CN113792955A

CN113792955A - Human resource supply and demand simulation method based on hierarchical path and differential search algorithm

Info

Publication number: CN113792955A
Application number: CN202110836036.3A
Authority: CN
Inventors: 杜红军; 姚敬军; 孙刚; 李明; 周武明; 于海燕; 郭长彪; 叶丹; 刘宇; 张莹; 张善; 刘鹏; 孙玉坤; 张福峰; 孙秀一
Original assignee: Anshan Power Supply Co Of State Grid Liaoning Electric Power Co; State Grid Corp of China SGCC
Current assignee: Anshan Power Supply Co Of State Grid Liaoning Electric Power Co; State Grid Corp of China SGCC
Priority date: 2021-07-23
Filing date: 2021-07-23
Publication date: 2021-12-14

Abstract

The invention provides a human resource supply and demand simulation method based on hierarchical paths and a differential search algorithm, which simulates and deduces retirement, natural staff reduction, recruitment personnel and enterprise internal post mobilization and promotion of different hierarchical units, different service types and different employment forms, and simulated data support multi-level and multi-dimensional statistical analysis. The method comprises an automatic simulation hierarchical path method and a personnel random extraction method which is based on a differential search algorithm and meets probability distribution conditions. Simulating the ways of increasing, decreasing and flowing of employees from top to bottom in a sequence of first-in and last-in from low to high; the feature selection of a differential search algorithm is adopted to perform dimension reduction work, and when public personnel are automatically extracted and distributed, the simulation process of automatic extraction and flow of the personnel is realized. The technical problem in a prediction model is solved, a reasonable scheme for simulating staff export, staff entry and staff internal flow is formulated through data analysis and technical construction, and accurate human resource balance analysis to departments and posts is achieved.

Description

Human resource supply and demand simulation method based on hierarchical path and differential search algorithm

Technical Field

The invention relates to the technical field of human resource supply and demand, in particular to a human resource supply and demand simulation method based on hierarchical paths and a differential search algorithm.

Background

At present, the process of establishing a human resource employment strategy of a plurality of enterprises in China is rough, a manual statistical prediction mode is adopted by Excel, a scientific method and a professional informatization tool are not provided for support, and the method is difficult to adapt to the processing and massive calculation of a large amount of basic information of the human resource employment strategy with large scale, numerous personnel, complex structure and frequent flow of the enterprises, so that the problems that the total amount prediction deviation is large, the analysis granularity is not fine, the medium-long term prediction is inaccurate, the analysis and evaluation of the personnel structure cannot be specific to departments and posts and the like exist generally, and the method is difficult to adapt to the development requirements of the company in new periods, new situations and new targets. A human resource demand configuration prediction model which is high in refinement degree and strong in operability and meets the analysis requirements of big data information needs to be constructed, and a technical means support is provided for a company to accurately make a labor employment strategy. However, the following technical problems still exist:

1) in actual work, the personnel entering and exiting and flowing all the year round are scattered, the simulation operation needs to be carried out all at once, reasonable logic needs to be researched and formulated, and the increase, decrease and change of the personnel of each level of unit are orderly processed so as to carry out the simulation operation. Therefore, an automatic simulation hierarchical path method is researched to solve the above problems.

2) Although the human resource planning of the existing power grid enterprises can think and conspire the human resource team development and the human resource management work global from a strategic height, the human resource planning prediction cannot be closely combined with the production operation development to form a specific and reasonable human resource allocation; the recruitment of new employees is configured according to the conditions of over-determined employees, but cannot be effectively matched with the quality requirements needed by over-lacking posts. Therefore, a scientific, reasonable and comprehensive human resource demand prediction method is urgently needed to solve the problem of human resource allocation. When the personnel automatically extract the flow, the most common technical problem in the simulation process of realizing the automatic flow extraction of the personnel is to study the random extraction problem of the personnel meeting the probability distribution condition. However, as the dimension increases, when the probability combinations are too many thousands (e.g., combination of gender (2), age group (8), school calendar (5), school level (7), school specialty (4) ═ 2240), the minimum probability distribution and the maximum probability distribution are different by ten thousand times, the above method is not applicable. Therefore, an optimization algorithm needs to be researched to solve the above problems.

In the prior art, a method for predicting labor demand increment by using LSTM modeling is proposed in patent publication No. CN109345021A, a regional labor demand increment prediction method is proposed in patent publication No. CN108428014A, a flexible labor data processing method, device and system based on block chains is proposed in patent publication No. CN112053090A, and a real-time labor management system and management method are proposed in patent publication No. CN104361445B, which do not solve the technical problems in the prediction model.

Disclosure of Invention

In order to solve the technical problems provided by the background technology, the invention provides a human resource supply and demand simulation method based on a hierarchical path and a differential search algorithm, which solves the technical problems in constructing a human resource deduction model and takes 'automatic simulation of the hierarchical path' and 'random extraction of personnel meeting probability distribution conditions based on the differential search algorithm' as the key simulation distribution technology. Through data analysis and technical construction, a reasonable scheme for simulating staff export, staff entry and staff internal flow is formulated, and accurate analysis of human resource balance to departments and posts is realized.

In order to achieve the purpose, the invention adopts the following technical scheme:

a human resource supply and demand simulation method based on hierarchical paths and a differential search algorithm simulates and deduces retirement, natural staff reduction, recruitment and enterprise internal post mobilization and promotion of different hierarchical units, different service types and different employment forms, and simulated data support multi-level and multi-dimensional statistical analysis.

The simulation method comprises an automatic simulation hierarchical path method and a personnel random extraction method which is based on a differential search algorithm and meets probability distribution conditions.

The method for automatically simulating the hierarchical path comprises the following steps: simulating the increase, decrease and flow paths of workers from top to bottom in the sequence of first-in and last-in from low to high; the first-in and last-out are operations of a staff exit first when the human resource planning simulation operation is carried out, namely, retirement operation and natural staff reduction operation are carried out first, and then staff entry work of transfer officer distribution, specialist student distribution, agricultural and electrician distribution, labor dispatching distribution and student research distribution is carried out; from top to bottom, simulation operation is carried out according to the sequence of retirement, natural staff reduction, two-line withdrawal, cadre subsidiaries, subsidiaries of collective enterprises, diversion officer allocation, specialist student allocation, agro-electrician allocation, labor dispatching allocation, student allocation and staff flow of the second part; the low-to-high mode refers to the flowing process from county level to provincial level company when the second part of staff flow simulation operation; wherein, two lines are returned, the subsidiaries of cadres and the subsidiaries of the collective enterprise flow to the first part of the employees;

through the employee export simulation operation, the employee entry simulation operation and the employee flow simulation operation, an employee condition database at the end of each year of the next five years is formed, various statistical query and analysis works can be carried out on the basis of the database, and data support is provided for manpower resource planning;

the random extraction method of the personnel meeting the probability distribution condition adopts the characteristic selection of a differential search algorithm to perform the dimension reduction work, and realizes the simulation process of automatic extraction and flow of the personnel when the personnel are randomly extracted and distributed.

Further, the simulation process of simulating the ways of increasing, decreasing and flowing of the employees from top to bottom in a sequence of first-in and last-out and from low to high comprises the following specific steps:

step 1: the simulation process starts from employee export, firstly, retirement is simulated, the system automatically carries out retirement reduction treatment according to the specialty of gender and the date of birth, simulated retired persons in five years in the future are formed, retirement time is marked, and simulated retirement operation is completed;

step 2: then simulating natural staff reducers, summarizing the conditions of the sex natural staff reducers of all age groups according to historical natural staff reducer conditions, calculating the natural staff reducer probability and the most possible natural staff reducers of each unit according to historical conditions, and simulating the number of the staff reducers and the number of the staff reducers of the natural staff reducers;

and step 3: after the employee export simulation is completed, performing a first part of employee flow simulation, firstly simulating two-line withdrawal, automatically calculating the simulated two-line withdrawal personnel every year in the next five years according to the gender, the age and the post system, marking the two-line withdrawal time, and completing the simulated two-line withdrawal operation;

and 4, step 4: corresponding cadre posts can be vacated after the simulation returns the two lines, simulation operation of cadre subsidiaries is carried out on the cadre posts, the cadre subsidiaries are selected from internal personnel or obtained by flowing from a lower-level unit, personnel meeting the conditions are extracted according to various requirements of cadre subsidiary plan, the annual number of the subsidiaries of each unit is small according to the previous situation, the personnel meeting the conditions can be randomly extracted to determine the personnel to be subsidiarily carried out by the staff during the simulation of the subsidiaries, and a final simulation cadre subsidiary result is formed;

and 5: the method is characterized in that a researcher simulating the collective enterprise performs internal personnel movement according to the simulation staff-reducing result of the collective enterprise, and randomly extracts personnel of the collective enterprise to simulate the operation of filling the vacant posts of the staff-reducing;

step 6: after the flow simulation of the first part of employees is finished, the simulation operation of the employee entrance is carried out;

601) the transfer officers at the employee entrances distribute simulation operations, and the number of units and persons receiving the transfer officers every year is fixed according to historical conditions, so that the simulation operations are realized by adding a corresponding number of persons to units meeting the conditions according to the plan of the subsidiaries, and the simulation operations of the transfer officers in the next five years are completed;

602) the specialist allocation simulation operation of the employee entrance is carried out according to the mender plan and the allocation plan of each city-level unit every year;

603) the rural power and electricity worker distribution simulation operation of the employee entrance is carried out according to the rural power and electricity worker researcher plan and the distribution plan of each city level unit every year;

604) the labor dispatching personnel distribution simulation operation of the staff entrance is carried out according to the repair personnel plan and the distribution plan of each city-level unit every year;

605) the student allocation simulation operation of the employee entrance is divided into two parts, namely a provincial department and a directly subordinate unit allocation simulation operation and a city level unit allocation simulation operation, wherein the provincial department and the directly subordinate unit are only allocated to students, and the city level unit is allocated to students and the students; the simulation operation of the provincial department and the affiliated units is performed according to the distribution plan and the demand plan of each unit;

and 7: after the simulation of the employee exit and entrance is completed, the simulation operation of the second part of employee flow is carried out, the simulation operation is based on the principle of from low to high, namely in four levels of county bureaus, basic level units, implementation mechanisms, directly subordinate units and provincial companies, the flow direction is from the skill posts to the management posts in the level, the skill posts flow to the upper level units, and the management posts flow to the upper level units; the flow mode is that according to the staff flow distribution plan and the staff flow demand plan, through the staff flow distribution processing function, the staff flowing through various paths are simulated and extracted by using the intelligent optimization algorithm selected by the seat distribution characteristics, and the staff flow simulation operation in the next five years is completed.

Further, the specialist distribution simulation operation of the employee portal specifically includes: the method comprises the following steps that a specialist is a farm electric station from a researcher to a county level, a researcher plan is to maintain the number of the researchers fixed in each city level unit every year, a distribution plan is to simulate and calculate the number of the missing members and the occupation ratio of the missing members of the farm electric station under each city level unit every year, and the distribution number of each farm electric station is calculated by combining a preset distribution coefficient; the specialist allocation simulation operation simulates and extracts specialist personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulates and allocates the specialist personnel to agricultural and electric farms subordinate to each city company, and completes the annual specialist allocation simulation operation in the next five years.

Further, the simulation operation of the rural power distribution of the employee entrance specifically comprises: the method comprises the steps that farmers and electricians go to county-level rural power farms, the recruiter plan is to simulate and maintain the number of recruiters of each city-level company every year according to adjustment parameters, the annual missing member index, the age index and the talent index are simulated and extracted, the number of recruiters of each city-level single farmer and electrician is calculated by combining parameter simulation of the recruiter coefficient and the configuration index, the distribution plan is to simulate and calculate the number of missing members and the missing member proportion of each city-level unit every year, and the number of distributed persons of each rural power farm is calculated by combining preset distribution coefficients; the agricultural electrician distribution simulation operation simulates and extracts the agricultural electrician by using an intelligent optimization algorithm selected by the seat position distribution characteristics, simulates and distributes the agricultural electrician to agricultural electric power farms subordinate to various city companies, and completes the annual agricultural electrician distribution simulation operation in the next five years.

Further, the labor dispatching personnel allocation simulation operation of the employee entrance specifically comprises: the labor service personnel are all members to be supplemented to county-level rural power stations, the member supplementing plan is to maintain the fixed number of the members to be supplemented in each urban unit every year, the distribution plan is to simulate and calculate the labor service missing member number and the missing member occupation ratio of the rural power stations in each urban unit every year, and the distribution number of each rural power station is calculated by combining a preset distribution coefficient; and the labor dispatching personnel allocation simulation operation simulates and extracts labor dispatching personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulates and allocates the labor dispatching personnel to agricultural and electric facilities subordinate to each city company, and completes the labor dispatching personnel allocation simulation operation every year in the next five years.

Further, the student allocation simulation operation of the employee portal specifically includes: the distribution plan is to maintain the number of distributed people of each professional of each unit, the demand plan is to cross multiply according to the detailed item ratio of each item according to the sex ratio, the academic calendar ratio and the school level ratio demand reference item of each professional to form the comprehensive ratio of the cross result of each post and each detailed item, and the simulated distribution operation of the province and the directly affiliated units simulates and extracts new students by using an intelligent optimization algorithm selected by the seat distribution characteristics to complete the simulated distribution of personnel; the simulation operation of the city level unit is to carry out the simulation distribution of the local students and the researchers according to the unit supplementing strategy, the professional supplementing strategy, the distribution plan and the demand plan of each unit, wherein the unit supplementing strategy is to extract the index of the absent number, the absent rate and the aging index of each simulation year according to the condition of each unit, combine the index with the labor guide line proportion and the parameters of the supplementing proportion to check out the supplementing proportion and the number of the supplementing members of each unit, the professional supplementing strategy is to further detail the supplementing strategy to each specialty of each unit on the basis of the unit supplementing strategy, check out the checking and determining the number of the supplementing members of each unit, the distribution plan is to simulate and calculate the absent number and the absent ratio of each city level unit to each professional position under the team, the distribution number of each professional position is calculated according to the gender ratio, the demand plan is to calculate the number of each professional, And the simulation distribution operation of the city level unit simulates and extracts newly-increased researchers and the undergraduates by using an intelligent optimization algorithm selected by using seat distribution characteristics, and completes the simulation distribution of personnel.

Furthermore, the difference search algorithm-based random extraction method for people meeting probability distribution conditions adopts feature selection of a difference search algorithm to perform dimensionality reduction work, dimensionality in dimensionality reduction is the number of probability combinations meeting extraction conditions in random extraction of people, and the extraction conditions comprise gender, age bracket, academic calendar, school hierarchy, academic specialty and the like; the concrete steps include:

1) initialization

The differential search algorithm randomly initializes [ N ] within the search space using the following formula_P,D]Artificial superindividual X of dimension_i,j：

X_i,j＝low_j+rand*(up_j-low_j)

i＝1,2,...,N_p

j＝1,2,...,D

Wherein N is_pRepresenting the number of elements in a superindividual (population size); d represents the dimension of the problem; up and low define the upper and lower bounds of the learned space, respectively;

2) migration operations

After initialization, the dwell vector S in the search area is searched_i,GRandom transformation is used for random generation, which is the key for successfully realizing the migration process in the DSA;

S_i,G＝X_i,G+scale*(donor-X_i,G)

scale＝randg(2*rand₁)*(rand₂-rand₃)

and donor ═ X_i,j|_{random_shuffing}

Wherein scale controls the magnitude of the change in position of the artificial organic individual, randg being a random value selected from the gamma distribution; rand₁,rand₂,rand₃Is at [0,1 ]]The random number selected in (1);

3) search operations

The search process for the dwell vector may be calculated by individual organisms in the super organism using the following process:

wherein, S'_i,j,GAn experimental vector representing the jth particle in the ith dimension of the G generation; r is_i,jIs an integer of 0 or 1;

4) selection operation

The selection operation is used to define the next generation, i.e. G + 1; based on the fitness value, intervening between the dwell vector population and the artificial organism population; the selection operation is described in detail as follows:

5) feature selection

Variable of 0-1

Logistic function

Feature selection

The algorithm can be used for constructing a multi-target optimization problem by taking the precision or the number of errors and dimensionality reduction as 2 optimization targets; whether each feature uses a random variable defined as a 0-1 distribution; mapping the independent variable in the algorithm updating process in a range of 0-1 by using a logistic function, so that the value of the characteristic with the function value larger than 0.5 is 1, and the characteristic needs to be reserved; the characteristic with the function value less than or equal to 0.5 is set to be 0, which represents that the characteristic is not reserved; in the subsequent research, a feature extraction technology is introduced to find the functional relationship among the features, so that some features are replaced by the functional relationship of other related features, and the purpose of reducing the dimension is achieved.

Compared with the prior art, the invention has the beneficial effects that:

1) the invention takes 'automatic simulation hierarchical path' and 'random extraction of personnel meeting probability distribution conditions' as the key technology of simulation distribution; through data analysis and technical construction, a reasonable scheme is made, employee outlets, employee inlets and employee internal flow are simulated, and accurate division and post of human resource balance analysis are realized;

2) in the deduction simulation method, the system establishes an automatic simulation hierarchical path tree according to the principle of first-in-last-in and top-down, orderly processes the entering and exiting and flowing of the employees of each level unit, which are dispersedly generated all year round in the actual work, from the dispersing generation to the one-time processing, and establishes reasonable software operation logic to ensure that the internal flowing, the employee exiting and the replenishment actions are orderly and reasonable;

3) the random personnel extraction method meeting the probability distribution condition adopts the characteristic selection of a differential search algorithm to perform the dimension reduction work, and realizes the simulation process of automatic personnel extraction and flow when personnel extraction and distribution are performed.

Drawings

FIG. 1 is an automatic simulation hierarchical path diagram of the present invention;

fig. 2 is a functional block diagram of a system implementation applying an embodiment of the method of the present invention.

Detailed Description

The following detailed description of the present invention will be made with reference to the accompanying drawings.

The following examples are given by Power saving of Liaoning, a national grid.

Method for automatically simulating hierarchical path

As shown in fig. 1, the method for automatically simulating hierarchical paths includes: simulating the ways of increasing, decreasing and flowing of employees from low to high according to the sequence of first-in and last-in from top to bottom; the first-in and last-out are operations of staff export when human resource planning simulation operation is carried out, namely retirement operation and natural staff reduction operation are carried out firstly, and then staff entry work of transfer officer distribution, special student distribution, agricultural and electrician distribution, labor dispatching distribution and local student distribution is carried out; the simulation is carried out according to the sequence from top to bottom in the figure 1, namely, the simulation operation is carried out according to the sequence of retirement, natural staff reduction, two-line withdrawal, cadre subsidiaries, subsidiaries of collective enterprises, diversion officers distribution, special students distribution, agro-electrician distribution, labor dispatching distribution, local students distribution and the flow of the second part of staff; the low-to-high state refers to a flowing process from county level to provincial level company when the second part of staff flow simulation operation; wherein, two lines are returned, the subsidiaries of cadres and the subsidiaries of the collective enterprise flow for the first part of the staffs.

Through the employee export simulation operation, the employee entry simulation operation and the employee flow simulation operation, an employee condition database at the end of each year of the next five years is formed, various statistical query and analysis works can be carried out on the basis of the database, and data support is provided for manpower resource planning.

The simulation process for simulating the paths of increasing, decreasing and flowing of the employees from top to bottom in a sequence of first-in and last-in from low to high specifically comprises the following steps:

and step 3: after the simulation of the employee export is completed, the flow simulation of the employee is performed for the first time, firstly, the second line quitting is simulated, the simulated second line quitting personnel in each year in the next five years is automatically calculated according to the sex, the age and the post system where the employee is located, the second line quitting time is marked, and the simulation of the second line quitting operation is completed;

the specialist assignment simulation operation of the employee entrance specifically comprises the following steps: the method comprises the following steps that a specialist is a farm power station from a researcher to a county level, a researcher plan is to maintain fixed number of the researcher in each city level unit every year, a distribution plan is to simulate and calculate the number of missing members and the proportion of the missing members of the farm power station in each city level unit every year, and the distribution number of each farm power station is calculated by combining a preset distribution coefficient; the specialist allocation simulation operation simulates and extracts specialist personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulates and allocates the specialist personnel to agricultural and electric farms of each city company, and completes the annual specialist allocation simulation operation in the next five years.

the rural power industry distribution simulation operation of the employee entrance specifically comprises the following steps: the method comprises the steps that farmers and electricians go to county-level rural power farms, the replenishment plan is to simulate and maintain the replenishment number of each city-level company every year according to adjustment parameters, simulate and extract annual shortage index, age index and talent index, combine the replenishment coefficient and the parameter simulation of configuration index to calculate the replenishment number of each city-level unit farmer and electrician every year, the distribution plan is to simulate and calculate the shortage number and shortage ratio of each city-level unit rural power farm every year, and combine the preset distribution coefficient to calculate the distribution number of each rural power farm; the agricultural electrician distribution simulation operation simulates and extracts the agricultural electrician by using an intelligent optimization algorithm selected by the seat distribution characteristics, simulates and distributes the agricultural electrician to agricultural electric power farms subordinate to various municipal companies, and completes the annual agricultural electrician distribution simulation operation in the next five years.

the labor dispatching personnel allocation simulation operation of the employee entrance specifically comprises the following steps: the labor service personnel are all subsidiaries to county-level rural power stations, the subsidiary plan is to maintain the number of the fixed subsidiaries in each urban unit every year, the distribution plan is to simulate and calculate the labor service absent personnel number and the absent personnel occupation ratio of the rural power stations in each urban unit every year, and the distribution number of each rural power station is calculated by combining a preset distribution coefficient; and (4) labor dispatching personnel allocation simulation operation, namely simulating and extracting labor dispatching personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulating and allocating the labor dispatching personnel to agricultural and electric facilities subordinate to each city company, and completing the annual labor dispatching personnel allocation simulation operation in the next five years.

the student research student allocation simulation operation of the employee entrance specifically comprises the following steps: the distribution plan is to maintain the number of distributed people of each professional of each unit, the demand plan is to cross multiply according to the detailed item ratio of each item according to the sex ratio, the academic calendar ratio and the school level ratio demand reference items of each professional to form the comprehensive ratio of the cross result of each post and each detailed item, and the simulation distribution operation of the province and the directly affiliated units simulates and extracts new students by using an intelligent optimization algorithm selected by the seat distribution characteristics to complete the simulation distribution of personnel; the simulation operation of the city level units is to carry out the simulation distribution of the local students and the researchers according to the unit supplementing strategy, the professional supplementing strategy, the distribution plan and the demand plan of each unit, wherein the unit supplementing strategy is to extract the number of absent persons, the absent person rate and the index of aging index of each simulation year according to the condition of each unit, combine the worker guide line occupation ratio and the parameters of the supplementing proportion to define the supplementing ratio and the number of the supplementing persons of each unit, the professional supplementing strategy is to further detail the supplementing strategy to each specialty of each unit on the basis of the unit supplementing strategy, define the number of the supplementing persons of each unit, the distribution plan is to simulate and calculate the number of absent persons and the absent person occupation ratio which are detailed to each professional position under the team each city level unit every year, combine the distribution coefficient to calculate the distribution number of each professional position, and the demand plan is to calculate the distribution number of each professional position according to the gender occupation ratio, the number of each professional, And the simulation distribution operation of the city level unit simulates and extracts new students and the undergraduates by using an intelligent optimization algorithm selected by seat distribution characteristics, and completes the simulation distribution of personnel.

And 7: after the simulation of the employee exit and entrance is completed, the simulation operation of the second part of employee flow is carried out, the simulation operation is based on the principle of from low to high, namely in four levels of county bureaus, basic level units, implementation mechanisms, directly subordinate units and provincial companies, the flow direction is from the skill posts to the management posts in the level, the skill posts flow to the upper level units, and the management posts flow to the upper level units; the flow mode is that according to the staff flow distribution plan and the staff flow demand plan, through the staff flow distribution processing function, the staff flowing through various paths are simulated and extracted by using the intelligent optimization algorithm selected by the seat distribution characteristics, and the staff flow simulation operation in the next five years is completed. The specific flow process is shown in the flow sequence 1-9 in FIG. 1.

Second, random personnel extraction method meeting probability distribution condition

Although the human resource planning of the existing power grid enterprises can think and conspire from the strategy height to the human resource team development and the human resource management work overall situation, the human resource planning prediction cannot be closely combined with the production operation development to form a specific reasonable human resource allocation; the recruitment of new employees is configured according to the conditions of over-determined employees, but cannot be effectively matched with the quality requirements needed by over-lacking posts. Therefore, a scientific, reasonable and comprehensive human resource demand prediction method is urgently needed to solve the problem of human resource allocation. When people automatically draw flow, the research on the random drawing problem of people meeting the probability distribution condition is the most common technical problem in the simulation process of realizing the automatic drawing flow of people. However, as the dimension increases, when the probability combinations are too many thousands (e.g., gender (2) X age group (8) X school calendar (5) X school level (7) X school professional (4) ═ 2240 combinations), the minimum probability distribution and the maximum probability distribution are different by ten thousand times, the above method is not applicable. Therefore, an optimization algorithm needs to be researched to solve the above problems.

Differential Search Algorithm (DSA) is a new population-based heuristic Evolutionary Algorithm (EA) developed by Civicioglu. Its inspiration comes from the migration process of the organisms that make up the super-organism (a metaphor for social insect populations such as bees, ants and termites) during a year's climate change. The migration process allows species to migrate from a habitat to a habitat that has greater natural resource capacity and diversity and is more efficient. In DSA, the search space is modeled as a food area, and each location in the search space represents an artificial super-organism.

The dimensionality in the dimensionality reduction is the probability combination number meeting the extraction conditions in the random extraction of the personnel, and the extraction conditions comprise sex, age group, academic calendar, school hierarchy, learned specialty and the like; such as: gender (2), age group X (8), school calendar X (5), school level X (7), school specialty X (4) ═ 2240 combinations. (see item 17 of the system function of the specific embodiment: staff flow assignment process for details).

The method specifically comprises the following steps:

1. initialization

DSA is randomly initialized in search space using the following formula [ N ]_P,D]Artificial superindividual X of dimension_i,j：

X_i,j＝low_j+rand*(up_j-low_j)

i＝1,2,...,N_p

j＝1,2,...,D

Wherein N is_pRepresenting the number of elements in the superindividual (population size). D represents the dimension of the problem. up and low define the upper and lower bounds of the learned space, respectively.

2. Migration operations

After initialization, the dwell vector S in the search area is searched_i,GRandom transformation is used for random generation, which is the key for successful realization of the migration process in DSA.

S_i,G＝X_i,G+scale*(donor-X_i,G)

scale＝randg(2*rand₁)*(rand₂-rand₃)

And donor ═ X_i,j|_{random_shuffing}

Where scale controls the magnitude of the artificial organic individual's position change and randg is a random value selected from the gamma distribution. rand₁,rand₂,rand₃Is at [0,1 ]]The random number of (1).

3. Search operations

wherein, S'_i,j,GThe experimental vector of the j-th particle in the i-th dimension of the G-th generation is shown. r is_i,jIs an integer of 0 or 1.

4. Selection operation

The selection operation is used to define the next generation, i.e., G + 1. Based on the fitness value, between the dwell vector population and the artificial organism population. The selection operation is described in detail as follows:

5. feature selection

Variable of 0-1

Logistic function

Feature selection

The algorithm can be used for constructing a multi-objective optimization problem by taking the precision (or error) and the number of dimension reduction as 2 optimization objectives; whether each feature uses a random variable defined as a 0-1 distribution; mapping the independent variable in the algorithm updating process in a range of 0-1 by using a logistic function, so that the value of the characteristic with the function value larger than 0.5 is 1, and the characteristic needs to be reserved; and a feature with a function value of 0.5 or less is set to 0, indicating that the feature is not retained. In subsequent research, a feature extraction technology is introduced to find a functional relationship among features, so that some features are replaced by functional relationships of other related features, and the purpose of reducing dimensions is achieved.

Third, System embodiment applying the method of the present invention

The human resource supply and demand simulation system applying the human resource supply and demand simulation method based on the hierarchical path and the differential search algorithm comprises the following functional modules: the method comprises the steps of basic setting, basic information, employee export, employee entry, employee internal flow and statistical analysis, simulation and deduction of retirement, natural deceased and recruiting personnel of different levels, different business types and different employment forms, and internal post mobilization and promotion of enterprises, wherein simulation data support multi-level and multi-dimensional statistical analysis, support export of analysis results and generate analysis reports of various formats of word, excel and pdf.

The cardinality setting is used for providing management and configuration of an index system and a knowledge system for data management and analysis, simulation and prediction; the method comprises the following steps: information classification dictionary, employment guide line, business outsourcing coefficient, employment configuration rate, graduate demand description, personnel reduction description, personnel flow demand description, flow basic condition, flow probability coefficient management and simulation path management;

the basic information is used for providing organization architecture and internal and external personnel information management; the method comprises the following steps: organization information management, post information management, personnel information management, transit officer repairment information, rural power and electric power repairment information, graduate repairment information and labor service dispatch repairment information;

the employee exports are used for providing planning-period retirement and other labor contract-releasing personnel staff reduction process simulation;

the employee entrance is used for providing a researcher process simulation for planning period transfer officers, graduates, agro-electricians and labor dispatching;

the staff internal flow is used for providing internal flow processing functions of leading personnel to quit two lines, leading personnel to carry out additional editing, saving management industry unit supplementaries and between each level of units, and simulating the staff internal flow process;

the statistical analysis is used for carrying out balance analysis and personnel structure analysis of the total amount of labor, the allocation rate of labor and demand of any level unit, any professional and any planning year in a graphic and text form according to different analysis emphasis points, and comprises the following steps: the change trend of the total amount of the used labor, the change of the requirement of the used labor, the change of the allocation rate of the used labor, the composition of internal personnel, the condition of each professional used labor, the condition of each unit used labor and the condition of personnel at the end of the planning period.

Fig. 2 is a block diagram of the functional structure of the system of the method of the present invention.

The contents included in the above functional modules are specifically explained as follows:

1. finger wire for worker

The employment guideline function is a criterion of the number of professional workers of each unit imported from the outside as basic data of the demand plan.

2. Business outsourcing factor

The outsourcing coefficient function is to maintain the coefficient of the outsourcing of each outsourcing specialty and prepare basic data for outsourcing distribution and demand planning. The adjustment factor for each production outsourcing specialty of each unit is recorded in units.

3. Graduate need description

The graduate demand description is based on unit and professional classification, maintains the description of the graduate demand of each professional, and provides basic data for the requirement of an increase and the distribution plan, wherein the description comprises the proportions of gender, each level of academic history, academic history type, academic specialty and the like.

4. Rate of employment allocation

The employment allocation rate is the proportion of the employment allocation of each maintenance unit, and the allocation rate is between 0 and 100 as basic data.

5. Description of flow requirements

The staff flow requirement description is the situation that maintenance staff flow among units, and the description mode is described according to items such as gender, age, school calendar, school hierarchy and the like. Each detailed item in each project has a certain proportion, and subsequent various demands and distribution plans are influenced by changing the proportion.

6. Probability of each unit flowing to provincial company

The liaison probability of each unit to the provincial company is the distribution probability of the flow of the maintenance-oriented unit, the basic electric bureau, the county bureau and other basic units to the provincial company.

7. Statistics of the number of persons in excess or in shortage in each unit

And (4) counting the conditions of the persons in the absence or the excess of the persons in each unit and department according to the information of the basic information base and the employment guide line, wherein the counting content comprises information such as the number of the persons in the absence or the excess of the persons in each age group and the like.

8. Existing personnel information base

The existing personnel information base has the function of storing various basic information of personnel, is a personnel information set planned according to historical information according to actual conditions, comprises a data set of various personnel information items such as names, sexes, academic calendars, technical titles, skills, ages and the like, and is a basic information base for realizing personnel requirements and distribution plans.

9. Graduate information base

The graduate information base is a database for storing graduate information of each year, the items of the database are consistent with the personnel information base, the graduate information base also comprises the annual information for distinguishing graduates of different years, and the graduate information base is used for providing basic information of personnel for the additional member distribution plan.

10. Labor dispatch information base

The labor dispatch information base is a database for storing basic information of labor dispatch personnel, and the contained information is similar to the personnel information base, because the labor dispatch personnel are not managed in detail like the staff members, so that the items and the information in the database are less than those in the existing personnel information base.

11. Employee retirement and other staff reduction

The retirement of staff and other staff reducers is a database for storing information of retirement and other staff reducers in each year.

12. Staff increase allocation plan

The staff member-adding distribution plan is divided into distribution plans for a transfer officer, a specialist, a researcher, a local student, an agricultural and electrical worker and a labor service dispatching person, and different distribution plans are provided for different persons. The distribution plan is a plan for clarifying each post to each unit section, and the plan includes plan items such as the number of persons who are scheduled, the actual number of persons, the number of persons who are absent, the proportion of persons who are absent, and the distribution coefficient of each post.

13. Staff increase demand plan

The staff member-increasing demand plan is divided into demand plans for a professional officer, a specialist, a research student, a local student, an agricultural and electrical worker and a labor service dispatching person, and different demand plans are provided for different persons. The demand plan is a plan for making clear each post in each unit department, and the plan comprises items such as gender ratio, academic calendar ratio, school level ratio and the like of each post, and the items are cross-multiplied according to detailed items of each item to form a comprehensive ratio of each post and a cross result of each detailed item.

14. Staff add-to-staff allocation process

The staff member-increasing allocation processing is to relatively and fairly allocate the staff members of the trans-employment officer, the specialist, the researcher, the local student, the agricultural and electrical workers and the labor dispatching personnel to the positions of the members lacking in the units by utilizing a position allocation algorithm and combining the staff member-increasing demand plan and the allocation plan.

15. Staff flow distribution plan

The staff member-added distribution plan is divided into distribution plans for staff members of off duty companies, collective enterprises, provincial company headquarters, affiliated units, city units and county companies, and different distribution plans are provided for different staff members. The distribution plan is a plan for clarifying each post of each unit department, and the plan includes plan items such as the number of persons who are fixed, the actual number of persons, the employment allocation rate, the number of persons who are absent, and the proportion of persons who are absent.

16. Staff flow requirement planning

The staff flow demand plan is divided into demand plans for staff on duty, collective enterprises, province company headquarters, direct units, city units and county public departments, and different demand plans are provided for different staff. The demand plan is a plan for making clear each post in each unit department, and the plan comprises items such as gender ratio, academic calendar ratio, school level ratio and the like of each post, and forms a comprehensive ratio of intersection results of each post and each detail item by cross multiplication according to the detail items of each item.

17. Staff flow allocation process

The staff flow distribution processing is an intelligent optimization algorithm selected by using a core algorithm, namely, a seat distribution characteristic, and the staff random extraction problem meeting the probability distribution condition realizes the simulation process of 'staff automatic flow extraction', and the following table illustrates the problem of randomly extracting 2 men according to probability distribution when the staff flow:

described in the table are simple examples of randomly extracting 2 males in two dimensions of gender and age according to a given probability distribution, and the distribution coefficient (also expected number of people in each age) is calculated by probability, wherein the number of people in '35 years and under' is 0.116, and if data processing is performed according to any alternative method, a distorted result is obtained when random extraction is performed. For example, 2 people are respectively extracted from the age groups of 56-60 and 51-55 according to the distribution coefficient sequence from large to small, then the distribution probability of other age groups is 0, the randomly extracted people in each group are expanded in the same proportion, 17 people meeting the probability distribution are extracted from the resource pool prepared for the second extraction, and 2 people are randomly extracted from the 17 people.

With the increase of dimensionality, when various probability combinations are over thousand (for example, the combination of gender (2), age (8), school calendar (5), school level (7), school professional (4) ═ 2240), the minimum probability distribution and the maximum probability distribution are different by ten thousand times, the above method is not available. The intelligent optimization algorithm selected by the bit allocation characteristics can solve the problems and realize the allocation processing of staff flow.

18. Statistics of total amount of work variation trend

The conditions of the number of the added and removed workers and the number of the supplemented workers of all the workers in the simulated year are counted to form the tables and graphs of the number of the added and removed workers of all the workers in each year, the conditions of different simulated years can be counted, and the conditions of mechanisms of all levels can be counted.

19. The work condition of each unit and each professional

And (4) counting the labor guideline and the labor personnel of each specialty of the simulation year to form the labor allocation rate, counting the conditions of different simulation years, and counting the conditions of mechanisms of each level.

20. Planning end-of-term personnel condition analysis

The statistics of the end-of-term personnel of the simulated year can be carried out according to the items such as age structure, gender, academic calendar, change situation and the like, each item can be further divided into different specialties for statistics, and the statistics content comprises the information of the increase and decrease of personnel and the change situation.

The above embodiments are implemented on the premise of the technical solution of the present invention, and detailed embodiments and specific operation procedures are given, but the scope of the present invention is not limited to the above embodiments. The methods used in the above examples are conventional methods unless otherwise specified.

Claims

1. A human resource supply and demand simulation method based on hierarchical paths and differential search algorithms is characterized in that the method simulates and deduces retirement, natural staff reduction, recruitment personnel and enterprise internal post mobilization and promotion of different hierarchical units, different service types and different employment forms, and simulated data support multi-level and multi-dimensional statistical analysis;

the method comprises an automatic simulation hierarchical path method and a personnel random extraction method which is based on a differential search algorithm and meets probability distribution conditions;

the method for automatically simulating the hierarchical path comprises the following steps: simulating the ways of increasing, decreasing and flowing of employees from top to bottom in a sequence of first-in and last-in from low to high; the first-in and last-out are the operations of a staff exit, namely, the retirement operation and the natural staff reduction operation are firstly carried out when the human resource planning simulation operation is carried out, and then staff entrance operations of the transformer officer distribution, the specialist student distribution, the agriculture and electrician distribution, the labor dispatching distribution and the student research distribution are carried out; from top to bottom, simulation operation is carried out according to the sequence of retirement, natural staff reduction, second line withdrawal, cadre subsidiaries, subsidiaries of collective enterprises, division of trans-officers, division of specialist students, division of agro-electricians, division of labor assignment, division of local students and staff flow of the second part; the low-to-high state refers to a flowing process from county level to provincial level company when the second part of staff flow simulation operation; wherein, two lines are returned, the subsidiaries of cadres and the subsidiaries of the collective enterprise flow to the first part of the employees;

through the employee export simulation operation, the employee entry simulation operation and the employee flow simulation operation, an employee condition database at the end of the annual period of the next five years is formed, various statistical query and analysis works can be carried out on the basis of the database, and data support is provided for manpower resource planning;

the difference search algorithm-based random extraction method for the personnel meeting the probability distribution condition adopts the feature selection of the difference search algorithm to perform the dimension reduction work, and realizes the simulation process of automatic extraction and flow of the personnel when the personnel are randomly extracted and distributed.

2. The human resource supply and demand simulation method based on the hierarchical path and the differential search algorithm as claimed in claim 1, wherein the simulation process of simulating the increase and decrease and flow paths of the employees from top to bottom in a sequence of first-in and last-in and from bottom to top comprises the following specific steps:

step 1: the simulation process starts from employee export, firstly, retirement is simulated, according to the specialty of gender and the date of birth, the system automatically carries out retirement reduction treatment, simulated retired persons in five years in the future are formed, retirement time is marked, and simulated retirement operation is completed;

step 2: then simulating natural staff reducers, summarizing the conditions of the sex natural staff reducers of all age groups according to historical natural staff reducer conditions, calculating the natural staff reducer probability and the most possible natural staff reducers of each unit according to historical conditions, and simulating the number of the natural staff reducers and the number of the staff reducers;

and step 3: after the simulation of the employee export is completed, performing the flow simulation of the first part of employees, firstly simulating the second line quitting, automatically calculating the simulated second line quitting staff every year in the next five years according to the sex, the age and the post system where the staff is located, marking the second line quitting time, and completing the simulation of the second line quitting operation;

and 4, step 4: corresponding cadre posts can be vacated after the simulation returns the two lines, simulation operation of cadre subsidiaries is carried out on the cadre posts, the cadre subsidiaries are selected from internal personnel or obtained by flowing from a lower-level unit, personnel meeting the conditions are extracted according to various requirements of cadre subsidiary plan, the annual number of the subsidiaries of each unit is small according to the previous situation, the personnel meeting the conditions can be randomly extracted to determine the personnel to be subsidiaries by the staff during the simulation of the subsidiaries, and a final cadre subsidiary simulation result is formed;

601) the transfer officers at the employee entrances distribute simulation operations, and the number of units and persons receiving the transfer officers every year is fixed according to historical conditions, so that the simulation operations are realized by adding corresponding number of persons to units meeting the conditions according to the plan of the subsidiaries, and the simulation operations of the transfer officers in the next five years are completed;

and 7: after the simulation of the employee exit and entrance is completed, the simulation operation of the second part of employee flow is carried out, the simulation operation is based on the principle of from low to high, namely in four levels of county bureaus, basic level units, implementation institutions, directly subordinate units and province companies, the flow direction is from the skill posts to the management posts in the level, the skill posts flow to the upper level units, and the management posts flow to the upper level units; the flow mode is that according to the staff flow distribution plan and the staff flow demand plan, through the staff flow distribution processing function, the staff flowing through various paths are simulated and extracted by using the intelligent optimization algorithm selected by the seat distribution characteristics, and the staff flow simulation operation in the next five years is completed.

3. The human resource supply and demand simulation method based on the hierarchical path and the differential search algorithm as claimed in claim 2, wherein the specialist assignment simulation operation of the employee portal specifically comprises: the method comprises the following steps that a specialist is a farm electric station from a researcher to a county level, a researcher plan is to maintain fixed number of the researcher in each city level unit every year, a distribution plan is to simulate and calculate the number of missing members and the proportion of the missing members in the farm electric station in each city level unit every year, and the distribution number of each farm electric station is calculated by combining a preset distribution coefficient; the specialist allocation simulation operation simulates and extracts specialist personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulates and allocates the specialist personnel to agricultural and electric farms subordinate to each city company, and completes the annual specialist allocation simulation operation in the next five years.

4. The human resource supply and demand simulation method based on the hierarchical path and the differential search algorithm as claimed in claim 2, wherein the simulation operation of the rural power and electricity distribution of the employee entrance specifically comprises: the method comprises the steps that farmers and electricians go to county-level rural power farms, the recruiter plan is to simulate and maintain the number of recruiters of each city-level company every year according to adjustment parameters, simulate and extract annual missing member indexes, age indexes and talent indexes, combine the parameter simulation of the recruiter coefficient and the configuration index to calculate the number of recruiters of the farmers and electricians of each city-level unit every year, the distribution plan is to simulate and calculate the number of missing members and the missing member proportion of the rural power farms of each city-level unit every year, and combine the preset distribution coefficient to calculate the number of distributed persons of each rural power farm; the agricultural electrician distribution simulation operation simulates and extracts the agricultural electrician by using an intelligent optimization algorithm selected by the seat distribution characteristics, simulates and distributes the agricultural electrician to agricultural electric power farms subordinate to various municipal companies, and completes the annual agricultural electrician distribution simulation operation in the next five years.

5. The human resource supply and demand simulation method based on hierarchical paths and differential search algorithm as claimed in claim 2, wherein the labor dispatching personnel allocation simulation operation of the employee entrance specifically comprises: the labor service personnel are all subsidiaries to county-level rural power stations, the subsidiary plan is to maintain the number of the fixed subsidiaries in each urban unit every year, the distribution plan is to simulate and calculate the labor service absent personnel number and the absent personnel occupation ratio of the rural power stations in each urban unit every year, and the distribution number of each rural power station is calculated by combining a preset distribution coefficient; and (4) labor dispatching personnel allocation simulation operation, namely simulating and extracting labor dispatching personnel by using an intelligent optimization algorithm selected by seat allocation characteristics, simulating and allocating the labor dispatching personnel to agricultural and electric facilities subordinate to each city company, and completing the annual labor dispatching personnel allocation simulation operation in the next five years.

6. The human resource supply and demand simulation method based on hierarchical paths and differential search algorithm as claimed in claim 2, wherein the assignment simulation operation of the student research on employee entry specifically comprises: the distribution plan is to maintain the number of distributed people of each professional of each unit, the demand plan is to cross multiply according to the detailed item ratio of each item according to the sex ratio, the academic calendar ratio and the school level ratio demand reference item of each professional to form the comprehensive ratio of the cross result of each post and each detailed item, and the simulation distribution operation of the province and the directly affiliated units simulates and extracts new students by using an intelligent optimization algorithm selected by the seat distribution characteristics to complete the simulation distribution of personnel; the simulation operation of the city level units is to carry out the simulation distribution of the local students and the researchers according to the unit supplementing strategy, the professional supplementing strategy, the distribution plan and the demand plan of each unit, wherein the unit supplementing strategy is to extract the number of absent persons, the absent person rate and the aging index of each simulation year according to the condition of each unit, combine the index of the employment index line and the supplement proportion parameters to check out the proportion of each unit supplementing persons and the number of the supplementing persons, the professional supplementing strategy is to further detail the supplementing strategy to each specialty of each unit on the basis of the unit supplementing strategy, check out the number of the supplementing persons of each unit specialty, the distribution plan is to simulate and calculate the number of absent persons and the absent person proportion of each city level unit to each professional position under the team, combine the distribution coefficient to calculate the number of the distribution of each professional position, the demand plan is to calculate the distribution number of each professional position according to the gender proportion, And the simulation distribution operation of the city level unit simulates and extracts new students and the undergraduates by using an intelligent optimization algorithm selected by seat distribution characteristics, and completes the simulation distribution of personnel.

7. The human resource supply and demand simulation method based on hierarchical paths and differential search algorithm as claimed in claim 1, wherein the human random extraction method based on differential search algorithm and satisfying probability distribution condition adopts feature selection of differential search algorithm to perform dimensionality reduction work, wherein dimensionality in dimensionality reduction is the number of probability combinations satisfying extraction conditions in human random extraction, and the extraction conditions include gender, age group, academic calendar, school hierarchy, academic specialty, etc.;

the method specifically comprises the following steps:

1) initialization

X_i,j＝low_j+rand*(up_j-low_j)

i＝1,2,...,N_p

j＝1,2,...,D

Wherein N is_pRepresenting the number of elements in a super-volume; d represents the dimension of the problem; up and low define the upper and lower bounds of the learned space, respectively;

2) migration operations

S_i,G＝X_i,G+scale*(donor-X_i,G)

scale＝randg(2*rand₁)*(rand₂-rand₃)

and donor ═ X_i,j|_{random_shuffing}

3) search operations

4) selection operation

5) feature selection

Variable of 0-1

Logistic function

Feature selection

The algorithm can be used for constructing a multi-objective optimization problem by taking the precision or the number of errors and dimensionality reduction as 2 optimization objectives; whether each feature uses a random variable defined as a 0-1 distribution; mapping the independent variable in the algorithm updating process in a range of 0-1 by using a logistic function, so that the value of the characteristic with the function value larger than 0.5 is 1, and the characteristic needs to be reserved; the characteristic with the function value less than or equal to 0.5 is set to be 0, which represents that the characteristic is not reserved; in subsequent research, a feature extraction technology is introduced to find a functional relationship among features, so that some features are replaced by functional relationships of other related features, and the purpose of reducing dimensions is achieved.