Human resource management method and system

Abstract

The invention relates to a human resource management method and system, and the method comprises the steps: obtaining the full data or incremental data of a plurality of business systems in real time, so as to construct a human resource data set, and the human resource data set at least comprises real-time human resource change data; importing the human resource data set into a human cost prediction model, and performing accounting through the human cost prediction model to obtain a human cost prediction result matched with the current production condition and the human resource change; and executing a corresponding human resource allocation strategy based on the human cost prediction result. According to the method, the real-time data of the multi-source service system is integrated, the human cost prediction model is constructed, and the human resource allocation strategy is dynamically generated, so that accurate prediction of human cost and optimal configuration of resources are realized, and the efficiency and the intelligent level of human resource management in an industrial scene are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human resource management, in particular to a human resource management method and system. BACKGROUND

[0002] Please refer to Figure 1 , the existing technology of human cost prediction is generally calculated according to the number of workers in the last month and the actual salary in the last month to calculate the average salary per person in the last month, and then the human cost in the next month is estimated according to the average salary per person in the last month and the number of workers in the next month. This process is usually realized by Excel. However, the existing technology has obvious shortcomings in accuracy, real-time and intelligent level due to the problems of ignoring dynamic factors, easy errors in manual data input, and real-time synchronization of data between systems, etc., which makes it difficult to cope with complex labor scenarios, and cannot meet the fine management needs of human cost in industrial scenarios. SUMMARY

[0003] In view of the above technical problems, the present application provides a human resource management method and system, which integrates real-time data of multiple business systems, builds a human cost prediction model and dynamically generates a human resource allocation strategy, realizes accurate prediction of human cost and optimal allocation of resources, and significantly improves the efficiency and intelligent level of human resource management in industrial scenarios.

[0004] To solve the above technical problems, the present application provides a human resource management method, which comprises the following steps:

[0005] Real-time acquisition of full data or incremental data of multiple business systems to build a human resource data set, wherein the human resource data set at least includes real-time human resource change data;

[0006] Importing the human resource data set into a human cost prediction model to obtain a human cost prediction result matched with the current production conditions and human resource changes through the human cost prediction model;

[0007] Based on the human cost prediction result, a corresponding human resource allocation strategy is executed.

[0008] In some embodiments, the real-time acquisition of full data or incremental data of multiple business systems to build a human resource data set comprises at least two of the following:

[0009] Acquiring standard working hours through an enterprise resource planning system;

[0010] Obtaining model plan output through a production planning and scheduling system;

[0011] Obtaining worker types, number of workers, attendance hours and hourly wage through a human resource system;

[0012] Collecting actual output man-hours through a manufacturing execution system.

[0013] In some embodiments, before the human resource dataset is imported into the human cost prediction model, the method comprises:

[0014] Preprocessing the human resource dataset based on a multi-level data warehouse.

[0015] In some embodiments, before the human resource dataset is imported into the human cost prediction model, the method further comprises:

[0016] Constructing the human cost prediction model comprises:

[0017] Using a multiple linear regression model to analyze the relationship between human cost and production variables, and / or using a time series model to predict human demand.

[0018] In some embodiments, the human cost prediction result matching the current production conditions and human resource changes is calculated through the human cost prediction model, comprising:

[0019] Determining historical production efficiency according to historical actual output man-hours and historical attendance man-hours;

[0020] Classifying the historical production efficiency through cluster analysis and determining the time sequence characteristics of the historical production efficiency through time sequence analysis to output the production efficiency with the highest matching degree to the current production conditions;

[0021] Generating planned attendance man-hours based on the production efficiency, the planned production quantity of the current model, and the standard man-hours;

[0022] Dynamically adjusting the number of workers of each type according to the real-time human resource change data;

[0023] Generating the human cost prediction result in combination with the planned attendance man-hours, the dynamically adjusted number of workers of each type, and the man-hour unit price.

[0024] In some embodiments, after the human cost prediction result matching the current production conditions and human resource changes is calculated through the human cost prediction model, the method further comprises:

[0025] Based on the type of the to-be-generated device, determining the difference between the human cost prediction result corresponding to a single device and the single device budget to obtain a single device human cost deviation;

[0026] If the single device human cost deviation exceeds a preset threshold, triggering an abnormal warning and generating a corresponding optimization suggestion.

[0027] In some embodiments, the method further comprises:

[0028] receiving a natural language query sentence input by a user, the natural language query sentence being used to query human resource analysis data;

[0029] inputting the natural language query sentence into an intelligent dialogue tool to obtain a structured query sentence;

[0030] calling the human cost prediction model based on the structured query sentence to output a human resource analysis result corresponding to the natural language query sentence.

[0031] The application also provides a human resource management system, comprising a data acquisition module, a model prediction module and a resource allocation module; wherein,

[0032] The data acquisition module is configured to acquire full data or incremental data of a plurality of business systems in real time to construct a human resource dataset, the human resource dataset comprising at least real-time human resource change data;

[0033] The model prediction module is configured to import the human resource dataset into a human cost prediction model to obtain a human cost prediction result matched with current production conditions and human resource changes through the human cost prediction model;

[0034] The resource allocation module is configured to execute a corresponding human resource allocation strategy based on the human cost prediction result.

[0035] The application also provides a computing device comprising a storage medium and a controller, the storage medium storing a computer program, the computer program being executed by the controller to implement the steps of the method described above.

[0036] The application also provides a storage medium storing a computer program, the computer program being executed by a processor to implement the steps of the method described above.

[0037] The human resource management method of the application comprises: acquiring full data or incremental data of a plurality of business systems in real time to construct a human resource dataset, the human resource dataset comprising at least real-time human resource change data; importing the human resource dataset into a human cost prediction model to obtain a human cost prediction result matched with current production conditions and human resource changes through the human cost prediction model; and executing a corresponding human resource allocation strategy based on the human cost prediction result. The application integrates real-time data of multiple business systems, constructs a human cost prediction model and dynamically generates a human resource allocation strategy to realize accurate prediction of human cost and optimal allocation of resources, thereby significantly improving the efficiency and intelligent level of human resource management in industrial scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a flowchart of a human resource management method according to the prior art.

[0039] Figure 2 is an architecture diagram of a human resource management method according to an embodiment of the present application.

[0040] Figure 3 is a flowchart of a human resource management method according to an embodiment of the present application.

[0041] Figure 4 is a structure diagram of a human resource management system according to an embodiment of the present application.

[0042] Figure 5 is a structure diagram of a computing device according to an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to facilitate the understanding of the present application, the present application will be described more fully below with reference to the accompanying drawings. The preferred embodiments of the present application are shown in the drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present application can be more thoroughly and completely understood.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terminology used in the description of the present application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. In the present application, "each" includes one and more than one.

[0045] In the field of intelligent manufacturing, due to the complexity and diversity of production models and worker types, the existing human resource accounting method is low in efficiency and has certain limitations, which are specifically shown in the following aspects: 1. The existing technology is based on fixed parameters for prediction, ignoring dynamic factors such as worker entry / exit, capacity fluctuation, etc., resulting in a large deviation between the predicted results and the actual situation; 2. Human resource data relies on manual input, but manual entry is prone to errors and difficult to check; 3. Business system data cannot be synchronized in real time with attendance, performance and other systems, and needs to be updated manually frequently, which is low in efficiency and lagging; 4. When facing multiple types of work, differentiated salary structure or social security policy, Excel and other tools are difficult to accurately predict due to model simplification; 5. It is unable to automatically identify abnormal situations such as sudden increase in human cost, threshold exceeding of resignation rate, etc., and needs to rely on manual analysis, with a lagging response; 6. When multiple people collaborate, version confusion is easy to occur, and modification records are difficult to track, affecting data consistency and collaboration efficiency. The above problems collectively result in obvious short boards in accuracy, real-time performance and intelligence level of the existing technology, which is difficult to meet the needs of fine management of human cost in industrial scenarios.

[0046] First embodiment

[0047] In view of the above shortcomings and deficiencies of the prior art, the present application proposes a human resource management method, which establishes a quantitative relationship model of "work efficiency-work asset value-yield" under the premise of fixed monthly planned yield, to quickly solve the optimal human resource investment scheme, realize the maximization of direct labor cost-yield ratio and the minimization of wage cost, meet the constraints of human resource management KPI (Key Performance Indicator) such as yield per capita, work hour utilization rate, good product rate, etc., and can quickly output analysis and decision report through intelligent dialogue tools.

[0048] Figure 2 An architecture diagram of a human resource management method provided by an embodiment of the present application is shown. As shown in the figure, the architecture of the human resource management method provided by the embodiment of the present application mainly includes the following modules: Figure 2As shown, the architecture can include a business system, a big data platform, a human cost prediction model, and a chatBI intelligent conversation tool, etc. The business system can include SAP (Systems, Applications, and Products in Data Processing), APS (Advanced Planning and Scheduling), HR (Human Resources), MES (Manufacturing Execution System), and other subsystems. The enterprise resource planning system can be used to provide standard working hour data. The production planning and scheduling system can be used to provide planned output data. The human resources system can be used to provide worker type, hourly wage, and other data. The manufacturing execution system can be used to provide finished work reporting data. The data of each subsystem in the business system is collected and imported into the big data platform. The big data platform includes data collection, data warehouse, data model construction and solidification, and other modules. The data warehouse includes a DM layer (Data Mart Layer), a DWD layer (Data Warehouse Detailed), a DWS layer (Data Warehouse Service), and an ODS layer (Operational Data Store). The data collection module automatically extracts human resource data from the SAP, HR, APS, and other business systems through full and incremental methods, and stores it in the ODS layer of the data warehouse. The data warehouse module standardizes the data in the ODS, DWD, DWS, DM, and other layers. The data model construction and solidification module models the business data according to different classification methods such as business domain, theme domain, and index group, and solidifies the calculation logic and data source. The human cost prediction model is imported after the business data collected by the big data platform is preprocessed. The human cost prediction model is used to calculate past production efficiency, next month's model output, next month's planned output hours, next month's planned attendance hours, and other data, and to calculate next month's average attendance hours according to the number of workers of each type in the next month to estimate next month's personnel wages. On this basis, further estimates of single device human cost deviation are made, and human cost prediction situations are analyzed, such as generating recruitment, loan analysis reports, personnel vacancy anomaly information, and cost overrun anomaly information, etc. The output data of the human cost prediction model can be empowered through the chatBI intelligent conversation tool, such as users calling the human cost prediction model through conversation with the chatBI intelligent conversation tool to realize factory human cost analysis, workshop anomaly cause positioning, cost overrun warning, and personnel vacancy warning, and other functions.

[0049] Figure 3is a flowchart of a human resource management method according to an embodiment. As shown in Figure 3 The human resource management method according to the embodiment of the application comprises the following steps:

[0050] In step S1, full data or incremental data of a plurality of business systems are acquired in real time to construct a human resource dataset, which at least comprises real-time human resource change data.

[0051] The business systems comprise at least two of enterprise resource planning systems, production planning and scheduling systems, human resource systems and manufacturing execution systems. The full data or incremental data of the plurality of business systems are extracted in real time to form the human resource dataset, which is stored in the ODS layer of the data warehouse. For example, the full data of the plurality of business systems can be extracted when the human resource dataset is first constructed. Subsequently, in response to data updates in the business systems, incremental data of the business systems are extracted. In response to instructions for invoking the human resource prediction model, incremental data of the plurality of business systems can be acquired in real time. The incremental data of the plurality of business systems can also be acquired periodically (e.g., daily, weekly, monthly). A data synchronization tool such as Kafka can be deployed in the data collection module of the big data platform to capture data of the business systems in real time. The human resource dataset comprises not only basic information of human resources such as employee records and post configuration, but also real-time human resource change data such as new employee information, resignation records, job transfer approval and performance score updates, so as to avoid predicting human resource costs based on fixed human resource data and improve the accuracy of human resource prediction data.

[0052] In S2, the human resource dataset is imported into a human resource cost prediction model to obtain a human resource cost prediction result matched with the current production conditions and human resource changes through the human resource cost prediction model.

[0053] The human resource dataset is imported into the human resource cost prediction model, which can be constructed by a prediction modeling method. Prediction modeling is a technology for quantitatively predicting future events or trends through historical data and mathematical statistical methods. In the embodiment, a mathematical model can be constructed according to the business relationships between various historical data of the human resource dataset, and real-time data can be imported to predict the future trend of human resource costs. In the human resource cost prediction model, by integrating multi-source business data, machine learning algorithms or statistical models (such as regression analysis, time series prediction, random forest, etc.) are used to model the variable relationships in the data, so as to calculate a human resource cost prediction result highly matched with the current production conditions (such as business scale, project cycle, budget constraints) and human resource dynamics (such as resignation rate, new employee hiring speed, overtime length).

[0054] Specifically, the total human cost is calculated by position type, and the formula is:

[0055] Total human cost = ∑(single position cost)

[0056] Wherein, the single position cost is determined by the number of people, hourly wage and working hours, and the formula is:

[0057] Single position cost = number of people × hourly wage × working hours

[0058] Wherein, the working hours need to be dynamically calculated in combination with the model planned production and standard working hours (or production efficiency).

[0059] Based on the planned production, single standard working hours and worker efficiency coefficient, the model working hours are dynamically generated, and the formula is:

[0060] Model working hours = (model planned production × single standard working hours) / worker efficiency coefficient

[0061] Wherein, the standard working hours can be sourced from the SAP system or historical quota data; the worker efficiency changes dynamically with factors such as production capacity, number of workers, season, etc., and the efficiency coefficient is dynamically adjusted according to factors such as production capacity utilization, number of workers, seasonal fluctuations, etc. The average value under the same conditions can be taken from historical data. The weighted average value of historical efficiency can be extracted from historical data similar to the current production conditions (such as order type, equipment status, personnel scale). For example, if the current production conditions are 80% similar to a certain season in history, the average efficiency coefficient of that season is preferred.

[0062] Link the model planned production with the MES system to obtain real-time order changes or equipment failure information, dynamically correct the planned production and efficiency coefficient. According to the real-time human resource change data (such as pre-warning of resignation, temporary recruitment), dynamically adjust the number of people in each position to ensure that the model input is synchronized with the actual labor state.

[0063] Finally, through the above structured modeling and dynamic adjustment, the model can output the human cost and total human cost prediction results of each position type, and can also generate human resource prediction results regularly (such as monthly, quarterly), improving the accuracy and business response flexibility of human cost accounting.

[0064] S3, based on the human cost prediction results, execute the corresponding human resource deployment strategy.

[0065] The human resource allocation strategy refers to the plan for adjusting human resources according to the human cost prediction results and actual business needs. The adjustment methods include, but are not limited to, post allocation, personnel changes, and incentive mechanisms. For example, flexible employment modes such as labor dispatch, part-time, or outsourcing can be used to supplement short-term human resource gaps. For example, cross-department training can be conducted for existing employees to reduce dependence on external recruitment and optimize internal resources. For example, for high-risk positions predicted to have high turnover, a differentiated performance-driven incentive plan can be designed to reduce talent turnover costs. In addition, for work content that is inefficient or has a high error rate, automation tools can be introduced to reduce costs and increase efficiency. In this way, the corresponding human resource allocation strategy is executed based on the human cost prediction results, which can achieve the optimal balance between human cost and organizational efficiency, ensuring business continuity and improving resource utilization efficiency.

[0066] In some embodiments, full data or incremental data of multiple business systems are acquired in real time to construct a human resource dataset, including at least two of the following:

[0067] Standard man-hours are collected through an enterprise resource planning system;

[0068] Model planned production is obtained through a production planning and scheduling system;

[0069] Worker types, worker numbers, attendance man-hours, and man-hour unit prices are obtained through a human resource system;

[0070] Actual output man-hours are collected through a manufacturing execution system.

[0071] Here, the enterprise resource planning system can be used to provide standard man-hour data. Standard man-hour data refers to the unit output time benchmark value required for a specific post, process, or task based on industry benchmarks, historical operation records, or engineering analysis.

[0072] The production planning and scheduling system is a management tool for production processes that optimally allocates raw materials and production capacity to meet various needs. Model planned production is the expected production quantity target for a specific product model based on market demand and production resource allocation.

[0073] The human resource system is a digital platform that integrates employee information management, recruitment, compensation, performance, and other functions, achieving full life cycle management of human resources. Worker types refer to specific skill categories or post responsibilities that workers engage in, such as welders, electricians, and assembly workers. Worker numbers refer to the number of employees assigned to a certain type of work in the production process, reflecting the size of the post manpower demand. Attendance man-hours refer to the total effective working hours of employees actually attending and used for production within a statistical period, usually measured in hours. Man-hour unit price refers to the human cost or remuneration standard per unit of man-hour, used to calculate labor costs or calculate labor efficiency.

[0074] Manufacturing execution system is a digital system connecting production planning and shop floor execution, which monitors production progress, resource allocation and quality control in real time. Actual output man-hour refers to the total man-hour consumed to complete actual production tasks, reflecting the effective labor input in the production process.

[0075] In some embodiments, before importing the human resource dataset into the human cost prediction model, the human resource management method comprises:

[0076] The human resource dataset is preprocessed based on the multi-level data warehouse.

[0077] Before importing the human resource dataset into the human cost prediction model, various types of business data in the human resource dataset need to be preprocessed. The data can be extracted, transformed and loaded from the source end to the destination end through an ETL (Extract-Transform-Load) tool (such as Informatica, Talend). In this embodiment, the source end is a plurality of business systems, and the destination end is a data warehouse. Data cleaning includes deleting duplicate records (such as eliminating redundancy by comparing employee ID, name, ID number and other fields), filling missing values (such as predicting missing data using mean, median or machine learning model), correcting abnormal values (such as identifying unreasonable salary or abnormal attendance length), and standardizing format (such as unifying date format and salary unit).

[0078] The data warehouse includes ODS layer, DWD layer, DWS layer and DM layer. Specifically, the human resource dataset is first extracted from the business system and stored in the ODS layer of the data warehouse. Key fields such as employee number, post category and work hour data are extracted from the ODS layer through the ETL tool. Then, data cleaning, dimension association (cross-business system data association) and standardization processing are performed in the DWD layer. The cleaned human resource data is aggregated to the DWM layer according to business requirements, such as monthly department human cost statistics. Finally, the human resource data is loaded into the DM layer to generate a data mart required for human cost prediction, which is directly called by the human cost prediction model. It should be noted that the data mart is a subset of the human resource dataset, which is stored in a multi-dimensional manner to meet the needs of a specific department or user, including defining dimensions, indicators to be calculated, dimension levels, etc., generating a data cube for decision analysis requirements.

[0079] In some embodiments, before importing the human resource dataset into the human cost prediction model, the human resource management method further comprises:

[0080] The human cost prediction model is constructed, comprising:

[0081] The relationship between human cost and production variables is analyzed by using a multiple linear regression model, and / or the human demand is predicted by using a time series model.

[0082] Here, the human cost prediction model needs to be built before importing the human resource dataset. The human cost prediction model can be built by a prediction modeling method. Specifically, first, the relationship between human cost and production variables (such as production and scheduling) is analyzed by using a multiple linear regression model. Historical data (such as the number of employees, hourly wage, production efficiency, equipment utilization rate, etc.) is collected and a regression equation is established, where the independent variable is the production variable and the dependent variable is the human cost.

[0083] At the same time, the future human demand is predicted by using a time series model, which can be ARIMA (Autoregressive Integrated Moving Average) or SARIMA (Autoregressive Integrated Moving Average). Historical human resource changes and / or historical human demand data are input into the time series model for decomposition to output future human demand data or human resource changes. For example, temporary employees are needed during the peak holiday period to avoid labor shortages.

[0084] In this way, by creating a hybrid model based on multiple linear regression and time series model, the multiple regression model reveals the impact of production variables on human cost, while the time series model captures the dynamic change rule of human cost. The dual model collaborative structure of the human cost prediction model can significantly improve the accuracy of human cost prediction.

[0085] In some embodiments, the human cost prediction model calculates the human cost prediction result matched with the current production conditions and human resource changes, including:

[0086] The historical production efficiency is determined according to the historical actual output hours and historical attendance hours;

[0087] The historical production efficiency is classified by cluster analysis, and the time sequence characteristics of the historical production efficiency are determined by time series analysis to output the production efficiency with the highest matching degree to the current production conditions;

[0088] Based on the production efficiency, the planned production quantity of the current model, and the standard working hours, the planned attendance working hours are generated;

[0089] According to the real-time human resource change data, the number of workers of each type is dynamically adjusted;

[0090] The human cost prediction result is generated in combination with the planned attendance working hours, the dynamically adjusted number of workers of each type, and the hourly wage.

[0091] In this embodiment, the historical production efficiency is first calculated. The historical production efficiency can be calculated based on the historical actual output working hours and the historical attendance working hours, and the formula is: historical production efficiency = actual output working hours / attendance working hours. Further, the historical production efficiency can be dynamically calculated by using a sliding time window (such as a weekly / monthly rolling window) to capture the short-term fluctuations and long-term trends of the historical production efficiency.

[0092] Secondly, the K-means or hierarchical clustering algorithm can be used to cluster and classify the historical production efficiency, such as dividing different types of products into high-efficiency groups, medium-efficiency groups, and low-efficiency groups, and combining time series analysis to extract the periodicity, trend, and seasonality characteristics of each group of data to derive the future production efficiency. For example, if the historical production efficiency of a certain type of product decreases by 30% during equipment maintenance and increases by 15% during the peak season, the time series matching algorithm is used to evaluate the similarity between the current production conditions and the historical efficiency production conditions, and the highest matching efficiency category is output as the current predicted production efficiency.

[0093] Subsequently, the planned attendance working hours are generated. According to the model production quantity and the standard working hours, the planned output working hours are determined, and the formula is: planned output working hours = model production quantity of the model * standard working hours of the model. The standard working hours of the model refer to the standard working hours required to produce a specific model product. Then, according to the current production efficiency and the planned output working hours, the planned attendance working hours are determined, and the formula is: planned attendance working hours = planned output working hours / current production efficiency.

[0094] Further, by collecting real-time human resource change data such as hiring, resignation, loan, and job transfer, the number of each type of worker in the next month is estimated, and the formula is: next month's number of each type of worker = this month's number of each type of worker + number of people to be hired - number of people to be resigned + number of people to be loaned - number of people to be transferred.

[0095] Finally, the planned attendance working hours, the dynamically adjusted number of workers of each type, and the hourly wage are combined to calculate the average attendance working hours, and the formula is: average attendance working hours = planned attendance working hours / next month's number of each type of worker. Based on the hourly wage, the next month's personnel salary prediction is generated, and the formula is: next month's personnel salary = (number of each type of worker) * average attendance working hours * hourly wage.

[0096] In this way, by integrating the human resource data of various business systems and dynamically adjusting according to the production plan and personnel changes, accurate human cost prediction results can be obtained.

[0097] In some embodiments, after the human cost prediction model calculates the human cost prediction results that match the current production conditions and human resource changes, the human resource management method further includes:

[0098] Based on the type of the device to be generated, the difference between the human cost prediction result corresponding to a single device and the single device budget is determined to obtain a single device human cost deviation;

[0099] If the single device human cost deviation exceeds the preset threshold, an abnormal warning is triggered and a corresponding optimization suggestion is generated.

[0100] First, the human cost deviation is calculated: according to the type of the device (such as mechanical processing equipment, automated production line, etc.), the corresponding single device human cost prediction value is determined, and compared with the preset single device budget, the difference formula is:

[0101] Human cost deviation = (each model) single device human cost prediction value - single budget value

[0102] Among them, the single device budget can be determined according to the industry standard or the benchmark value of the historical cost data, and the single device human cost prediction value can be output by the human resource prediction model, and the formula is:

[0103] (Each model) single human cost prediction value = each model planned output working hours * planned output working hours unit price

[0104] Planned output working hours unit price = estimated next month personnel salary / planned output working hours

[0105] Second, set the preset threshold, which can be based on the type of the device and industry experience, and adopt hierarchical threshold rules, such as setting the threshold value as ± 5% of the budget value, for a budget value of 200,000, the preset threshold is 10,000. If the deviation exceeds the preset threshold, an abnormal warning is triggered and an optimization suggestion is generated simultaneously. If the estimated single human cost deviation is too large in the positive direction, that is, the single device human cost prediction value is greater than the single device budget value, it means that the human cost is over budget, and measures need to be taken to reduce human input or borrow personnel to other factories, workshops. If the estimated single human cost deviation is too large in the negative direction, that is, the single device human cost prediction value is less than the single device budget value, it means that the human input is too low and the personnel is short, which may lead to failure to achieve production target, product quality not up to standard, and personnel overload. In this way, real-time monitoring and dynamic correction of device human cost can be realized to ensure that the cost deviation is controlled within a reasonable range.

[0106] In some embodiments, the human resource management method further comprises:

[0107] Receiving a natural language query sentence input by a user, the natural language query sentence being used to query human resource analysis data;

[0108] Inputting the natural language query sentence into an intelligent dialogue tool to obtain a structured query sentence;

[0109] The human cost prediction model is called based on the structured query statement to output the human resource analysis result corresponding to the natural language query statement.

[0110] In this embodiment, the chatBI (business intelligence) intelligent conversation tool can be used to convert the user input question into a structured query (such as SQL) through natural language processing (NLP), understand the context based on the generative AI, generate answers or explain data trends, connect enterprise data sources, and present the results in the form of charts, tables, etc. Of course, deep learning-based natural language processing models such as BERT, GPT-4, or Llama series can also be used.

[0111] First, the natural language query statement input by the user is parsed by the intelligent conversation tool, for example, "What is the human cost prediction value and budget deviation of the machining equipment in Q2 2025?". The intelligent conversation tool performs semantic understanding, extracts key elements (such as time range, equipment type, query target), and locates the scope (such as "machining equipment") and indicators (such as "human cost prediction value" and "budget deviation") involved in the query through entity recognition technology. Then, the parsed natural language intent is converted into a structured query statement (such as SQL or a custom query syntax). For example, for the above query, the system generates a structured instruction similar to the following:

[0112] SELECT equipment type, predicted human cost, budget value, ABS(predicted human cost - budget value) AS human cost deviation

[0113] FROM human cost prediction table

[0114] WHERE equipment type ='machining equipment' AND time range = '2025-Q2'

[0115] Subsequently, the structured query statement is input into the pre-set human cost prediction model, and human resource data matching the structured query statement is extracted, such as model planned production, standard working hours, and historical efficiency coefficient. The model calculates the human cost prediction value, the human cost prediction value of a single device, and the human cost deviation compared with the preset budget value.

[0116] Of course, in response to different natural language query statements input by the user, the model can also output human cost analysis results such as recruitment, loan analysis reports, personnel vacancy anomaly information, cost overrun anomaly information, etc. In addition, business personnel can perform operation analysis through chatBI, such as factory human cost analysis, workshop human cost anomaly cause positioning, workshop personnel vacancy or human cost overrun early warning, model human cost deviation too large early warning, etc. Finally, the model output result is returned to the user through a visual interface or an API interface. In this way, through the cooperation of natural language processing, structured query generation and prediction model, flexible query and accurate response of human resource analysis data are realized, which significantly reduces the user operation threshold and improves the decision-making efficiency.

[0117] In the human resource management method of the present application, first, in view of the trend of expansion of industrial environment scale and complication of business scenarios, the system automatically collects and integrates multi-source human cost data through a big data platform, uses an ETL tool to realize data cleaning and standardization, and significantly improves production and operation efficiency. Secondly, different positions, hourly wages and other conditions are divided into unified job grades through cluster analysis, and historical data are classified in combination with factors such as season, output, and number of people, to realize cost granular management; further, time series analysis is used to mine the time dependence law of human cost, to capture trend, seasonal and periodic patterns. On this basis, a prediction model is constructed, which correlates human cost with variables such as output and scheduling, to dynamically predict future demand. Finally, through an intelligent dialogue tool, the user can directly query human cost data, generate decision reports or obtain analysis insights through natural language dialogue, without the need to write SQL or use professional BI tools, breaking through the limitations of traditional Excel and other static tools in dynamic factor integration, real-time, complex scenario processing and risk warning, and realizing the automation, intelligentization and real-time collaboration of human cost management. In this way, through the human resource management method of the present application, the complex, dynamic and fine-grained needs of human cost management in industrial scenarios are effectively solved.

[0118] The human resource management method of the present application comprises: acquiring full data or incremental data of a plurality of business systems in real time to construct a human resource data set, the human resource data set at least comprising real-time human resource change data; importing the human resource data set into a human cost prediction model to obtain a human cost prediction result matched with current production conditions and human resource changes through the human cost prediction model; and executing a corresponding human resource deployment strategy based on the human cost prediction result. The present application integrates real-time data of multiple business systems, constructs a human cost prediction model and dynamically generates a human resource deployment strategy, realizes accurate prediction of human cost and optimal allocation of resources, and significantly improves the efficiency and intelligent level of human resource management in industrial scenarios.

[0119] Second embodiment

[0120] Figure 4 is a structural schematic diagram of a human resource management system according to an embodiment. As shown in Figure 4 the human resource management system 20 of the present application includes a data acquisition module 201, a model prediction module 202 and a resource allocation module 203; wherein,

[0121] The data acquisition module 201 is configured to acquire full data or incremental data of a plurality of business systems in real time to construct a human resource data set, wherein the human resource data set at least includes real-time human resource change data.

[0122] The model prediction module 202 is configured to import the human resource data set into a human cost prediction model, and obtain a human cost prediction result matched with the current production condition and the human resource change through the human cost prediction model.

[0123] The resource allocation module 203 is configured to execute a corresponding human resource allocation strategy based on the human cost prediction result.

[0124] In some embodiments, the full data or incremental data of a plurality of business systems is acquired in real time to construct a human resource data set, including at least two of the following:

[0125] The standard working hours are collected through the enterprise resource planning system;

[0126] The model output is obtained through the production planning scheduling system;

[0127] The worker type, the number of workers, the attendance working hours and the working hour unit price are obtained through the human resource system;

[0128] The actual output working hours are collected through the manufacturing execution system.

[0129] In some embodiments, before the human resource data set is imported into the human cost prediction model, the data acquisition module 201 is further configured to:

[0130] Preprocess the human resource data set based on a multi-level data warehouse.

[0131] In some embodiments, before the human resource data set is imported into the human cost prediction model, the model prediction module 202 is further configured to:

[0132] Construct the human cost prediction model, including:

[0133] Use a multiple linear regression model to analyze the relationship between human cost and production variables, and / or use a time series model to predict human demand.

[0134] In some embodiments, the human cost prediction result matched with the current production condition and the human resource change is obtained through the human cost prediction model, including:

[0135] determine historical production efficiency according to historical actual output man-hours and historical attendance man-hours;

[0136] classify the historical production efficiency through cluster analysis, and determine time sequence characteristics of the historical production efficiency through time sequence analysis, to output a production efficiency with the highest matching degree to current production conditions;

[0137] generate planned attendance man-hours based on the production efficiency, current model planned output and standard man-hours;

[0138] dynamically adjust the number of workers of each type according to real-time human resource change data;

[0139] generate human cost prediction results in combination with the planned attendance man-hours, the dynamically adjusted number of workers of each type and the man-hour unit price.

[0140] In some embodiments, after obtaining the human cost prediction results matched to the current production conditions and human resource changes through the human cost prediction model, the model prediction module 202 is further configured to:

[0141] determine a difference between the human cost prediction results of a single device and the budget of a single device based on the type of the device to be generated, to obtain a single device human cost deviation;

[0142] if the single device human cost deviation exceeds a preset threshold, triggering an abnormal warning and generating corresponding optimization suggestions.

[0143] In some embodiments, the human resource management system 20 further includes a data query module (not shown in the figure), which is configured to:

[0144] receive a natural language query sentence input by a user, the natural language query sentence being used to query human resource analysis data;

[0145] input the natural language query sentence into an intelligent dialogue tool to obtain a structured query sentence;

[0146] invoke the human cost prediction model based on the structured query sentence, to output human resource analysis results corresponding to the natural language query sentence.

[0147] For specific implementation manners of the present embodiment, reference can be made to the related descriptions of the first embodiment, which will not be repeated here.

[0148] Based on the same inventive concept as the foregoing embodiments, the present embodiment provides a computing device, as shown in Figure 5 the computing device includes a processor 410 and a memory 411 storing a computer program; and Figure 5The processor 410 in the figure is not used to refer to the number of processors 410 being one, but is only used to refer to the positional relationship of the processor 410 relative to other devices. In actual application, the number of processors 410 can be one or more; similarly, Figure 5 The memory 411 in the figure also has the same meaning, that is, it is only used to refer to the positional relationship of the memory 411 relative to other devices. In actual application, the number of memories 411 can be one or more. When the processor 410 runs the computer program, the human resource management method described above is implemented.

[0149] The computing device can also include at least one network interface 412. The various components in the computing device are coupled together by a bus system 413. It can be understood that the bus system 413 is used to realize the connection communication between the components. In addition to including a data bus, the bus system 413 also includes a power bus, a control bus and a status signal bus. However, for the sake of clear illustration, all kinds of buses are marked as the bus system 413 in the figure. Figure 5

[0150] ​The memory 411 can be a volatile memory or a non-volatile memory, and can include both a volatile and a non-volatile memory. The non-volatile memory can be a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a Compact Disc Read-Only Memory (CD-ROM). The magnetic surface memory can be a magnetic disk memory or a magnetic tape memory. The volatile memory can be a Random Access Memory (RAM) used as an external cache. By way of example, but not limitation, many forms of RAM can be used, such as a Static Random Access Memory (SRAM), a Synchronous Static Random Access Memory (SSRAM), a Dynamic Random Access Memory (DRAM), a Synchronous Dynamic Random Access Memory (SDRAM), a Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), an Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), a Sync Link Dynamic Random Access Memory (SLDRAM), a Direct Rambus Random Access Memory (DRRAM).The memory 411 described in the embodiments of the present application is intended to include, but is not limited to, these and any other suitable type of memory.

[0151] The memory 411 in the embodiments of the present application is used to store various types of data to support the operation of the computing device. Examples of these data include: any computer programs for operating on the computing device, such as operating systems and application programs; contact data; phonebook data; messages; pictures; videos; and the like. Among them, the operating system contains various system programs, such as framework layer, core library layer, driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application program can contain various application programs, such as media players, browsers, etc., for implementing various application services. Here, the program for implementing the method of the embodiments of the present application can be contained in the application program.

[0152] Based on the same inventive concept as the foregoing embodiments, the present embodiment also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer readable storage medium can be a ferromagnetic random access memory (FRAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc. The computer readable storage medium can also be various devices including one or any combination of the above memories, such as a mobile phone, a computer, a tablet device, a personal digital assistant, etc. The computer program stored in the computer readable storage medium is run by a processor to implement the human resource management method applied to the computing device described above. The specific step flow implemented by the computer program executed by the processor is described in the embodiments of the present application. Figure 1 The description of the embodiments shown in the foregoing is not repeated here.

[0153] Each technical feature of the above-described embodiments can be combined arbitrarily, and to make the description concise, each technical feature of the above-described embodiments is not described in all possible combinations, however, as long as the combination of these technical features does not exist contradictory, it should be considered as the scope of the present disclosure.

[0154] In this document, the terms "comprise", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0155] The above description is only specific embodiments of the application, but the protection scope of the application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the application, which should be covered by the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.

Claims

1. A human resource management method, characterized in that: The method comprises the following steps: Acquire full data or incremental data from multiple business systems in real time to construct a human resources data set, which at least includes real-time human resources change data; Importing the human resource data set into a human resource cost prediction model, and calculating through the human resource cost prediction model to obtain a human resource cost prediction result that matches current production conditions and human resource changes; Execute corresponding human resource deployment strategies based on the human resource cost prediction results.

2. The method according to claim 1, characterized in that The real-time acquisition of full or incremental data from multiple business systems to construct a human resources dataset includes at least two of the following: Collect standard working hours through the enterprise resource planning system; Obtain the planned production volume of the model through the production planning and scheduling system; Obtain workers' job types, number of workers in each job type, attendance hours, and hourly rates through the human resources system; Actual output hours are collected through the Manufacturing Execution System.

3. The method according to claim 1 or 2, characterized in that Before importing the human resource dataset into the human resource cost prediction model, the method includes: The human resources data set is preprocessed based on a multi-level data warehouse.

4. The method according to claim 1, wherein Before importing the human resource dataset into the human resource cost prediction model, the method further includes: Constructing the labor cost prediction model includes: Use multiple linear regression models to analyze the relationship between labor costs and production variables, and / or use time series models to predict labor demand.

5. The method according to claim 2, characterized in that The labor cost forecast results calculated by the labor cost forecast model and matching the current production conditions and human resource changes include: Determine historical production efficiency based on historical actual output working hours and historical attendance working hours; Classifying the historical production efficiency through cluster analysis, and determining the time series characteristics of the historical production efficiency through time series analysis, so as to output the production efficiency that best matches the current production conditions; Generate planned attendance hours based on the production efficiency, the planned output of the current machine model and the standard working hours; Dynamically adjust the number of people in each job type based on the real-time human resources change data; The labor cost forecast result is generated by combining the planned attendance working hours, the dynamically adjusted number of workers and the working hour unit price.

6. The method according to claim 5, characterized in that After obtaining a labor cost forecast result that matches current production conditions and human resource changes through the labor cost forecast model, the method further includes: Based on the type of the device to be generated, determining the difference between the labor cost prediction result corresponding to the single device and the single device budget to obtain the labor cost deviation of the single device; If the labor cost deviation of the single device exceeds the preset threshold, an abnormal warning is triggered and corresponding optimization suggestions are generated.

7. The method according to claim 1, characterized in that The method further comprises: receiving a natural language query statement input by a user, wherein the natural language query statement is used to query human resources analysis data; Inputting the natural language query statement into an intelligent dialogue tool to obtain a structured query statement; The human cost prediction model is called based on the structured query statement to output a human resource analysis result corresponding to the natural language query statement.

8. A human resource management system, characterized in that: It includes data acquisition module, model prediction module and resource allocation module; among them, The data acquisition module is used to acquire full data or incremental data from multiple business systems in real time to construct a human resources data set, which at least includes real-time human resources change data; The model prediction module is used to import the human resource data set into the human resource cost prediction model, and calculate the human resource cost prediction results that match the current production conditions and human resource changes through the human resource cost prediction model; The resource allocation module is used to execute a corresponding human resource allocation strategy based on the human resource cost prediction result.

9. A computing device, characterized in that The method comprises a storage medium and a controller, wherein a computer program is stored on the storage medium, and when the computer program is executed by the controller, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.

Images