Search

CN-122021999-A - Short-term labor market supply and demand matching optimization method based on matching theory

CN122021999ACN 122021999 ACN122021999 ACN 122021999ACN-122021999-A

Abstract

The invention discloses a short-term labor market supply and demand matching optimization method based on a matching theory, which comprises the steps of S1, generating a post attribute standardization result set and a personnel attribute standardization result set, S2, generating an initial preference priority matrix, S3, executing an improved Gale-shape matching algorithm on the basis of the initial preference priority matrix to obtain an initial matching result, S4, calculating a matching satisfaction evaluation index result set for the initial matching result according to the post preference vector set and the personnel preference vector set, S5, generating an initial matching solution set by taking the initial matching result as a seed solution, obtaining an optimized matching result, and obtaining a final optimized matching result, and S6, transmitting the final optimized matching result or the dynamic matching result to a short-term labor dispatch management system. The invention obviously improves the adaptation precision and the generation efficiency of the new scheme.

Inventors

  • SHENG XIAOYUAN
  • LUO ZHI

Assignees

  • 武汉大学

Dates

Publication Date
20260512
Application Date
20251223

Claims (11)

  1. 1. A short-term labor market supply and demand matching optimization method based on a matching theory is characterized by comprising the following steps: S1, acquiring a short-term labor post information data set and a labor personnel information data set, and performing attribute standardization processing to generate a post attribute standardization result set and a personnel attribute standardization result set; s2, building a post preference vector set based on the post attribute standardization result set, building a personnel preference vector set based on the personnel attribute standardization result set, and generating an initial preference priority matrix; S3, executing an improved Gale-shape matching algorithm on the basis of the initial preference priority matrix to obtain an initial matching result; s4, calculating a matching satisfaction evaluation index result set according to the post preference vector set and the personnel preference vector set for the initial matching result; s5, generating an initial matching solution set by taking the initial matching result as a seed solution, obtaining an optimized matching result, judging whether the optimized matching result reaches a convergence criterion, if yes, taking the optimized matching result as a current effective matching scheme, otherwise, feeding back the optimized matching result to a cuckoo search algorithm, and continuing to execute global search iteration until the convergence criterion is met, so as to obtain a final optimized matching result.
  2. 2. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 1, further comprising: s6, monitoring a post preference dynamic change event in the labor market, if the post or personnel preference is detected to be changed, outputting a dynamic matching result based on the current effective matching scheme, and transmitting the final optimized matching result or the dynamic matching result to a short-term labor dispatch management system.
  3. 3. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 1, wherein the step S1 comprises the following steps: S11, constructing a short-term labor post information data set And staff information data set ; S12, carrying out standardized processing on the short-term labor post information data set and the labor personnel information data set; s13, after the standardization process is completed, each post Form post attribute standardization results for each staff member Forming a personnel attribute standardization result by the attribute vector of (1); s14, combining all post attribute standardization results into a post attribute standardization result set The attribute standardization results of all people are combined into a personnel attribute standardization result set Wherein And The representations respectively correspond to posts And personnel Is used to normalize the attribute vector.
  4. 4. A short-term labor market supply and demand matching optimization method based on a matching theory according to claim 3, wherein each post in the short-term labor post information data set The staff information data set comprises three attributes, namely, a staff type code of the post, a working time length of the post and a salary level of the post, wherein each staff in the staff information data set comprises three attributes, namely, a skill level code of the staff, an acceptable staff type code and a desired salary level.
  5. 5. A short-term labor market supply and demand matching optimization method based on a matching theory according to claim 3, wherein the step S2 comprises the following steps: s21, standardizing result set based on post attribute For each post Building post preference vectors Post preference vector Representation post Preference score set for all staff, each post matching preference score Representation post For labor staff Is a post-matching preference score for a person Normalizing results by post attributes Normalizing results with personnel attributes The normalized Euclidean distance inverse ratio is calculated, the higher the post matching preference score is, the more the post is prone to matching the labor staff, the post matching preference score is divided by a theoretical maximum distance value for normalization processing, and a post preference vector set is formed ; S22, standardizing result set based on personnel attribute For each staff member Constructor preference vector Personal preference vector Representing personnel A set of preference scores for all short-term labor posts, wherein each person matches a preference score Representing personnel To the post Person-to-post person-to-person matching preference scores Normalizing results by personnel attributes Standardized results with post attributes Inversely proportional calculation and normalization are carried out on the normalized Euclidean distance between the two to form a personnel preference vector set ; S23, collecting post preference vectors With person preference vector set All scoring values in the set are respectively ranked from high to low to obtain an initial preference priority matrix Initial preference priority matrix Each element of (2) Representation post For personnel Ranking of (c).
  6. 6. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 5, wherein the step S3 comprises the following steps: s31, priority matrix based on initial preference Multidimensional preference sequence relation between middle posts and personnel, and multi-species post dynamic preference adjustment factor set is constructed Each post dynamic preference adjustment factor in the multi-species post dynamic preference adjustment factor set Representing short-term labor positions Dynamic adjustment degree of stability of personnel preference; s32, initializing each short-term labor post The state of (2) is a vacant state, and the maximum number of matching people of each post is defined as Each post current matching personnel set is Initializing each staff member The state is unmatched, and the current matching position is And setting dynamic iteration turns with positions for making matching requests as All posts and personnel are not matched at the beginning; S33, in each dynamic iteration round Middle and short term labor post Dynamic preference adjustment factor based on multiple kinds of posts Preference score for post matching Performing dynamic correction to obtain a post matching preference score after dynamic adjustment : ; Wherein, the The preference scores are matched for the initially calculated posts, For multiple kinds of post dynamic preference adjustment factors, Is the post To the first For the labor staff in turn Accumulating the number of refusal times Increasing characterization to advance along with matching rounds; S34, labor staff In each round of dynamic iteration, according to the personal preference vector set Person-matching preference scores in (a) Performing preference correction on the received matching request, and defining a personnel preference threshold The threshold value is adjusted according to the current situation of the matched posts and the historical preference satisfaction degree of the personnel, and the judging conditions for receiving the matching request are as follows: When working staff If any post is not matched, if the post after dynamic adjustment matches the preference score And person matching preference scores Receive the post Is matched with the matching request of the (a); When working staff Matched post If it is new Person matching preference score of (a) Exceeding the matched post Person matching preference score of (a) And dynamically adjusted post matching preference scores Labor staff Accept new post Is matched with the request and released from the original post Matching of (2); S35, updating the post dynamic preference adjustment factor set when each round of dynamic iteration is finished And a set of person preference thresholds Wherein the post dynamic preference adjustment factor Dynamically up-regulating as post cumulative refused times increases, personnel preference threshold Receiving dynamic correction of the variance of the post dynamic preference score according to personnel history; s36, iterating continuously according to the mode from S33 to S35 until the number of matching persons meeting all posts reaches the maximum number of matching persons Or the personnel cannot obtain the more preferable posts, and finally obtain the initial matching result set ; S37, initial matching result set Satisfying the improved stable matching constraint, which is characterized in that the possibility of recombination of any unmatched post-personnel combination under the condition that the post matching preference score and the personnel matching preference score after comprehensive dynamic adjustment exceed the matched combination is avoided, and the initial matching result set simultaneously maintains the initial preference priority matrix The determined multidimensional preference order.
  7. 7. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 6, wherein the step S4 comprises the following steps: s41, respectively calculating post matching satisfaction scores for each pair of short-term labor post and labor staff matching pairs based on the initial matching result set, the post preference vector set and the personnel preference vector set Matching satisfaction scores with people Post matching satisfaction score indicates post For matched personnel Is a person matching satisfaction score representing the staff member For matched post Is satisfied with the degree of (3); s42, post matching satisfaction score The values are taken from the posts in the set of post preference vectors For labor staff Is a post matching preference score of (2) Namely, the matching score of the post side score is directly used as the matching satisfaction score of the post side score, and the person matching satisfaction score The values are taken from the staff in the set of staff preference vectors To the post Person matching preference score of (a) I.e. the matching score of the personal score is directly used as the matching satisfaction score thereof; s43, matching satisfaction score at post Matching satisfaction scores with people Building a match-to-overall satisfaction score based on (a) The overall satisfaction score of the matching pair is formed by weighting the post matching satisfaction score and the personnel matching satisfaction score proportionally, and the post matching satisfaction score weight coefficient is that Person matching satisfaction degree scoring weight coefficient II is , + =1, Weight coefficient one And weight coefficient two The value of (2) is used for reflecting the priority degree of the post side and the personnel side in the short-term labor market; s44, scoring overall satisfaction of all matching pairs in the initial matching result set Sequentially recording, and summarizing to form a matching satisfaction evaluation index result set 。
  8. 8. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 1, wherein the step S5 comprises the following steps: s51, collecting initial matching results Constructing an initial set of matching solutions for a basis Each matching solution Consists of a post and personnel matching pair which meets the post capacity limit and personnel uniqueness matching constraint and is in an initial matching result set Introducing structural disturbance on the basis; S52, defining a post shortage coefficient set Each post in the post shortage coefficient set Is the coefficient of urgency of (a) Indicating how rare the post matches the number of rounds in the current round: ; Wherein, the Is the post The set of matching human choices may be accepted, For the total number of labor staff, the more scarce the posts are, the shortage coefficient The closer to 1, the more preferably the post should optimize its matching structure; s53, based on position shortage degree coefficient Introducing a post offset probability function in generating each matched solution disturbance structure : ; Wherein, the And Respectively a weight coefficient three and a weight coefficient four, Post offset probability function Determining the probability that the post matching pair is adjusted in each round of disturbance; S54, setting each cuckoo search iteration round as Aggregating matching solutions As the input of the current solution set, a new round of matching solution set is generated according to the position shortage coefficient and Levy distribution disturbance The matching pair transformation direction is guided and controlled by combining the history matching path entropy memory matrix; s55, for each matching solution set Constructing a multidimensional satisfaction distribution vector And introducing a matching distribution skewness index Characterizing satisfaction degree balance degree of the matched solution set: ; Wherein, the To match solution sets Is included in the overall average satisfaction score of (1), matching distribution skewness index A smaller value indicates a more uniform degree of satisfaction, Scoring the overall satisfaction; s56, constructing an improved fitness function As evaluation basis for cuckoo search: ; Wherein, the Respectively a positive weight coefficient five and a positive weight coefficient six; s57, according to the improved fitness function Performing matching solution updating and replacing operations on the values of the number of the matching sequences, and selecting the matching solution with the optimal fitness value in the current round And judging whether the convergence criterion is satisfied, i.e. the current optimal solution is continuous The fitness improvement in the round iteration is smaller than a set threshold value or the maximum iteration round number is reached If the convergence condition is satisfied, the fitness function is finally improved Outputting the final optimized matching result ; S58, if the convergence criterion is not met, recording the history matching path of the round to a memory matrix, and collecting the matching solutions Inputting the search iteration to the next round of cuckoo, and continuously executing optimization until the condition is met; S59, optimizing the matching result finally And outputting.
  9. 9. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 2, wherein the step S6 comprises the following steps: S61, constructing a post preference dynamic change monitoring mechanism in the implementation process of the matching scheme, and collecting preference change data of short-term labor posts and labor staff in the labor market in real time, wherein the preference change data comprise preference changes of posts on job types, working time and salary levels, and preference updates of staff on skill levels, acceptable job types and expected salary attributes; S62, judging whether a post preference dynamic change event or a personnel preference dynamic change event exists or not based on preference change data acquired in real time, and triggering a dynamic re-matching mechanism if a preference vector set of any one of the post or personnel is detected to change; S63, optimizing the matching result according to the current final result Extracting post set with preference change as incremental optimization basic solution With people collection Optimizing the matching structure from the final matching result Is stripped to form an unchanged part matching set Match set with part needing to be matched again ; S64, performing an iterative cuckoo search algorithm on posts and personnel in the part matching set to be matched again, and utilizing the unchanged part matching set Initializing a new search solution by the provided structure information to generate a dynamic matching result set Matching the dynamic matching result set with the unchanged part matching set Merging to form updated final matching scheme ; S65, updating the final matching scheme Performing task mapping and post assignment, and dividing finally output multi-work-task short-term task dispatch types according to the consistency degree of the post attribute and the personnel attribute in the post matching; S66, matching the final task with the result And the corresponding dispatch task type is transmitted to a short-term labor dispatch management system through an interface and is used for post dispatch notification, personnel matching confirmation and task execution dispatch to complete final task allocation between the short-term labor post and the labor personnel.
  10. 10. The short-term labor market supply and demand matching optimization method based on the matching theory according to claim 9, wherein the multi-work-operation short-term labor task dispatch type output rule is as follows: The post salary level and personnel expected salary difference is less than or equal to 10%, post salary codes are consistent with personnel skill codes, the working time error is less than or equal to 2 hours, and tasks are sent for priority accurate matching; The post salary level and personnel expected salary difference is less than or equal to 20 percent, post salary codes are similar to personnel skill codes, and tasks are compatible matching dispatch; type C, station salary level and personnel expected salary difference >20% or station salary code and personnel skill code are not matched, but station shortage degree coefficient Personnel are in await job assignment's state and tasks are dispatched for emergency replenishment matches.
  11. 11. A computer system comprising a memory, a processor to store a computer program executable on the processor, wherein the processor executes the computer program to implement the short-term labor market demand-supply matching optimization method based on the matching theory as claimed in any one of claims 1 to 10.

Description

Short-term labor market supply and demand matching optimization method based on matching theory Technical Field The invention relates to the technical field of labor services, in particular to a short-term labor service market supply and demand matching optimization method based on a matching theory. Background With the rise of a flexible labor mode, a short-term labor market increasingly shows important roles in coping with seasonal production demand fluctuation, temporary project task execution and regional human resource adjustment, however, the problem of supply and demand matching of short-term labor is always a key bottleneck for restricting the improvement of labor market efficiency, in a traditional short-term labor dispatching process, supply and demand information is mostly released through off-line intermediaries, manual platform registration or on-line simple recruitment pages, enterprises screen labor staff one by one according to post demands, and staff select posts according to subjective experience, so that the manual matching mode has remarkable defects. Firstly, the lack of systematic preference modeling between the posts and the personnel, the matching basis is often limited to 'salary-skill' simple comparison, the multi-dimensional requirement difference of the posts on the type of work and the working time is ignored, so that the matching precision is low, the satisfaction is low, secondly, the existing automatic matching models mainly adopt a static scoring mode to generate a single priority list, the bidirectional preference dynamic change characteristics between the posts and the personnel are not fully considered, and the matching stability and satisfaction are evaluated without introducing a mechanism, so that the generated matching scheme is easily influenced by parameter disturbance and has insufficient stability. In further practice, partial platforms try to perform matching optimization based on heuristic rules or machine learning algorithms, but two defects generally exist, namely, firstly, a coordination mechanism between 'post matching priority' and 'personnel matching willingness' is lack of description, global optimization is difficult to ensure when multi-work and multi-constraint short-term posts are processed, secondly, under the background that market preference fluctuation is frequent, the existing models mostly adopt static matching structures, a mechanism for quickly responding to post or personnel preference change is lack, and dynamic evolution of the matching structures is difficult to realize. In addition, the current matching strategy often ignores the uniformity of individual satisfaction distribution in a matching structure, so that part of posts are very easy to be subjected to long-term unmanned matching due to the characteristic of 'cold door', or the 'matching starvation' phenomenon of matching failure after multiple attempts by personnel is caused, the overall matching efficiency is influenced, the service reliability of a dispatching management system is weakened, and the core appeal of a 'quick, accurate and adjustable' matching mode by a modern labor market cannot be met, so that a more scientific and dynamically-responded supply and demand matching optimization method is needed to break through the technical bottleneck of the existing labor dispatching system. Disclosure of Invention The invention aims to provide a short-term labor market supply and demand matching optimization method based on a matching theory, and the invention remarkably improves the adaptation precision and the generation efficiency of a new scheme. According to the embodiment of the invention, the short-term labor market supply and demand matching optimization method based on the matching theory comprises the following steps: S1, acquiring a short-term labor post information data set and a labor personnel information data set, and performing attribute standardization processing to generate a post attribute standardization result set and a personnel attribute standardization result set; s2, building a post preference vector set based on the post attribute standardization result set, building a personnel preference vector set based on the personnel attribute standardization result set, and generating an initial preference priority matrix; S3, executing an improved Gale-shape matching algorithm on the basis of the initial preference priority matrix to obtain an initial matching result; s4, calculating a matching satisfaction evaluation index result set according to the post preference vector set and the personnel preference vector set for the initial matching result; S5, generating an initial matching solution set by taking the initial matching result as a seed solution, obtaining an optimized matching result, judging whether the optimized matching result reaches a convergence criterion, if yes, taking the optimized matching result as a current effective matching scheme, otherwise,