Intelligent Human Resource Allocation Methods And Applicaions For Hierarchical Organization

Human resource allocation is a process which based on the organization’s strategic goals,through premise activities such as competency model construction,personnel ability assessment,and specific allocation activities such as human resource planning,personnel selection,and team formation to arrange various types of personnel on the appropriate posts in a timely and reasonable manner.This process exists in all kinds of organizations in society,some of which are huge in size,complex in post setting,diverse in personnel types,and have hierarchical organizational structures,such as multi-national and cross-regional companies,large state-owned enterprises,and the military.In the existing related research,this type of organization is defined as a hierarchical organization.In today’s competitive social environment,human resources are the core asset of any organization,especially in hierarchical organizations,it is more necessary to rationally allocate human resources to meet the needs of their daily operations and strategic development.The traditional human resource allocation process is mainly based on decision makers’ and mainly have the problems such as different standards in job requirements,qualitative and subjective evaluation of personnel,and insufficient fit for personnel and job.With the development of information technology and the arrival of the era of big data,the cost of data storage has been continuously reduced,and the methods of data generation have become more diverse.Some powerful hierarchical organizations have gradually acquired and stored a large amount of human resource data.Based on these data,through intelligent optimization algorithms,data mining,natural language processing and other artificial intelligence technologies,the abovementioned problems can be effectively solved,making all activities in the process of human resource allocation more intelligent,and therefore realizing the comprehensive optimization of the ‘person-job’ system in the hierarchical organization.From the perspective of system engineering,this paper models and deconstructs the process of intelligent human resource allocation under the background of hierarchical organization,and studies the intellectualization of key activities in the whole process of human resource allocation based on the human resource data provided by G enterprise,which is a large state-owned enterprise in the domestic military industry.The main research work and contributions of this paper are listed as follows:1.A method of constructing a competency model dually driven by experts’ knowledge and HR data is proposed.The traditional competency model building process mostly relies on experts’ knowledge and experience,and is carried out through Delphi method,behavioral event interview method,questionnaire survey method,etc.This kind of method may cause the inaccuracy and invalid of competency model due to insufficient expertise and experience.With the gradual accumulation of a large amount of HR data in some organizations,by analyzing and mining the historical data of outstanding employees for a specific position,the implied characteristic can be found,thereby providing new support for the construction of the competency model.Based on the above analysis,this paper proposes a method for building a competency model driven by both expertise and data,which systematically integrates the experts’ knowledge and HR data in the process of building a competency model,making the process more scientific.In the part of case application,this paper verifies the effectiveness of the method by constructing the competency model of the middle-level manager of the strategic development department of the G enterprise’s Beijing headquarters.2.A method of accurate personnel profile is proposed.In order to have a more comprehensive and accurate description of the ability and behavioral characteristics of personnel,this paper focuses on how to make accurate personnel profile based on the structured data and unstructured data accumulated in hierarchical organizations and proposes a method for accurate personnel profile.This method firstly constructs the label system for personnel profile,and divides these labels into three types,namely,basic fact labels,quantitative ability labels and textual ability labels.For the generation of quantitative ability labels,this paper proposes three methods,including methods based on structured data and specific rules,methods based on human group behavior clustering,and methods based on interpersonal relationship measurement.Aiming at the generation of text-based competency labels,this paper models this process as a named entity recognition problem,and considers the scarcity of labeled data due to the privacy of human resources text data,a Transfer Learning model for Named Entity Recognition in Human Resource domain(TL-HRNER)is therefore proposed.In order to better test the performance of the TL-HRNER algorithm,this paper conducts entity recognition experiments based on a recruitment dataset.The experimental results also demonstrate that compared with other NER algorithms,the TL-HRNER method can achieve the best performance.This paper also takes part of the desensitized data of a middle-level manager of the strategic development department of the Beijing headquarters in G enterprise as an example,showing the whole process of accurate personnel profile based on the above methods.3.A human resource planning method based on meta-heuristic algorithm is proposed.In a hierarchical organization,human resource planning is usually carried out by controlling the flow of personnel inside and outside the organization or between units or positions at different levels within the organization.This paper first constructs a hierarchical personnel flow network to depict this process,and models the optimization problem of this network as a multi-stage integer programming problem.In order to solve this problem,inspired by modern intelligent optimization theory,this paper proposes two intelligent optimization operators,namely move operator and exchange operator,as well as a Late Acceptance Hill Climbing algorithm with TAbu and REtrieval(TARE-LAHC)strategy.Based on the above methods,this paper models and solves the human resource planning problem in the process of human resource allocation in the third quarter of 2020 in G enterprise.The objective function is set by considering from both organizational and personal development perspectives during the process of modeling.This paper also compares the TARE-LAHC algorithm with other classical algorithms,and the experimental results show that the TARE-LAHC algorithm can achieve the best solution.4.A data-driven Personnel selection method under group decision-making environment is proposed.After the human resource planning is completed,personnel will flow for the reasons such as recruitment,resignation,promotion,etc.,which resulting in vacancies for personnel selection.This paper focuses on a new type of personnel selection problem,namely,the data-driven personnel selection problem under group decision-making environment.This paper proposes a DAta-driven PErsonnel selection framework for Group decision-making(DAPEG)to solve this kind of problem.The decision-making process of the method is divided into three stages,including the determination of job qualifications and candidate profiles,preliminary screening of candidates and final decision-making of the expert group.The second stage integrates fuzzy C-means clustering algorithm and MULTIMOORA with improved Borda rule to automatically and intelligently process the HR data input to realize the preliminary screening of candidates.The DAPEG method introduces dual hesitant fuzzy numbers in the third stage to help decision makers evaluate and rank candidates from both positive and negative perspectives.It is worth mentioning that the weight setting problem under the group decision-making environment are all involved in the second and third stages.This paper also proposes a Linear Group Best-Worst Method(LGBWM)method for better solving this problem.Based on the DAPEG method,this paper solves the problem of personnel selection problem in the strategic development department,marketing department,logistics support department and R&D department of G enterprise’s Beijing headquarters in the process of human resource allocation in the third quarter of 2020.The experiments demonstrate that the method can effectively solve the data-driven personnel selection problem under the group decision-making environment.5.A team formation method based on multi-objective optimization algorithm is proposed.Compared with personnel selection,team formation not only needs to consider the matching degree between personnel and positions in the team,but also needs to consider the degree of cooperation between team members,which is a more complex human resources specific allocation activity.This paper studies two types of team formation problems in hierarchical organizations,namely team personnel replenishment problem and new team formation problem.First,a calculation method of person-job matching degree oriented to team formation scenarios is proposed.The method is flexible and can effectively deal with the evaluation of person-job fit degree for different types of positions in the team.Then,in order to maximize the overall person-job matching degree of the team and the cooperation degree of members in the team,this paper models these two types of the team formation problem as a multi-objective optimization problem.According to the characteristics of the problem,a NSGA-Ⅱ algorithm for Team Formation(TF-NSGA-Ⅱ)is proposed.This method is improved on the basis of the classic NSGA-Ⅱ algorithm,adding the accelerated individual fitness calculation mechanism,the greedy strategy based individual selection mechanism and the non-dominated sorting mechanism based early rejection strategy to make it more efficient and suitable for solving team formation problems.In the part of case application and experiment,this paper first takes the member supplementary problem of the leader team in the North China head office in the process of human resource allocation of G enterprise in the third quarter of 2020 as an example to show the solving process based on the above method.Then,based on specific rules,this paper generates a large-scale team supplementary example and a new team formation example to test the performance of the TF-NSGA-Ⅱ algorithm.The results show that when solving these two examples,compared with vanilla NSGA-Ⅱ algorithm,the solution efficiency of this algorithm is improved by 13.02% and 4.12% respectively.