Evaluation and Classification of Overseas Talents in China Based on the BWM for Intuitionistic Relations

International competition is becoming more and more severe, as a result of the development of economic globalization. Human resources have become part of the core of national competitiveness. This paper will focus on the evaluation of overseas talents in the Chinese situation. As the world’s largest developing country, China has unique advantages in attracting overseas talents. First of all, the Chinese culture provides an open and inclusive environment that is suitable for the development of overseas talents. China has always advocated cultural exchange between different nationalities, different countries and different areas. Moreover, China has always had a good tradition of respecting talents, recruiting remarkable people without stereotyping. After the reform and opening-up policy initiated in 1978, China has opened its door to welcome international friends from all over the world. Secondly, the rapid development of the Chinese economy and society has provided a broad space for overseas talents to start businesses. More and more overseas students and immigrants are being drawn to China and are becoming a new force in Chinese social development. Based on data from the Chinese Service Center for Scholarly Exchange, the ratio of overseas students who returned to work has reached 70%–80% percent. Thirdly, the Chinese Government attaches great importance to the role of overseas talents. At present, China is in an economic transition stage, and innovation has become the driving force. Overseas talents’ innovation behavior is an essential influencing factor in the nation’s innovative economic development [1]. Therefore, the Chinese Government strongly supports the introduction of overseas talents and expects them to contribute to technology industries. A series of talent introduction policies have been issued in succession, resulting in step by step improvement. In 2008, the central committee staff office of the Chinese Communist Party implemented a national plan named the Recruitment Program of Global Experts/Thousand-Talent Scheme. Moreover, some detailed plans followed this policy, for instance, the Young Overseas High-level Talents Introduction Plan, starting from 2011, whose purpose is to introduce the outstanding young talents of around 35 years of age; the Foreign High-level Experts Introduction Plan, starting from 2012, whose purpose is to introduce the overseas non-Chinese-origin experts who should work in China for three years continuously and nine months per year at least.

However, the gradual expansion of overseas talents in China creates new requirements for the management of overseas talent. In order to evaluate and classify foreign talents who have come to China, the Chinese government is required to build up a complete, systematic and scientific talents evaluation system. Therefore, this study aims to construct a targeted overseas talent evaluation index system based on the Chinese context and explores a combination of geometry consistency and the non-linear best-worst-method model. Furthermore, the study may enrich talent evaluation theory and provide a reference for overseas talents’ management by the Chinese Government and other developing countries.

According to the theory of talent evaluation, building a reasonable talent evaluation system is the foundation of human resource utilization. Academics have continuously explored human resource management. Lai [2] described the employee selection process as a multi-objective decision making problem. Iwamura and Lin [3] explained that the employee selection process required the accomplishment and aggregation of different factors. Labib, Williams and O’Connor [4] suggested an employee selection process that uses the Analytic Hierarchy Process (AHP) with four stages. As Chou, Sun and Yen [5] said, the pure AHP model for human resource evaluation and management has some shortcomings. On the one hand, the traditional evaluation method, such as the pure AHP method, mainly used explicit information to describe the evaluations. On the other hand, the criteria’s weights, which have great influence on the final results, are mostly obtained by the preference of decision makers [6]. To overcome these drawbacks, some scholars introduced fuzzy theory to the AHP methodology. Golec and Kahya [7] presented a hierarchical structure for selecting and evaluating the right employee based on five-scale measure tables. Zhong [8] analyzed the performance of high-level talents with the fuzzy evaluation method, which described the decisions by numbers from 1–9 and their reciprocals. Zhao et al. [9] studied the creative talents evaluation index system based on the nine-scale AHP method. Chou, Hsu and Yen [10] analyzed a country’s competitiveness in terms of their technological human resources with three kinds of criteria. In the last two years, some Chinese scholars did research on human resource and relevant policies [11,12,13,14,15]. Although these methods have further developed the traditional evaluation methods through the introduction of fuzzy theory, they are still not practical enough to deal with the case in our paper. As our study objective is overseas talents, they can be considered as the alternatives. The relevant government departments can organize some experts who are decision makers to evaluate alternatives using several criteria. That makes this problem turn into a multi-criteria decision making problem, which can be solved by Multi-Criteria Decision Making (MCDM) methods. Many scholars have studied this field [16,17,18,19]. We decided to utilize one kind of MCDM method to discuss the overseas talents’ evaluation and classification issues. Different useful tools have been produced for MCDM problems, such as the TOPSIS method [20], pairwise comparison [21] and elimination and choice translation reality [22]. Among them, pairwise comparison is a more practical tool for solving the overseas talents’ evaluation problem. Lazim and Norsyahida [23] studied pairwise comparison to deal with the MCDM problem. Suppose there is a set $\{x_1,x_2,\dots ,x_3\}$ with n alternatives. We can compare them pairwise and obtain the following original decision matrix under a certain criterion: where $x_{ij}$ stands for the comparison result of $x_i$ over $x_j$. Generally speaking, Decision Makers (DMs) can express their decision results based on Multiplicative Preference Relations (MPRs) [24] and Fuzzy Preference Relations (FPRs) [25]. However, DMs may sometimes not be sure about the intensities of preferences. This means that their results may be fluctuating in a certain range during the decision making process. In 1987, Satty and Vargas introduced Interval-Valued Multiplicative Preference Relations (IVMPRs) [26]. In 2004, Xu introduced Interval-Valued Fuzzy Preference Relations (IVFPRs) [27]. These two extensions allowed DMs to use interval valued numbers [28] to express their comparison results. However, IVFPRs and IVMPRs pay more attention to the preferred degree of alternative $x_i$ over $x_j$, ignoring the non-preferred degree. In order to cope with this issue, the concept of Intuitionistic Multiplicative Preference Relations (IMPRs) [29] and Intuitionistic Fuzzy Preference Relations (IFPRs) [30] appear as extensions of the traditional multiplicative preference relations. An IMPR is given as $R={({\alpha }_{ij})}_{n\times n}$, where ${\alpha }_{ij}=({\rho }_{{\alpha }_{ij}},{\sigma }_{{\alpha }_{ij}})$ is an Intuitionistic Multiplicative Number (IMN) given on the $1/9-9$ scale, where ${\rho }_{{\alpha }_{ij}}$ stands for the preferred degree and ${\sigma }_{{\alpha }_{ij}}$ stands for the non-preferred degree of alternative $x_i$ over $x_j$, including the hesitation information. An IFPR is given as $A={(a_{ij})}_{n\times n}$, where $a_{ij}=[a_{ij}^{-},a_{ij}^{+}]$ is an interval multiplicative number in which the preferred degree is located. In order to evaluate overseas talents fully, we not only consider the preferred degree of one talent over another, but also discuss the non-preferred degree.

Many scholars have introduced various kinds of preference relations by studying decision matrix X [27,31,32,33,34]. Their applications are about different fields of multi-criteria decision making problems, such as management information systems, forecast theory and economic systems [29,35,36]. In general, we should make $n(n-1)$ times of comparisons to finish the pairwise decision matrix X. Most studies about the original decision matrix have a similar assumption that is $x_{ij}=x_{ji}$. Under this assumption, the number of pairwise comparisons decreased from $n(n-1)$ to $n(n-1)/2$. If the number of alternatives is small, the pairwise comparison method is applicable. For example, Celik, Kandakoglu and Deha Er [37] proposed a multi-stage evaluation model under multiple criteria in order to manage the academic personnel selection and development in maritime education and training institutions. However, if the alternatives’ number is large, such as 20 candidates and 190-times of comparison, most published methods are not suitable. We need a method to decrease the comparison times, as well as ensuring reliability. Rezaei [38] proposed a decision making method: the Best-Worst-Method (BWM) Rezaei, Wang and Tavasszy [39] use BWM to propose an integrative approach that includes capabilities and willingness as two dimensions for evaluating and subsequently segmenting suppliers. Rezaei [40] applied a linear BWM to a car choosing problem. Compared to the other MCDM methods, BWM has not been applied and published widely. Rezaei et al. [41] apply the BWM to find the best suppliers from among the qualified suppliers. BWM’s biggest highlight is the pairwise comparison times dropping from $n(n-1)/2$ to $2n-3$. As BWM noted, it is not hard for people to pick out the best one and the worst one among the alternatives under a certain criterion. Determining how much the best one is superior to the others and how much the others are superior to the worst are the difficult parts. BWM is an effective method to deal with comparison times’ problems. The BWM expressed the comparison results by numbers from $\{1,2,3,4,5,6,7,8,9\}$ and ignored the reciprocals of each pair to avoid the trouble caused by unequal distance between fractional comparisons. Rezaei [38] discussed a practical application of a college students’ cell phone selection problem. The result showed that without considering the reciprocals, the method was credible, as well. However, ignoring the reciprocals is not sufficient for some other situations, especially for the talent assessment problems, in which unequal distance between fractional comparisons does exist. This paper proposes a non-linear BWM model to cope with the overseas talents evaluation problem. As discussed above, this model will be built under the intuitionistic relations environment. This means that compared with the original BWM, decision makers not only assess how much the best one is superior to the others, but also assess how much the others are superior to the worst. So do the others to the worst. Then, constructing the non-linear BWM would simultaneously increase the efficiency and practicability of our method.

Besides, the consistency degree of decision matrix X is a very important property, which also attracts many scholars’ attention. In the process of pairwise comparison, generalized consistency means that if alternative $x_i$ is better than $x_k$ and $x_k$ is better than $x_j$, then $x_i$ is better than $x_j$. Furthermore, if the three preferred degrees of $x_i$ over $x_k$, $x_k$ over $x_j$ and $x_i$ over $x_j$ satisfy the preference relations, then they satisfy consistency. Wang and Chen [42] discussed comparison consistency depending on the incomplete fuzzy linguistic preference relations. Liu et al. [43] introduced a goal programming model to deal with the incomplete interval multiplicative preference relation problems, which was based on the consistency property. Then, we also introduce a weight-determining system, based on geometry consistency defined in this paper. This methodology will help to calculate the weight vector of decision makers and criteria. For evaluating the overseas talent problem, we consider the weighting issues under the intuitionistic preference relations environment, which contains preferred degree and non-preferred degree. Decision makers are weighed according to their geometry consistent degree, the criteria’s weights obtained subsequently. Further, decision makers and criteria can obtain their weights by the consistency degree. We use an example from Jiang et al. [44], in which they focused on calculating the missing elements and did not give enough consideration to the consistency of decision matrices, to demonstrate the validity of our method.

The main contributions of this paper are two-fold. Firstly, we construct an evaluation system and introduce a geometry consistency degree to measure the criteria weights, with its validity demonstrated by a published example. Secondly, we construct a non-linear BWM model for intuitionistic preference relations and show its practicability by solving an overseas talent evaluation and classification problem. Both of the two points contribute to MCDM problems. The rest of the paper is organized as follows. Section 2 introduces definitions about the weight determination. In Section 3, we give the concrete procedures about BWM for intuitionistic relations. In Section 4, we examine a real example, which also belongs to an ongoing project led by the relevant department of Tianjin. The paper concludes in Section 5 with a summary of the advantages and drawbacks of our developments and suggesting some further research for the future.

Saaty and Vargas gave the concept of interval-valued multiplicative relation in 1987, as follows:

Xia et al. proposed that we can write the interval valued fuzzy number $u_{ij}=[u_{ij}^{-},u_{ij}^{+}]$ in another way: ${\alpha }_{ij}=({\rho }_{{\alpha }_{ij}},{\sigma }_{{\alpha }_{ij}})=(u_{ij}^{-},1/u_{ij}^{+})$. This introduced the definition of intuitionistic multiplicative preference relations.

Obviously, Definition 2 developed the Definition 1 in the aspect of considering the hesitance, making the decision results closer to the practical problems.

Depending on Definitions 2 and 3, we give the next definition:

Deriving from Definition 5, we obtain the following propositions about fuzzy-based geometric index.

(1)

$f_{a_{ij}}+f_{a_{ij}^c}=1$.

(2)

It is obvious that, $G_a^f>0$; $f_{a_{ij}}=\frac{1}{2}$, if ${\rho }_{a_{ij}}={\sigma }_{a_{ij}}$; $f_{a_{ij}}>\frac{1}{2}$, if ${\rho }_{a_{ij}}>{\sigma }_{a_{ij}}$; $f_{a_{ij}}<\frac{1}{2}$, if ${\rho }_{a_{ij}}<{\sigma }_{a_{ij}}$; $G^f$ indicates which kind of preference relation of $x_i$ over $x_j$ is dominant, the preferred degree or the non-preferred degree.

(3)

Let $a^1=({\rho }_{a^1},{\sigma }_{a^1})$ and $a^2=({\rho }_{a^2},{\sigma }_{a^2})$ be two IMNs. If ${\rho }_{a^1}\ge {\rho }_{a^2}$ and ${\sigma }_{a^1}\le {\sigma }_{a^2}$, then $f_{a^1}\ge f_{a^2}$.

Jiang et al. [44] proposed the method that decomposed an intuitionistic multiplicative preference number $U={({\alpha }_{ij})}_{n\times n}=({\rho }_{ij},{\sigma }_{ij})$ into two matrices: a preferred matrix $C={(c_{ij})}_{n\times n}$ describing the positive information and a non-preferred matrix $D={(d_{ij})}_{n\times n}$ describing the negative information given by DM with respect to alternative $x_i$ over $x_j$ [44]. Then, we develop the next definition to calculate the missing elements of an incomplete IMP’s corresponding interval fuzzy preference relation matrix $P={(p_{ij})}_{n\times n}$ by calculating the missing elements of $P^{-}={(c_{ij})}_{n\times n}$ and $P^{+}={(d_{ij})}_{n\times n}$, respectively.

From Equation (7), we rewrite it in the following way:

Based on Equation (9), we introduce the consistency degree $CD$ of CIFPR $P={(p_{ij})}_{n\times n}$: for expert $e_t$, $t\in \{1,2,\dots ,m\}$, where $\zeta =\sqrt{p_{ij}^{-}{p}_{ij}^{+}}\sqrt{(1-p_{ik}^{-})(1-p_{ik}^{+})}\sqrt{(1-p_{kj}^{-})(1-p_{kj}^{+})}$, $\eta =\sqrt{(1-p_{ij}^{-})(1-p_{ij}^{+})}\sqrt{p_{ik}^{-}{p}_{ik}^{+}}\sqrt{p_{kj}^{-}{p}_{kj}^{+}}$ and $i\le j$, $i,j,k\in \{1,2,\dots ,n\}$.

In addition, we can get the weight values for expert $e_t$ with respect to the consistency degree by: where $S_t=C{D}_t/min\{C{D}_t|t=1,\dots ,m\}$

Apply Equation (4) to get $G_t^f={(f_{{\alpha }_{ij}})}_{n\times n}$ to calculate the the score value $R{S}_l^t={({\sum }_{j=1}^n{f}_{{\alpha }_{ij}})}_{n\times 1}$. Then, we can obtain the weight value of criteria $c_l$: where $R{S}_i=({\sum }_{t=1}^m{w}_{t}R{S}_i^t)$.

Firstly, calculate the unknown elements of the decision results of the matrices $A_1$, $A_2$ and $A_3$. By Definition 4, obtain the corresponding fuzzy matrices. By Definition 8, decomposethe corresponding fuzzy matrices into two parts. By Definition 7, calculate the missing elements of $A_1$: ${\sigma }_{15}^{*}=0.6289$, ${\sigma }_{23}^{*}=0.4854$, ${\rho }_{13}^{*}=4.4977$, ${\rho }_{23}^{*}=2.2496$, ${\rho }_{34}^{*}=0.7777$; the missing elements of $A_2$: ${\sigma }_{15}^{*}=2.8110$, ${\rho }_{13}^{*}=0.4543$, ${\rho }_{24}^{*}=0.7340$; the missing elements of $A_3$: ${\sigma }_{35}^{*}=2.7139$, ${\rho }_{12}^{*}=3.3753$, ${\rho }_{35}^{*}=0.1734$. Secondly, calculate the consistency degree of each DM’s decision matrix by applying Equation (9) and obtaining: $C{D}_1=0.1786$, $C{D}_2=0.0550$, $C{D}_3=0.0389$. By Equation (11), we obtain the weight values for every expert: $w_1=0.4824$, $w_2=0.3034$, $w_3=0.2142$. Finally, apply Equation (12) to the weight vector of the criteria ${(3.5340,2.1758,1.3782,1.9237,3.4884)}^{\prime }$, and then, the ranking result is $x_1>x_5>x_2>x_4>x_3$.

The final results calculated by Jiang et al. [44] is: $S({\alpha }_1^{\prime })=6.0127$, $S({\alpha }_2^{\prime })=0.5342$, $S({\alpha }_3^{\prime })=0.2766$, $S({\alpha }_4^{\prime })=0.4939$, $S({\alpha }_5^{\prime })=6.0250$, with a ranking $x_5>x_1>x_2>x_4>x_3$. Obviously, there exists a difference about the ranking position about alternatives $x_1$ and $x_5$, while their final scare values are really close to each other. Y. Jiang et al. focused on calculating missing values and ignored the consistency degree’s influence. Our paper uses a simple way to calculate the unknown elements and takes the decision matrices’ consistency degree into consideration, with simpler calculating processes and a reasonable conclusion.

In order to cope with the overseas talent evaluating problem, we will introduce intuitionistic preference relations to express decision results. Additionally, a non-linear BWM model is constructed to rank alternatives

Tianjin, an economic center in the northern part of China, has constantly focused on the introduction of overseas talents. The relevant government agencies of the city have implemented some policies with regard to problems with respect to the introduction of overseas talents, such as how to evaluate and classify them based on the assessments and how to fairly offer them equitable treatment, which has become the main bottle neck. One important problem is improving assessment systems, including evaluation and classification. Here, we mainly discuss the evaluation part. Test samples are drawn out from the overseas talents who participated in the selection of the Tian Jin Haihe Friendship Award in 2015. To ensure the objectivity of the evaluation process, we omit specific information, such as the names of decision makers and overseas talents. We chose 20 overseas talents who came to the city for working or cooperating via the introduction by the government. With the purpose of making general statements for the 20 overseas talents, we list five criteria to evaluate them. The government assigned a group of three experts to evaluate the alternatives with each criterion.

We evaluated candidates from different aspects, considering some necessary qualities. In the early days, Harvard professor McClelland [48] proposed the concept of competency. He expounded on the relationship between individual and position from the quality and ability aspects. Spreitzer et al. [49] developed the prospector as an evaluation tool, which evaluated the managers’ potential to lead national and international enterprises. He researched managers from 21 countries and belonging to six major industries, who came to work in the United States, as multinational professional managers. The results of this research demonstrated the effectiveness of the prospector. Subsequently, more systematic talent evaluation problems were considered by researchers, including some Chinese scholars who did research mostly in the Chinese situation. Luo [50] considered that good internal ability and continuous creative ability were key points for the definition of talent. In addition, the creative achievements obtained by the talents should have a positive impact on the progress of society. Yan and Chen [51] conducted research based on case analysis and came to the conclusion that the marketer’s assessment was largely from three aspects, consisting of personality, knowledge and ability. Recently, Xu et al. [52] studied the talent evaluation index in regards to the high-level talents’ introduction in DongHu High-tech Zone. They constructed a high-level personnel evaluation index system containing four indexes, consisting of performance, ability, contribution and potential.

In this study, the criteria are sourced from Xu’s evaluation system [52], which was used to evaluate high-level talents in China. In addition, we collect advice from a sample of 15 experts, who are good at talent evaluation, in order to extract factors of overseas talents’ assessment. This group includes university professors, headhunting company managers, company executives and government review experts. We first explained the problem to these experts, and then, they were given some documents describing the criteria for evaluating overseas talents and a comprehensive list of the characteristics of overseas talents. Finally, after the interviews, we analyzed the information and created the index system of overseas talents’ evaluation, which contains basic quality, capacity, contribution, development potential and internationalization, as shown in Table 5.

It is important to notice that internationalization is a key-influencing factor for overseas talents [53]. On the one hand, global perspective refers to international experience and cognition of the world technology market [54]. Someone who has a global perspective may gain an insight into development opportunities, which may come from the differences between countries, such as the business opportunities that come from the technology and application differences between different countries, as well as differences in market demand and supply. On the other hand, overseas talents need to have good cross-cultural adaptability, which can help them to easily tackle the obstacles of cultural differences. Cross-cultural adaptability is a capability that appropriately deals with interactions among people from different countries and cultures [55]. Specifically, overseas talents with good cultural adaptability can identify similarities and differences between the new culture and their own culture and can actively coordinate the cultural conflicts, so that they can integrate into the new culture.

After constructing the criteria system, in order to evaluate and classify the chosen 20 overseas talents, we follow the subsequent decision making procedures, depending on the definitions and models from Section 2 and Section 3.

Step 1: From Definition 2, decision maker $e_t$, $t\in \{1,2,3,\}$, gives his/her decision matrix about criteria $c_1$, $l\in \{1,2,3,4,5\}$ (standing for basic quality, capacity, contribution, development potential and internationalization, respectively).