

[0013] In one feasible implementation, after determining the execution order or combination of the constraint parameter correction and the objective function parameter correction based on the above deviation type determination result, the method further includes: The above-mentioned constraint parameter correction or the above-mentioned objective function parameter correction is carried out in stages according to the preset correction step size or correction ratio. After each correction phase is completed, the above planning deviation index is recalculated, and it is determined whether the preset stability conditions are met. When the above stability conditions are met, the subsequent correction phase is terminated.

[0014] Secondly, this invention also proposes a human resource planning and decision support method for group enterprises, including: The acquisition unit is used to acquire information on the group company's organizational structure, human resource allocation, and job competency requirements. The first building unit is used to build a digital twin model of the group's human resources based on the organizational structure information, human resource allocation information and job competency requirements information of the aforementioned group enterprises. The aforementioned digital twin model of the group's human resources is used to depict the human resource mapping relationship between the group level, subsidiary level and job level. The simulation unit is used to receive organizational restructuring events or human resource planning requests based on the aforementioned group human resource digital twin model, and to simulate and extrapolate the corresponding organizational restructuring plan or human resource allocation plan in the aforementioned group human resource digital twin model to generate planning impact assessment results. The second building unit is used to construct the local human resource allocation objective function for each subsidiary based on the above planning impact assessment results; The solution unit is used to combine the global constraints at the group level and collaboratively solve the above-mentioned local human resource allocation objective function through distributed game optimization to form the group-level human resource planning decision results. The evidence storage unit is used to reliably store the above-mentioned group-level human resource planning decision results and their corresponding inference parameters and solution process information. The update unit is used to update the aforementioned group-level human resources digital twin model and the aforementioned local human resources allocation objective function based on the execution feedback information after the aforementioned group-level human resources planning and decision-making results are implemented.

[0015] In summary, this embodiment constructs a digital twin model of group human resources, unifying the human resource structure at the group, subsidiary, and job levels. Within this model, it simulates organizational restructuring events or human resource planning requests, enabling human resource planning to assess the impact of different organizational adjustment or configuration schemes in a virtual environment. Compared to existing technologies that rely on post-event manual adjustments or static rule configurations, this embodiment achieves synchronous mapping and conflict prediction between the planned scheme and the actual organizational structure when the organizational structure dynamically changes. This significantly improves the adaptability and implementability of the planning scheme and reduces the risk of human resource allocation conflicts caused by organizational adjustments. Based on the generated planning impact assessment results, this embodiment constructs local human resource allocation objective functions for each subsidiary and, combined with global constraints at the group level, collaboratively solves the objective functions of each subsidiary through distributed game optimization. In this process, subsidiaries do not need to directly report complete raw human resource data to participate in group-level planning decisions. This achieves unified planning and collaborative optimization at the group level while ensuring subsidiary data security and privacy protection, effectively alleviating the contradiction between cross-subsidiary data sharing and privacy protection in existing technologies. This embodiment transforms the human resource planning problem into a computable objective function and constraints, and employs a distributed game optimization mechanism for solution. This enables planning decisions to automatically coordinate strategies and determine convergence under multi-objective and multi-constraint conditions. Compared to existing technologies that rely on manually summarizing lagging data for decision-making, this embodiment significantly improves the real-time nature and scientific rigor of human resource decisions, allowing the group to quickly formulate reasonable planning decisions when facing business changes, authority adjustments, or changes in resource boundaries. After generating group-level human resource planning decision results, this embodiment reliably stores the decision results and their corresponding derivation parameters and solution process information, making the formation process of planning decisions traceable and verifiable. By managing the evidence of the decision basis and solution process, it avoids the problems of unclear sources of planning decisions, tampered parameters, or difficulty in verifying the execution basis in existing technologies, providing reliable data support for subsequent audits, compliance management, and decision review. After the planning decision is executed, this embodiment collects actual changes in organizational structure and human resource allocation execution results as execution feedback information, and updates the group's human resource digital twin model and local human resource allocation objective function based on the execution feedback. Compared with existing technologies that lack the ability to evaluate the effects and continuously optimize after planning is completed, this embodiment can continuously correct the model parameters and optimization objectives, so that the human resource planning model can continuously align with the actual operating status of the group, providing a more accurate and reliable basis for subsequent planning decisions.In summary, this embodiment, through digital twin modeling, distributed collaborative optimization, trusted evidence storage, and closed-loop feedback updates, has realized the transformation of human resource planning for group enterprises from static management to dynamic deduction, collaborative decision-making, and continuous optimization. It has significant technical effects in improving planning accuracy, decision-making efficiency, execution controllability, and result credibility, and can effectively overcome the shortcomings of existing technologies in group human resource planning and decision support.

[0016] Other advantages, objectives and features of this application will be apparent in part from the description which follows, and in part from what those skilled in the art will understand through study and practice of this application. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0018] Figure 1 A flowchart illustrating a human resource planning and decision support method for group enterprises, provided as an embodiment of this application; Figure 2 A schematic diagram of a method for constructing a digital twin model of group human resources is provided in this application embodiment; Figure 3 A schematic flowchart illustrating a specific method for constructing an objective function and the aforementioned global constraints, provided in an embodiment of this application; Figure 4 A schematic diagram of a method for generating group-level human resource planning decision results is provided in an embodiment of this application; Figure 5 This is a schematic diagram of a trusted evidence preservation method provided in an embodiment of this application; Figure 6 A schematic diagram of a method for updating a group human resources digital twin model and the above-mentioned local human resources allocation objective function provided in this application embodiment; Figure 7 This application provides a schematic diagram of a method for modifying model parameters and objective function before implementation. Figure 8 This is a flowchart illustrating a method for determining the type of deviation indices of the aforementioned planning deviation indicators, provided in an embodiment of this application. Figure 9 This application provides a schematic diagram of a method for modifying model parameters and objective function according to an embodiment of the present application. Figure 10This is a schematic diagram of a human resource planning and decision support structure for group enterprises, provided as an embodiment of this application. Detailed Implementation

[0019] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The technical solutions of the embodiments of this application will now be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0020] Please see Figure 1 This is a flowchart illustrating a human resource planning and decision support method for group enterprises provided in this application embodiment, which may specifically include: S110. Obtain information on the group company's organizational structure, human resource allocation, and job competency requirements. S120. Based on the organizational structure information, human resource allocation information and job competency requirements of the aforementioned group enterprises, a digital twin model of the group's human resources is constructed. The aforementioned digital twin model of the group's human resources is used to depict the human resource mapping relationship between the group level, subsidiary level and job level. S130. Based on the above-mentioned group human resources digital twin model, receive organizational structure adjustment events or human resources planning requests, and simulate and deduce the corresponding organizational adjustment plan or human resources allocation plan in the above-mentioned group human resources digital twin model to generate planning impact assessment results. S140. Based on the above planning impact assessment results, construct the local human resource allocation objective function for each subsidiary. S150. Combining the global constraints at the group level, the objective function of the above-mentioned local human resource allocation is solved collaboratively through distributed game optimization to form the group-level human resource planning decision results. S160. The above-mentioned group-level human resource planning decision-making results and their corresponding inference parameters and solution process information shall be reliably stored. S170. After the above-mentioned group-level human resource planning decision results are implemented, the above-mentioned group human resource digital twin model and the above-mentioned local human resource allocation objective function are updated based on the implementation feedback information.

[0021] For example, the first step is to acquire the group company's organizational structure information, human resource allocation information, and job competency requirements information. The organizational structure information includes the organizational relationships at the group level, subsidiary level, and job level. The human resource allocation information includes the personnel size, personnel structure, and job occupancy of each organizational unit. The job competency requirements information is used to characterize the competency elements and qualification requirements for different positions. This unified collection of information provides foundational data support for subsequent human resource planning.

[0022] After acquiring the aforementioned basic information, a digital twin model of the group's human resources is constructed based on the group's organizational structure, human resource allocation, and job competency requirements. This digital twin model abstractly models different levels of organizational units within the group and establishes mapping relationships between organizational units, positions, and human resources. This allows the group's actual human resource structure to be fully reproduced in a virtual environment, thus providing a platform for simulating organizational adjustments and human resource allocation plans.

[0023] After the group's human resources digital twin model is constructed, when an organizational restructuring event or human resources planning request is received, the corresponding organizational restructuring plan or human resources allocation plan is simulated and analyzed within the digital twin model. By analyzing the results of different restructuring plans operating in the digital twin model, a planning impact assessment result is generated. This planning impact assessment result is used to reflect the impact of organizational restructuring changes or human resources adjustments on staff size, job matching, and the rationality of resource allocation.

[0024] Based on the planning impact assessment results, corresponding local human resource allocation objective functions are constructed for each subsidiary. These local human resource allocation objective functions characterize the optimization objectives of each subsidiary in terms of workforce size control, job competency matching, and human resource cost constraints, thereby transforming the planning impact assessment results into decision inputs that can be used for subsequent optimization solutions.

[0025] After constructing the local human resource allocation objective function, global constraints are formed by combining the overall strategic goals and resource boundary conditions at the group level. Under the premise of satisfying these global constraints, the local human resource allocation objective functions of each subsidiary are collaboratively solved through a distributed game optimization method. Through this distributed game optimization process, each subsidiary can satisfy the overall constraints of the group while considering its own optimization goals, thereby forming a group-level human resource planning decision.

[0026] After obtaining the group-level human resource planning decision results, the results, their corresponding derivation parameters, and solution process information are reliably stored. By storing the decision results and their formation process, the human resource planning decisions become traceable and verifiable, providing a technical basis for subsequent audits, reviews, or compliance management.

[0027] After the group-level human resource planning and decision-making results are implemented, information on actual changes in organizational structure and human resource allocation during the planning and execution process is collected as execution feedback information. Based on this execution feedback information, the group's human resource digital twin model and the corresponding local human resource allocation objective functions of each subsidiary are updated, so that the digital twin model and optimization objectives can continuously reflect the actual operating status of the group's human resources, thereby providing a more accurate basis for the next round of human resource planning and decision support.

[0028] This embodiment constructs a digital twin model of group human resources, unifying the human resource structure at the group, subsidiary, and job levels. Within this model, it simulates organizational restructuring events or human resource planning requests, enabling human resource planning to assess the impact of different organizational adjustment or configuration schemes in a virtual environment. Compared to existing technologies that rely on post-event manual adjustments or static rule configurations, this embodiment achieves synchronous mapping and conflict prediction between the planned scheme and the actual organizational structure when the organizational structure dynamically changes. This significantly improves the adaptability and implementability of the planning scheme and reduces the risk of human resource allocation conflicts caused by organizational adjustments. Based on the generated planning impact assessment results, this embodiment constructs local human resource allocation objective functions for each subsidiary and, combined with global constraints at the group level, collaboratively solves the objective functions of each subsidiary through distributed game optimization. In this process, subsidiaries do not need to directly report complete raw human resource data to participate in group-level planning decisions. This achieves unified planning and collaborative optimization at the group level while ensuring subsidiary data security and privacy protection, effectively alleviating the contradiction between cross-subsidiary data sharing and privacy protection in existing technologies. This embodiment transforms the human resource planning problem into a computable objective function and constraints, and employs a distributed game optimization mechanism for solution. This enables planning decisions to automatically coordinate strategies and determine convergence under multi-objective and multi-constraint conditions. Compared to existing technologies that rely on manually summarizing lagging data for decision-making, this embodiment significantly improves the real-time nature and scientific rigor of human resource decisions, allowing the group to quickly formulate reasonable planning decisions when facing business changes, authority adjustments, or changes in resource boundaries. After generating group-level human resource planning decision results, this embodiment reliably stores the decision results and their corresponding derivation parameters and solution process information, making the formation process of planning decisions traceable and verifiable. By managing the evidence of the decision basis and solution process, it avoids the problems of unclear sources of planning decisions, tampered parameters, or difficulty in verifying the execution basis in existing technologies, providing reliable data support for subsequent audits, compliance management, and decision review. After the planning decision is executed, this embodiment collects actual changes in organizational structure and human resource allocation execution results as execution feedback information, and updates the group's human resource digital twin model and local human resource allocation objective function based on the execution feedback. Compared with existing technologies that lack the ability to evaluate the effects and continuously optimize after planning is completed, this embodiment can continuously correct the model parameters and optimization objectives, so that the human resource planning model can continuously align with the actual operating status of the group, providing a more accurate and reliable basis for subsequent planning decisions.In summary, this embodiment, through digital twin modeling, distributed collaborative optimization, trusted evidence storage, and closed-loop feedback updates, has realized the transformation of human resource planning for group enterprises from static management to dynamic deduction, collaborative decision-making, and continuous optimization. It has significant technical effects in improving planning accuracy, decision-making efficiency, execution controllability, and result credibility, and can effectively overcome the shortcomings of existing technologies in group human resource planning and decision support.

[0029] In one feasible implementation, such as Figure 2 As shown, step S120 above constructs a digital twin model of the group's human resources based on the group's organizational structure information, human resource allocation information, and job competency requirements information, including: S1201. Perform hierarchical parsing on the above organizational structure information to generate group-level organizational nodes, subsidiary-level organizational nodes, and job-level organizational nodes. S1202. Associate personnel size constraint parameters, job competency matching constraint parameters, and permission boundary constraint parameters with each of the above-mentioned organizational nodes to form an organizational node constraint set; S1203. Based on the hierarchical and business collaboration relationships between the above-mentioned organizational nodes, establish human resource mapping rules to constrain the above simulation and deduction. S1204. Based on the above-mentioned organizational nodes, the above-mentioned organizational node constraint set, and the above-mentioned human resource mapping rules, generate the above-mentioned group human resource digital twin model.

[0030] For example, the organizational structure information of the group company is first parsed hierarchically to generate group-level organizational nodes, subsidiary-level organizational nodes, and job-level organizational nodes. The group-level organizational nodes represent the group headquarters and its functional domains, the subsidiary-level organizational nodes represent each legal entity or business unit, and the job-level organizational nodes represent the set of jobs and staffing units under each organizational unit. Through this hierarchical parsing, the system obtains the topological relationship of the organizational structure, which can be represented by a directed hierarchical graph G. It means that, among them A set of nodes, containing the group's node set. Subsidiary node set With job node set This is a set of edges used to represent hierarchical relationships, such as the connection between a group and its subsidiaries, or between a subsidiary and its job position.

[0031] After node generation, the system further associates personnel size constraints, job competency matching constraints, and permission boundary constraints with each organizational node to form an organizational node constraint set, which is used to elevate the structure into a constrained computational model. For example, for any job level organizational node... The personnel size constraint parameter can be defined as the range of job positions. ,in This indicates the minimum number of employees required for this position. This indicates the maximum number of employees allowed for this position within the budget or management span; for subsidiary-level organizational nodes. The total number of employees in subsidiaries can be defined. and the upper limit of the subsidiary's human resource cost budget ,in and These represent the lower and upper limits of the subsidiary's personnel size under the business continuity and cost boundaries, respectively. This represents the maximum total human resource cost that the subsidiary can afford. Correspondingly, the job competency matching constraint parameter can be used to characterize the feasible matching conditions between the job competency requirement vector and the personnel competency vector. For example, the job competency matching constraint parameter can be used to characterize the feasible matching conditions between the job competency requirement vector and the personnel competency vector. Capability requirements are characterized as ,in For the number of capability dimensions, The components are used to represent the minimum requirements for different competency elements for that position; [The remaining text appears to be a fragment and requires further context for accurate translation.] Ability representation Then a matching degree function can be defined.

[0032] Where "•" represents the vector dot product, Represents the L2 norm, For measurement personnel With position The degree of similarity in the ability space; furthermore, a job ability matching constraint threshold can be set. They were required to be assigned to a position. any person satisfy ,in This is used to reflect the minimum skill requirement for a position. Authority boundary constraint parameters are used to characterize the authorization boundaries between a position and personnel in a compliance or internal control sense. For example, the position can be... The set of permissions is represented as ,staff The authorized set is represented as Then it requires or ,in This represents the set of permissions that are not allowed. The above constraints are used to avoid unauthorized configurations or conflicts of responsibility.

[0033] After forming the set of organizational node constraints, the system establishes human resource mapping rules for constraint simulation based on the hierarchical and business collaboration relationships between organizational nodes. This connects node parameters with cross-node flows / dependencies, ensuring that the simulation process conforms to both organizational structure logic and cross-departmental and cross-company collaboration constraints. For example, a job affiliation mapping can be established. This is used to map job nodes to their respective subsidiary nodes; and to establish personnel affiliation mappings. This is used to map employees to their current job positions and define the personnel allocation decision matrix. ,in Indicates employees Have they been assigned to a position? The feasibility of personnel configuration can then be constrained using mapping rules. For example, the rule that each employee can only be configured with one position can be represented as... The fact that the number of positions is not exceeded can be expressed as follows: The minimum configuration requirement for a position can be expressed as follows: Capability matching constraints can be expressed as: Permission boundary constraints can be represented as Furthermore, a job collaboration matrix can be defined. This is used to characterize the degree of synergy dependency between positions; defining the subsidiary synergy matrix. It is used to characterize the strength of synergistic coupling between subsidiaries, so as to uniformly express constraints such as cross-unit resource sharing, key position complementarity, and shared service capacity during simulation.

[0034] Finally, in S1204, the system generates a digital twin model of the group's human resources based on the aforementioned organizational nodes, organizational node constraint sets, and human resource mapping rules, making it a unified model object that can be used for subsequent simulations and deductions. For example, the group's human resources digital twin model can be formally represented as follows:

[0035] in Represents an organizational hierarchy topology diagram. This represents the set of constraint parameters associated with each organizational node (including personnel size constraint parameters, job competency matching constraint parameters, and permission boundary constraint parameters). This represents the set of human resource mapping rules (including attribution mapping, configuration feasibility rules, and collaborative dependency rules). This represents the configuration state variables or configuration decision variables of the organization, positions, and personnel. Through the above formal structure, the system can make consistent feasibility judgments and impact assessments on schemes such as organizational splitting, merging, adding or removing positions, and cross-company transfers in subsequent simulations. It also provides a unified data structure and constraint basis for planning optimization solutions at the group level, thereby improving the computability, interpretability, and implementability of human resource planning for group enterprises.

[0036] In one feasible implementation, such as Figure 3 As shown, the specific construction steps for the above objective function and the above global constraints include: S210. Based on the above planning impact assessment results, construct subsidiary objective functions for each subsidiary to characterize the rationality of personnel size, job matching degree, or human resource cost constraints. S220. Based on the overall strategic goals and resource boundary conditions of the group, a group-level global constraint function is constructed to limit the feasible solution space of the objective functions of the aforementioned subsidiaries; wherein, the objective functions of the aforementioned subsidiaries and the aforementioned global constraint function together constitute the basis for solving the distributed game optimization problem.

[0037] For example, such as Figure 3 As shown, the purpose of steps S210 and S220 is to transform the impact assessment conclusions obtained from the previous simulation into calculable optimization objectives and verifiable group boundary constraints, so that subsequent distributed game optimization can achieve a unified solution between the local optimum of the subsidiary and the global feasibility of the group.

[0038] Specifically, in S210, the system first constructs a subsidiary objective function for each subsidiary based on the planning impact assessment results. This subsidiary objective function quantitatively characterizes the subsidiary's comprehensive optimization needs in areas such as the rationality of its workforce size, job competency matching, and human resource cost control, thus mapping the assessment results into optimizable decision objectives. Subsequently, in S220, the system constructs a group-level global constraint function based on the group's overall strategic goals and resource boundary conditions. This function limits the feasible solution space of each subsidiary's objective function, ensuring that any solved configuration strategy meets the group-level hard boundaries such as staffing, budgeting, compliance, and key capability assurance. Therefore, the aforementioned subsidiary objective functions and the global constraint function together constitute the foundation for solving the distributed game optimization problem.

[0039] For ease of explanation, let's assume there are a total of [number] people within the group. The first subsidiary The human resource allocation strategy variable of the subsidiary is denoted as ,in This can represent a set of decision variables such as the allocation vector for staffing positions, the number of key positions to be filled, and the combination ratio of internal transfers and external recruitment; the overall group strategy is denoted as... The results of the planning impact assessment can be organized into a set of measures related to gaps, risks, and cost pressures, such as a job gap vector. (Characterizing the gap size of each position or job family), capability risk vector (Characterizing the risk level resulting from insufficient key capabilities) and cost pressure scalar or vector. (This represents the degree of budget constraints or the risk of cost overruns). Based on this, the system constructs the first... Objective function of subsidiaries An exemplary form can be written as

[0040] in, This is a cost item for the reasonableness of personnel size, used to measure the degree of deviation between the subsidiary's configured personnel size and its business needs or staffing range; This is the job matching cost item, used to measure the degree of mismatch between job skill requirements and personnel skill supply. This is a human resource cost item, used to measure the overall costs introduced by the configuration plan, such as salary costs, recruitment and training costs, or outsourcing costs. These are weighting coefficients used to reflect the subsidiary's preference strength among scale rationality, matching degree, and cost control. These weighting coefficients can be adaptively set based on the planning impact assessment results; for example, the larger the gap, the higher the weighting coefficient. To strengthen matching priority, the higher the cost pressure, the higher the priority will be. To strengthen the cost constraint orientation.

[0041] As a further example, the cost of reasonable personnel size can be characterized by "deviation from the target size," for example, by letting This indicates that the subsidiary is in strategy The total number of personnel below, If the reference scale is derived from the planning impact assessment, then it can be defined as follows: in, This can be derived from business volume forecasts, capacity requirements, or projections, and used to guide staffing levels towards a reasonable range. Job matching costs can be used to penalize key position vacancies or skill deficiencies; for example, let... This indicates the set of key positions in the subsidiary, making Indicates job position In strategy The gap amount or mismatch degree can then be defined.

[0042] in, The importance weight of the position is used to prioritize filling key positions; when When defined as the number of people in shortfall, the above formula represents a weighted penalty for the shortfall; when When defined as the degree of skill mismatch, the above formula represents a weighted penalty for skill deviation. The human resource cost term can be directly represented by a comprehensive cost function, for example:

[0043] di in which Let be the inequality constraint vector. Further exemplarily, the upper limit constraint on the total group staffing can be expressed as:

[0044] in This represents the maximum available staffing level for the group; the group's total human resource cost budget constraint can be expressed as: in This represents the upper limit of the group's annual or periodic human resources budget; the key competency coverage constraint can be used to ensure that the supply of key competencies for key positions within the group is not lower than the minimum requirement. For example, let the group's key competency set be... ,make Representation Strategy Lower ability The coverage or satisfaction level can be written as: in This represents the minimum capability threshold; compliance and authority boundary constraints can be used to restrict personnel allocation or authorization for sensitive positions, and can be exemplified as follows: in Personnel Have you been assigned to a position? Indicates the set of authorized personnel. This represents the set of job authority permissions. The constraints mentioned above are used to avoid unauthorized configurations and conflicts of responsibility. Through the above global constraint function, the system can ensure that each subsidiary pursues its own objective function optimization without exceeding the hard boundaries at the group level, achieving collaborative optimization within the feasible region.

[0045] Therefore, in this implementation, S210 maps the planning impact assessment results into a subsidiary objective function composed of scale, matching, and cost terms, thus achieving a calculable expression of the optimization needs of each subsidiary; S220 formalizes the group's strategic objectives and resource boundary conditions into a global constraint function, thus uniformly limiting the feasible solution space of each subsidiary. Together, these two elements form the basis for solving the distributed game optimization problem, enabling the subsequent solution process to reflect the differentiated human resource optimization objectives of each subsidiary while also meeting the group-level requirements for staffing, budgeting, key capabilities, and compliance boundaries. This enhances the interpretability, feasibility, and global consistency of human resource planning decisions for group enterprises.

[0046] In one feasible implementation, such as Figure 4 As shown, step S150 above, combined with global constraints at the group level, uses a distributed game optimization approach to collaboratively solve the objective function of local human resource allocation, resulting in a group-level human resource planning decision, including: S1501. Under the premise of satisfying the above global constraint function, the strategy variables corresponding to the objective function of each of the above subsidiaries are solved in a distributed manner using an iterative update mechanism. S1502. In each iteration, determine whether the preset equilibrium convergence condition is met based on the changes of the above-mentioned strategy variables. S1503. When it is determined that the above equilibrium convergence condition is met, stop the iterative solution and take the corresponding strategy combination as the above group-level human resource planning decision result.

[0047] For example, such as Figure 4 As shown, step S150 is used to coordinate the conflicting relationship between the optimal and feasible aspects of each subsidiary within the unified governance boundary of the group, thereby outputting a feasible group-level human resource planning decision.

[0048] Specifically, in S1501, the system first treats each subsidiary as a decision-making entity, under the premise of satisfying the global constraint function at the group level. It defines subsidiary objective functions related to local business objectives, human resource cost control, and job competency matching, and describes the human resource allocation decision items of each subsidiary using strategy variables. Then, an iterative update mechanism is used to allow each subsidiary to adjust its own strategy variables round by round based on the current group constraints and the strategies of other subsidiaries, until in S1502 it is determined that the changes in the strategy variables have reached the preset equilibrium convergence condition. Finally, in S1503, the converged combination of strategy variables is output as the group-level human resource planning decision result.

[0049] For ease of understanding, the above distributed game optimization problem can be abstracted into a multi-agent optimization with global constraints: Assume that the group has a total of The first subsidiary The strategy variables of the subsidiaries are denoted as It can represent the allocation decision vector, including the number of positions, the number of personnel transferred, and the intensity of capacity enhancement for key positions; the group strategy vector is obtained by concatenating the strategies of all subsidiaries. . No. The objective function of a subsidiary can be expressed as a minimization form.

[0050] in, Indicates the first The cost (or loss) of local human resource allocation for subsidiaries. Indicates except the first The strategy set of subsidiaries other than the parent company; This represents a comprehensive measure of human resource costs (such as salaries, outsourcing, recruitment and training costs, etc.). This section indicates job vacancy / competency risk, used to depict the risks arising from the inability to meet competency requirements for key positions or the resulting job vacancies. This indicates cross-company coupling terms, used to characterize the mutual impact caused by shared talent pools within the group, cross-company transfers, and shared service center resource usage; The weighting coefficient is used to reflect the differences in cost, risk and synergy preferences among different subsidiaries. The weight can be determined by the group's strategic priorities, business scale or operating objectives.

[0051] Global constraint functions at the group level are used to limit the feasible solution space and can be written as follows: in, Represents the set of inequality constraints. This represents the set of equality constraints. For example, inequality constraints may include group total constraints. (in This represents the maximum total available staffing for the group. (All 1 vectors) Group total human resource cost budget constraint (in (This refers to the budget cap), and the boundaries of permissions / compliance constraints (e.g., certain positions must meet qualification thresholds, key positions must have dual backups, etc., which can be uniformly coded into...). Equality constraints can be used to express hard configuration relationships that must be satisfied, such as the conservation of supply capacity of shared service centers and demand of subsidiaries, or the balance of personnel allocation across companies.

[0052] During the iterative update of S1501, a feasible distributed update method with penalties or Lagrange multipliers can be adopted, allowing each subsidiary to automatically consider the cost of global constraints when solving locally. For example, an augmented Lagrange function can be constructed.

[0053] in, is a Lagrange multiplier vector used to map the degree of constraint violation to "constraint price"; This is the penalty coefficient, used to increase the severity of penalties for constraint violations; This indicates that the portion violating the constraint is truncated, ensuring that penalty is applied only to the excess portion. Then the... The subsidiary in The iterative strategy update can be written as:

[0054] This indicates that the current strategy of other subsidiaries is being fixed. Compared with the current constrained price In the case of the first Each subsidiary obtains its next-round strategy by solving a local subproblem; simultaneously, the group can update the multipliers, for example: in, This represents the aggregated group strategy after this iteration. The multiplier update strengthens the penalty for directions that violate constraints in the next iteration, thereby driving the strategy towards convergence to the feasible region. Through this type of iterative update mechanism, the system can gradually balance the local goals of subsidiaries with the global constraints of the group under distributed conditions.

[0055] In S1502, the equilibrium convergence condition can be determined by the change in the policy variable or the change in the objective function. For example, a policy convergence threshold can be set. When the following conditions are met: At this point, it is assumed that the strategies of each subsidiary have changed sufficiently in two adjacent iterations, and the system has entered a stable state; or the global cost convergence criterion is used, when... Convergence is determined at time, where This represents the threshold for cost variation. Furthermore, to ensure feasibility, constraints can be satisfied using criteria, such as... ,in To constrain violations of tolerance thresholds.

[0056] In S1503, when the above equilibrium convergence condition is satisfied, the system stops iteratively solving and the finally converged strategy combination is applied. The output is the group-level human resource planning and decision-making results, among which Including final staffing decisions for each subsidiary And the corresponding group-wide constraints on pricing or resource allocation. This decision-making outcome can simultaneously meet the overall constraints of the group level, such as staffing, budgeting, and compliance, and on this basis, optimize the goals of each subsidiary in terms of cost, capability matching, and employment risks as much as possible, thereby achieving collaborative decision-making and feasible implementation in organizational restructuring or human resource planning scenarios for group enterprises.

[0057] In one feasible implementation, such as Figure 5 As shown, step S160 above reliably verifies the above-mentioned group-level human resource planning decision results and their corresponding derivation parameters and solution process information, including: S1601. Extract the summary of the above-mentioned group-level human resource planning decision-making results, the above-mentioned inference parameters and the above-mentioned solution process information, and generate evidence data. S1602. Write the above-mentioned evidence data into an immutable evidence storage medium to form a decision evidence record corresponding to the above-mentioned group-level human resources planning decision results. S1603. Through the above-mentioned decision-making record, the formation process of the above-mentioned group-level human resource planning decision results can be traced and verified.

[0058] For example, in S1601, the system first performs a summary extraction operation on the group-level human resource planning decision results, inference parameters, and solution process information. The information is standardized and organized according to a preset data structure, and a summary value is calculated, which is then used to generate evidence-based data. The group-level human resource planning decision results may include the staffing levels of each subsidiary, key position filling strategies, cross-company transfer plans, and budget allocation results. The inference parameters may include digital twin model version identifiers, organizational node constraint parameters, and input events and scenario parameters for simulation. The solution process information may include the iteration rounds of distributed game optimization, convergence criteria, strategy variable trajectories, and global constraint satisfaction status. Through summary extraction, the system compresses large volumes of variable-length decision data into verifiable consistency fingerprints, enabling subsequent integrity protection of key decision elements at a lower storage cost.

[0059] In step S1602, the system writes the aforementioned evidence data into an immutable evidence storage medium, thereby forming a decision evidence record that corresponds one-to-one with the group-level human resource planning decision result. For example, the immutable evidence storage medium can be a log ledger, an audit chained storage structure, or an evidence repository employing an append-only mechanism, all of which can at least guarantee the irreversible alteration, verifiable consistency, and chronological continuity of the evidence data after writing. Through this write operation, the system achieves the binding and solidification of "decision result—deduction parameters—solution process information," ensuring that any decision evidence record uniquely corresponds to a specific human resource planning decision process, thereby preventing arbitrary replacement of the decision basis or selective presentation of the solution process afterward.

[0060] In S1603, the system achieves traceable verification of the formation process of group-level human resource planning decisions through the aforementioned decision evidence records. Specifically, when a review or audit is required, the system can retrieve the corresponding decision data, deduction parameters, and solution process information, and recalculate the summary value. The recalculated summary value is then compared with the summary value in the decision evidence records. When the comparison is consistent, it indicates that the group-level human resource planning decision and its formation process have not been tampered with, and its formation path can be fully traced. When the comparison is inconsistent, the system can output an anomaly prompt and locate the data item or process segment related to the summary, providing a basis for subsequent risk management, responsibility definition, and decision correction. Thus, this implementation method, through the extraction of summaries of key decision elements, tamper-proof writing, and consistency verification and backtracking, constructs a credible evidence storage and traceable verification mechanism for group-level enterprise human resource planning decisions, improving the transparency, auditability, and compliance credibility of planning decisions from a technical perspective.

[0061] In one feasible implementation, such as Figure 6 As shown, step S170, after the execution of the aforementioned group-level human resource planning decision results, updates the aforementioned group human resource digital twin model and the aforementioned local human resource allocation objective function based on the execution feedback information, including: S1701. Collect information on actual changes in organizational structure and human resource allocation results generated during the execution of the above-mentioned group-level human resource planning and decision-making results, and use this as the above-mentioned execution feedback information. S1702. Compare and analyze the above-mentioned implementation feedback information with the above-mentioned planning impact assessment results to generate planning deviation indicators; S1703. When the above-mentioned planning deviation index exceeds the preset threshold, the organizational node constraint parameters in the above-mentioned group human resources digital twin model or the above-mentioned local human resources configuration objective function shall be corrected, and the above-mentioned distributed game optimization shall be re-executed based on the corrected model.

[0062] For example, in S1701, the system continuously collects information on actual changes in organizational structure and the results of human resource allocation as execution feedback information during the execution of group-level human resource planning decisions. The information on actual changes in organizational structure may include actual structural changes such as organizational mergers / splitting, additions or reductions of positions, changes in the responsibilities of key positions, and adjustments to authority boundaries. The results of human resource allocation may include execution result data such as personnel on-the-job status, completion of inter-company transfers, recruitment and resignation status, completion of training, and actual human resource costs. Through the above collection operations, the system obtains the "real-world trajectory of the planning scheme," providing a quantifiable basis for subsequent deviation assessment.

[0063] In S1702, the system compares and analyzes the execution feedback information with the planning impact assessment results to generate a planning deviation index. This index quantifies the degree of deviation between "planning forecasts" and "execution reality," and can characterize the overall deviation or be broken down to organizational, subsidiary, and job levels to support targeted corrections. For example, the deviation index can be constructed as a vector consisting of personnel size deviation, job matching deviation, and cost deviation.

[0064] in, This indicates a discrepancy in personnel size. This indicates a job mismatch. This represents cost variance. As a further example, staffing variance can be calculated from the difference between the planned number of employees and the actual number, for example:

[0065] in, This indicates the target staff size for the group or a subsidiary within the execution period, as given by the planning and decision-making results. This indicates the actual number of personnel obtained from the statistics of the execution feedback information. To prevent small positive numbers with a denominator of zero, job mismatch can be used to characterize changes in the degree of key job vacancies or competency mismatches, and can be exemplarily written as:

[0066] in, This represents a set of key positions. Indicates job position Importance weight, Indicating positions in planning assessment The extent to which the expected gap or capacity is not met. Indicating the position in the execution feedback The actual gap or degree of capacity shortfall; when the actual gap is higher than the planned gap, the above formula is passed. Cumulative penalties are applied to any excesses to highlight the risk of key positions failing to meet targets. Cost variance can be expressed as the relative deviation between actual labor costs and planned costs, for example:

[0067] in, This indicates the budgeted or projected human resource costs corresponding to the planning and decision-making outcomes. This represents the actual human resource cost statistically obtained from the execution feedback information. To facilitate the formation of a single criterion, the above deviation vector can be further summarized into a comprehensive deviation index:

[0068] in, The weighting of the deviations is used to reflect the group's priority in focusing on deviations in scale, matching, and cost.

[0069] In S1703, when the planning deviation index exceeds a preset threshold, the system initiates a targeted correction mechanism: on the one hand, it corrects the organizational node constraint parameters in the group's human resources digital twin model; on the other hand, it adjusts the objective function for local human resources allocation in each subsidiary, and re-executes distributed game optimization based on the corrected model, thereby outputting updated group-level planning decision results. For example, the preset threshold can be expressed as... ,when Time-triggered corrections. For corrections to organizational node constraint parameters, personnel size constraint parameters, job competency matching constraint parameters, and permission boundary constraint parameters can be updated based on the source of deviation: for example, when personnel size deviation... When the number of employees remains positive and the actual number is lower than the planned number, the lower limit parameter for the number of employees in a position or subsidiary can be increased to improve the rigidity of the guarantee, or the upper limit parameter can be decreased to reflect the tightening budget; this can be expressed as an example linear correction method.

[0070] in, This represents a node in an organization or a job position. and These represent the planned number of people and the actual number of people at that node, respectively. To adjust the step size coefficient, which controls the parameter update amplitude and avoids oscillations caused by over-adjustment, the matching threshold or weight of key positions can be adjusted based on the failure of key positions to meet the requirements. For example, the position... Matching threshold or gap penalty weight Update the plan to prioritize meeting the skill requirements of this position in the next round of planning. For modifications to permission boundary constraint parameters, tighten or refine the permission set boundaries based on risks of unauthorized access, conflicts of responsibility, or compliance audit results encountered during execution, to improve the compliance and executability of the configuration scheme.

[0071] Meanwhile, the objective function for local human resource allocation can be modified by adjusting the weights of the scale, matching, and cost terms in the objective function to achieve adaptive calibration of the optimization direction. For example, if the... The objective function of the subsidiary is:

[0072] Then when there is a job mismatch When the shortage of key positions is significant and has not been addressed, an increase can be made. To enhance the optimization of matching items; when cost deviation When the cost is significant and the actual cost exceeds the budget, it can be increased. To enhance cost control orientation; an exemplary weight update method can be expressed as: in, The step size coefficient is updated for the weights to adjust the sensitivity of the weights to changes in deviation. When a certain type of deviation is small or has stabilized, the corresponding weights can be kept unchanged or decayed to avoid overfitting to disturbances in a specific stage.

[0073] After revising the aforementioned constraint parameters and objective function parameters, the system re-executes distributed game optimization based on the revised group-level human resource digital twin model and the revised local human resource allocation objective function. This results in updated group-level human resource planning decisions, ensuring that the new decisions not only meet the latest organizational structure and resource boundary conditions but also better align with the patterns of personnel onboarding, job matching, and cost control revealed by real-world execution feedback. Through this implementation method, the system upgrades from one-off planning to continuous adaptive planning, significantly improving the stability, feasibility, and long-term effectiveness of human resource planning decisions in dynamic organizational environments for group enterprises.

[0074] In one feasible implementation, such as Figure 7 As shown, before modifying the constraint parameters in the aforementioned group human resources digital twin model or the aforementioned local human resources allocation objective function, the following also applies: S310. Determine the type of deviation for the above planning deviation indicators to distinguish whether the above planning deviation indicators are caused by mismatch of organizational structure constraints or by deviation of human resource allocation optimization goals. S320. When the above planning deviation indicators are determined to be of the organizational structure constraint mismatch type, priority is given to triggering the correction of personnel size constraint parameters, job competency matching constraint parameters or authority boundary constraint parameters in the above group human resources digital twin model. S330. When the above planning deviation index is determined to be of the type of human resource allocation optimization target deviation, the objective function parameters in the above local human resource allocation objective function shall be corrected first. S340. Based on the above deviation type determination results, determine the execution order or combination of the above constraint parameter correction and the above objective function parameter correction.

[0075] For example, in S310, deviation type determination is performed on the planning deviation index. The core idea is to decompose the planning deviation into structural deviation components that reflect whether organizational structure constraints are still effective, and target deviation components that reflect whether the optimization target weights or orientations have shifted. When structural deviation is dominant, it indicates that the deviation is mainly caused by a mismatch between constraints such as organizational node constraints, authority boundaries, or staffing intervals and the actual organizational state. In this case, the constraint parameters of the group's human resources digital twin model should be corrected first. When target deviation is dominant, it indicates that the overall organizational structure constraints are still usable, but the weight settings of the optimization targets and the trade-off between cost-matching-scale are no longer suitable for the current operating state. In this case, the objective function parameters in the local human resources allocation objective function should be corrected first, and the order or combination of the two types of corrections should be further determined in S340 to form a robust closed-loop update strategy.

[0076] Let the planning deviation index vector obtained in step S1702 be... ,in, This indicates a discrepancy in personnel size. This indicates a job mismatch. This indicates a deviation in labor costs. To determine the type of deviation, the system can further construct two attribution scores: a structure mismatch score. Target offset score For example, a structural mismatch score, used to measure whether constraints are systematically violated or ineffective, can be determined by both the degree of constraint violation and the intensity of organizational structural changes, for example:

[0077] in, Indicates the actual execution configuration The degree of constraint violation under the given conditions can be defined as: here This refers to a group-level global constraint function or a set of constraints derived from a digital twin model. This indicates that only the portion exceeding the limit that violates the constraints will be counted. It is a norm 2; This indicates the intensity of organizational structure changes, used to characterize the scale of structural changes during execution, such as merging and splitting organizational units, adding or removing positions, and adjusting authority boundaries. It can be represented by a weighted count of change events or the percentage of change nodes. For example: in, To organize the set of nodes, For the number of nodes, As an indicator function, the more extensive the structural changes, the... The larger; , which is a weighting coefficient used to adjust the contribution ratio of constraint violation degree and structural change intensity in structural mismatch score.

[0078] The target deviation score measures whether, given that constraints are largely satisfied, the current target trade-offs result in a persistent critical deviation. It can be determined by both the structural characteristics of the deviation index and the persistence of the deviation. For example, it can be defined as follows:

[0079] in, The comprehensive deviation index can be obtained by summing up according to weights. Weights for the total deviations; This is a deviation persistence indicator used to characterize whether the deviation exhibits a systematic deviation over multiple consecutive execution cycles. It can be defined as the mean deviation or trend slope within a sliding window, for example: in, To calculate the window length, For the first The overall deviation over each period, , where is the weighting coefficient, is used to reflect the relative importance of single-period deviation and multi-period persistent deviation. The above definition makes it more likely that when the constraint violation is not high but the deviation persists for a long time, the system will judge it as a shift in the optimization objective rather than a constraint mismatch.

[0080] In S310, the system can determine the type of deviation based on the two scores mentioned above. For example, a discrimination threshold can be set. When satisfied When the planning deviation index is determined to be a type of organizational structure constraint mismatch; when the conditions are met... When the planning deviation index is determined to be a deviation from the human resource allocation optimization target, the difference between the two is less than 1. When the condition is met, it can be determined to be a mixed type or an uncertain type, and a combined correction strategy is adopted in S340 to improve robustness. Here This is used to avoid frequent switching when scores are close, thereby improving the stability of closed-loop updates.

[0081] In S320, when an organizational structure constraint mismatch is identified, the system prioritizes correcting the constraint parameters in the group's human resources digital twin model. These constraint parameters may include personnel size constraint parameters, job competency matching constraint parameters, and authority boundary constraint parameters. The correction logic involves locating nodes or positions where "constraints are systematically breached or no longer applicable" by executing feedback information, and tightening, relaxing, or reconstructing the corresponding constraint boundaries to ensure that subsequent deductions and solutions are performed within the updated feasible domain. For example, when the actual number of employees in a certain position consistently falls below the planned lower limit, the system can adjust the personnel size lower limit for that position. The system can be adjusted in conjunction with the job supply strategy, or the corresponding collaborative dependency constraints can be reconfigured to improve executability; when unauthorized configurations or conflicting responsibilities occur, the system can adjust the permission boundary set. Convergent adjustments are made to reduce compliance risks.

[0082] In S330, when it is determined that the human resource allocation optimization objective is deviated, the system first triggers a correction to the objective function parameters in the local human resource allocation objective function. The core of this correction is to adjust the "scale-matching-cost" trade-off structure, preventing the optimization process from iterating along the direction that causes the deviation to persist. For example, if the first... The objective function of the subsidiary is

[0083] When there is a significant shortage or mismatch in key positions, promotion can be implemented. and appropriately reduce This enhances the optimization orientation towards matching items; when cost deviations are significant and budget constraints are tighter, it can improve... and appropriately reduce This strengthens the cost control orientation. The above objective function parameter modification can be represented by an exemplary proportional update as follows:

[0084] in, To update the step size coefficient, which is used to control the sensitivity to weight changes, These are the matching bias and cost bias, respectively, ensuring that the weight updates are consistent with the magnitude of the bias.

[0085] In S340, the system determines the execution order or combination of constraint parameter correction and objective function parameter correction based on the deviation type determination result, in order to achieve a balance between correction effect and solution stability. For example, when the system is determined to be of the organizational structure constraint mismatch type, a constraint-first, then objective-second order can be adopted, i.e., first correct the constraint parameters of the digital twin model to form a new feasible region, and then perform fine-tuning of the objective function parameters within the new feasible region. When the system is determined to be of the optimization objective offset type, an objective-first, then constraint-second order can be adopted, i.e., first adjust the objective function to restore the solution direction to a reasonable level, and then perform local corrections on a small number of constraint boundaries if necessary. When the system is determined to be of the mixed or uncertain type, a combined correction strategy can be adopted, such as updating constraint parameters and objective function parameters simultaneously with a small step size, and recalculating the deviation index and discrimination score after each round of correction until the deviation falls back to within the threshold or reaches the stability condition. Through the above-mentioned deviation type determination and correction path selection mechanism, this implementation method can transform the deviation correction process into an interpretable, controllable, and convergent closed-loop update process, thereby improving the adaptability and long-term stability of the human resource planning system of group enterprises in a dynamic organizational environment.

[0086] In one feasible implementation, such as Figure 8 As shown, step S310 above determines the deviation type of the above planning deviation index, including: S3101. Based on the differences between the above-mentioned execution feedback information and the above-mentioned planning impact assessment results, a structural deviation analysis is conducted on the above-mentioned planning deviation indicators. S3102. When the above-mentioned differences in distribution characteristics are mainly concentrated in the range of organizational level changes, the range of job staffing deviations, or the range of authority coverage, the above-mentioned planning deviation indicators are judged as organizational structure constraint mismatch types. S3103. When the above-mentioned differences in distribution characteristics are mainly concentrated in the range of changes in personnel allocation ratio, job matching degree, or human resource cost weight, the above-mentioned planning deviation indicators are judged as human resource allocation optimization target deviation types.

[0087] For example, in S3101, the difference between the execution feedback information and the planning impact assessment results is first expressed in a distributed manner. That is, not only is the overall deviation calculated, but also the levels, job groups, or constraint dimensions in which the deviation is concentrated are characterized, thereby realizing structural deviation analysis. In S3102 and S3103, the system further completes the type determination based on the main concentration range of the difference distribution: if the difference is mainly concentrated in the range of organizational level changes, job staffing deviations, or authority coverage, it indicates that the planning deviation is more due to the constraint boundaries or organizational topology no longer matching the actual execution, and is therefore determined to be of the organizational structure constraint mismatch type. If the difference is mainly concentrated in the range of personnel allocation ratio, job matching degree, or human resource cost weight changes, it indicates that the organizational constraints are still largely usable, but the trade-off direction of the optimization goal is inconsistent with the actual execution pattern, and is therefore determined to be of the human resource allocation optimization goal deviation type.

[0088] For ease of explanation, let the predicted values ​​of several key indicators in the planning impact assessment results be denoted as . The actual value of the corresponding indicator in the execution feedback information is recorded as follows: Then, we can define a difference vector: in, The different components correspond to differences in different dimensions, such as differences related to changes in organizational hierarchy, deviations in job positions, differences related to authority coverage, differences related to staffing ratios, differences related to job matching, and differences related to changes in cost weights. To extract the distribution characteristics of these differences, the system can group the difference components by dimension to form several difference sets and calculate the proportion or energy contribution of each group of differences in the overall differences. For example, the set of differences related to organizational structure can be defined as follows:

[0089] in, This indicates the difference in the range of organizational hierarchical changes (e.g., the difference between the actual percentage of organizational nodes that underwent hierarchical adjustments and the predicted percentage). This indicates the difference between the actual number of positions and the predicted number of positions (e.g., the difference between the actual number of positions exceeding or falling below the lower limit and the predicted number of positions). This represents the difference between the scope of authority coverage (e.g., the difference between the number of actual unauthorized, conflicting, or insufficient authority events and the predicted risk); correspondingly, the set of goal-oriented related differences can be defined as: in, This indicates the difference between the actual implementation and the planned forecast in terms of staffing ratios (such as the difference between the ratios of internal transfers, external recruitment, and outsourcing). This indicates the difference in job matching (e.g., the difference between the actual and predicted average matching or gap in key positions). This represents the difference in cost weight changes (e.g., the difference between the weight drift caused by a significant increase in cost control priority during actual implementation and the planned setting). Based on the above grouping, the system can calculate the structural difference intensity and the target difference intensity separately, for example, by defining them using the normalized energy ratio:

[0090] in, It is a norm 2. To prevent tiny positive numbers with a denominator of zero, This indicates the contribution of organizational structure-related differences to the overall differences. This indicates the contribution of goal-oriented differences to the overall differences; the two quantities mentioned above together constitute the core expression of the differences distribution characteristics. In addition to energy proportion, the system can further introduce concentration to characterize whether differences are highly concentrated in a few dimensions, for example, defining a difference concentration index:

[0091] in, Difference vector The One portion, The larger the value, the more concentrated the differences are in a few key dimensions, which is more conducive to cause identification and targeted correction.

[0092] After completing the structural deviation analysis described above, the system determines in S3102 whether the difference distribution characteristics are mainly concentrated in the set of differences related to organizational structure. For example, a judgment threshold can be set. When satisfied This indicates that the overall discrepancy is mainly contributed by differences related to the range of organizational hierarchy changes, deviations in job staffing, or differences in authority coverage, thus classifying the planning deviation index as a type of organizational structure constraint mismatch. Here... This is used to avoid frequent switching when the contributions of the two types of differences are close, and to improve the stability of the determination. Accordingly, in S3103, when the following conditions are met... When the overall difference is mainly contributed by differences in personnel allocation ratios, job matching degrees, or variations in the weighting of human resource costs, the planning deviation indicator is classified as a deviation from the human resource allocation optimization target. If both differences are less than [a certain value], then [the deviation is considered]. If the situation is not clear, it can be treated as a mixed or uncertain type, and a combined correction strategy or the introduction of a longer time window for difference statistics can be used in subsequent steps to enhance the reliability of the judgment.

[0093] Through the above implementation methods, S3101 achieves structural deviation analysis of planning deviation indicators by grouping, normalizing, and characterizing the differences between execution feedback and planning evaluation, so that the deviation is no longer just a large value, but can be interpreted as where the deviation is concentrated. S3102 and S3103 further complete the type determination based on the difference distribution characteristics, so that the system can prioritize the selection of more targeted correction objects in the subsequent update stage. That is, when the organizational structure constraint is mismatched, the constraint parameters of the digital twin model are corrected first, and when the target is offset, the objective function parameters are corrected first. This significantly improves the effectiveness and convergence stability of closed-loop correction and reduces the risk of planning oscillation caused by over-correction.

[0094] In one feasible implementation, after determining the execution order or combination of the constraint parameter correction and the objective function parameter correction based on the above deviation type determination result, the method further includes: S410. Perform the above-mentioned constraint parameter correction or above-mentioned objective function parameter correction in stages according to the preset correction step size or correction ratio. S420. After each correction phase is completed, recalculate the above planning deviation index and determine whether the preset stability conditions are met. S430. When it is determined that the above stability conditions are met, the subsequent correction phase is terminated.

[0095] For example, in S410, the system performs phased adjustments to constraint parameters or objective function parameters based on a preset adjustment step size or adjustment ratio. The preset adjustment step size limits the maximum magnitude of a single parameter change, and the preset adjustment ratio limits the proportion of a single adjustment to the total adjustment. This phased execution method reduces the impact of a single adjustment when deviations are large or the model is highly sensitive, improving the stability of subsequent solutions and operations. For example, when the adjustment target is the personnel size constraint parameter, job competency matching constraint parameter, or permission boundary constraint parameter of an organizational node, the system can break down the parameter update into multiple rounds of small-step updates. When the adjustment target is the objective function parameter (e.g., the weight coefficient of cost, matching, or size item) in the local human resource allocation objective function, the system also adjusts the weights round by round according to the preset ratio, gradually bringing the objective orientation back to a reasonable range.

[0096] In S420, after each correction phase, the system recalculates the planning deviation index and determines whether it meets the preset stability conditions. This allows for a comprehensive evaluation of the correction results in the current phase across two dimensions: "effectiveness of the correction" and "system stability." Specifically, the recalculated planning deviation index reflects the improvement effect of the correction on personnel size deviation, job matching deviation, or labor cost deviation in this phase. The preset stability conditions constrain the correction process from triggering new oscillations or uncertainties. For example, the planning deviation index is required to show a monotonically decreasing trend compared to the previous phase or to reach a preset threshold. Simultaneously, the number of iterations in the optimization process, the fluctuation range of strategy variables, or the degree of constraint violation are required to remain within acceptable ranges. By re-evaluating after each phase, the system can achieve "correction and verification simultaneously," thus avoiding over-adjustment caused by continuing subsequent correction phases when the deviation has clearly converged, or timely termination and reverting to a more robust correction strategy when instability arises during correction.

[0097] In S430, when the preset stability condition is met, the system terminates the subsequent correction phase. That is, after confirming that the current phase has achieved effective correction and the system operation has become stable, the remaining correction actions are no longer pursued. This reduces the computational overhead and strategy disturbances caused by ineffective adjustments and improves the controllability and convergence efficiency of the closed-loop update process. Thus, this implementation method, through a combination of phased correction, phased deviation recalculation, and early termination based on stability determination, enables the correction of constraint parameters and objective function parameters to be gradual and stopable at the execution level. This significantly improves the stable convergence capability and continuous adaptability of the human resource planning decision-making system for group enterprises in a dynamic organizational environment.

[0098] The second aspect, such as Figure 10 As shown, this invention also proposes a human resource planning and decision support system for group enterprises, including: Acquisition unit 21 is used to acquire organizational structure information, human resource allocation information, and job competency requirements information of the group enterprise; The first building unit 22 is used to build a digital twin model of the group's human resources based on the organizational structure information, human resource allocation information and job competency requirements information of the aforementioned group enterprises. The aforementioned digital twin model of the group's human resources is used to depict the human resource mapping relationship between the group level, subsidiary level and job level. The simulation unit 23 is used to receive organizational restructuring events or human resource planning requests based on the above-mentioned group human resource digital twin model, and to simulate and simulate the corresponding organizational restructuring plan or human resource allocation plan in the above-mentioned group human resource digital twin model to generate planning impact assessment results. The second construction unit 24 is used to construct the local human resource allocation objective function for each subsidiary based on the above planning impact assessment results; Solving unit 25 is used to combine the global constraints at the group level and solve the above local human resource allocation objective function in a collaborative manner through distributed game optimization to form the group-level human resource planning decision results; The evidence storage unit 26 is used to reliably store the above-mentioned group-level human resource planning decision results and their corresponding inference parameters and solution process information; The update unit 27 is used to update the aforementioned group-level human resources digital twin model and the aforementioned local human resources allocation objective function based on the execution feedback information after the aforementioned group-level human resources planning decision results are executed.

[0099] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A human resource planning and decision support method for group enterprises, characterized in that, include: Obtain information on the group company's organizational structure, human resource allocation, and job competency requirements. Based on the organizational structure information, human resource allocation information, and job competency requirement information of the group enterprise, a digital twin model of the group's human resources is constructed. The digital twin model of the group's human resources is used to depict the human resource mapping relationship between the group level, subsidiary level, and job level. Based on the group's human resources digital twin model, it receives organizational restructuring events or human resources planning requests, and simulates and extrapolates the corresponding organizational restructuring plan or human resources allocation plan in the group's human resources digital twin model to generate planning impact assessment results. Based on the planning impact assessment results, construct the local human resource allocation objective function for each subsidiary; By combining the global constraints at the group level, the objective function of local human resource allocation is solved collaboratively through a distributed game optimization method to form the group-level human resource planning decision results; The results of the group-level human resource planning decision-making, along with their corresponding deduction parameters and solution process information, are reliably stored. After the group-level human resource planning decision results are executed, the group human resource digital twin model and the local human resource allocation objective function are updated based on the execution feedback information.

2. The human resource planning and decision support method for group enterprises according to claim 1, characterized in that, The construction of a digital twin model of the group's human resources based on the group's organizational structure information, human resource allocation information, and job competency requirements information includes: The organizational structure information is parsed hierarchically to generate group-level organizational nodes, subsidiary-level organizational nodes, and job-level organizational nodes; Each organizational node is associated with personnel size constraint parameters, job competency matching constraint parameters, and permission boundary constraint parameters to form an organizational node constraint set; Based on the hierarchical and business collaboration relationships between the organizational nodes, establish human resource mapping rules to constrain the simulation. Based on the organizational nodes, the set of constraints of the organizational nodes, and the human resource mapping rules, a digital twin model of the group's human resources is generated.

3. The human resource planning and decision support method for group enterprises according to claim 1, characterized in that, The specific steps for constructing the objective function and the global constraints include: Based on the planning impact assessment results, objective functions are constructed for each subsidiary to characterize the rationality of personnel size, job matching degree, or human resource cost constraints. Based on the group's overall strategic goals and resource boundary conditions, a group-level global constraint function is constructed to limit the feasible solution space of the objective functions of each subsidiary; wherein, the objective functions of the subsidiaries and the global constraint function together constitute the basis for solving the distributed game optimization problem.

4. The human resource planning and decision support method for group enterprises according to claim 3, characterized in that, The process of collaboratively solving the local human resource allocation objective function through distributed game theory optimization to form a group-level human resource planning decision result includes: Under the premise of satisfying the global constraint function, an iterative update mechanism is used to solve the strategy variables corresponding to the objective function of each subsidiary in a distributed manner; In each iteration, the pre-set equilibrium convergence condition is determined based on the changes in each of the strategy variables. When the equilibrium convergence condition is met, the iterative solution is stopped, and the corresponding strategy combination is taken as the decision result of the group-level human resource planning.

5. The human resource planning and decision support method for group enterprises according to claim 1, characterized in that, The reliable storage of the group-level human resource planning decision results and their corresponding derivation parameters and solution process information includes: The summary of the group-level human resource planning decision results, the inference parameters, and the solution process information is extracted to generate evidence data. The evidence data is written into an immutable evidence carrier to form a decision evidence record corresponding to the group-level human resources planning decision results; The decision-making record allows for traceable verification of the formation process of the group-level human resource planning decision-making results.

6. The human resource planning and decision support method for group enterprises according to claim 1, characterized in that, After the group-level human resource planning decision is executed, the updating of the group's human resource digital twin model and the local human resource allocation objective function based on the execution feedback information includes: Collect information on actual organizational structure changes and human resource allocation results generated during the execution of the group-level human resource planning and decision-making results, as the execution feedback information; The execution feedback information is compared and analyzed with the planning impact assessment results to generate a planning deviation index; When the planning deviation index exceeds the preset threshold, the organizational node constraint parameters in the group human resources digital twin model or the local human resources configuration objective function are corrected, and the distributed game optimization is re-executed based on the corrected model.

7. The human resource planning and decision support method for group enterprises according to claim 6, characterized in that, Before modifying the constraint parameters in the group's human resources digital twin model or the objective function for local human resources allocation, the following steps are also included: The planning deviation index is determined by the type of deviation to distinguish whether the planning deviation index is caused by mismatch of organizational structure constraints or by deviation of human resource allocation optimization target; When the planning deviation index is determined to be of the organizational structure constraint mismatch type, the correction of personnel size constraint parameters, job competency matching constraint parameters or permission boundary constraint parameters in the group human resources digital twin model is triggered first. When the planning deviation index is determined to be of the type of human resource allocation optimization target deviation, the objective function parameters in the local human resource allocation objective function are corrected first. Based on the deviation type determination result, the execution order or combination of the constraint parameter correction and the objective function parameter correction is determined.

8. The human resource planning and decision support method for group enterprises according to claim 7, characterized in that, The step of determining the deviation type of the planning deviation index includes: Based on the distribution characteristics of the differences between the execution feedback information and the planning impact assessment results, a structural deviation analysis is performed on the planning deviation index. When the differential distribution characteristics are mainly concentrated in the range of organizational level changes, the range of job staffing deviations, or the range of authority coverage, the planning deviation index is determined to be an organizational structure constraint mismatch type. When the differential distribution characteristics are mainly concentrated in the range of changes in personnel allocation ratio, job matching degree, or human resource cost weight, the planning deviation index is determined to be a human resource allocation optimization target deviation type.

9. The human resource planning and decision support method for group enterprises according to claim 7, characterized in that, After determining the execution order or combination of the constraint parameter correction and the objective function parameter correction based on the deviation type determination result, the method further includes: The correction of the constraint parameters or the correction of the objective function parameters is carried out in stages according to the preset correction step size or correction ratio. After each correction phase is completed, the planning deviation index is recalculated, and it is determined whether the preset stability condition is met. When the stability condition is determined to be met, the subsequent correction phase is terminated.

10. A human resource planning and decision support system for group enterprises, used to execute the method described in any one of claims 1 to 9, characterized in that, include: The acquisition unit is used to acquire information on the group company's organizational structure, human resource allocation, and job competency requirements. The first construction unit is used to construct a digital twin model of the group's human resources based on the group's organizational structure information, human resource allocation information, and job competency requirements information. The digital twin model of the group's human resources is used to depict the human resource mapping relationship between the group level, subsidiary level, and job level. The simulation unit is used to receive organizational restructuring events or human resource planning requests based on the group's human resource digital twin model, and to simulate and simulate the corresponding organizational restructuring plan or human resource allocation plan in the group's human resource digital twin model to generate planning impact assessment results. The second construction unit is used to construct the local human resource allocation objective function for each subsidiary based on the planning impact assessment results; The solution unit is used to collaboratively solve the local human resource allocation objective function by combining the global constraints at the group level and through a distributed game optimization method, so as to form the group-level human resource planning decision results. The evidence storage unit is used to reliably store the group-level human resource planning decision results and their corresponding inference parameters and solution process information. The update unit is used to update the group-level human resources digital twin model and the local human resources configuration objective function based on the execution feedback information after the group-level human resources planning decision results are executed.

Images