How to Evaluate College Students’ Green Innovation Ability—A Method Combining BWM and Modified Fuzzy TOPSIS

Abstract

Currently, enterprises are paying more and more attention to green innovation in order to reduce the negative impact on the environment and achieve sustainable development. Different from evaluating the green innovation ability of enterprises, how to evaluate a person’s green innovation ability from the perspective of open innovation is a new direction. This study aims at proposing a novel framework to evaluate college students’ green innovation ability. First, a set of criteria to evaluate college students’ green innovation ability is determined by experts using the panel consensus approach. Second, the best worst method (BWM) is used to calculate the weights of criteria. Lastly, modified fuzzy technique for order of preference by similarity to ideal solution technique (TOPSIS) technique is adopted to rank college students’ green innovation ability. Moreover, a case study is presented to verify the validity of the proposed method. The results provide implications for academic, industry, and policymakers. Specifically, through the evaluation of college students’ green innovation ability, universities can find the inadequacies in culturing students’ green innovation ability and strengthen education in the aspects corresponding to the criteria with high global weights. Companies can select the talented internship students with green innovation ability as employees and should realize that it is good for them to cultivate college students with universities. Policymakers of the education department can trace the quality of education in college senior students’ green innovation ability every year. In addition, they should pay more attention to establish policies regarding those criteria with high global weights.

1. Introduction

In the background of tightening resource constraints and the process of innovation-driven growth strategy, green innovation has become the core content and effective driving force of green growth mode [1]. In 2008, the Environmental Innovation Research Group of China International Cooperation Council on Environment and Development (CICCED) put forward the concept of “green innovation” in its report on “Mechanism Innovation and Harmonious Development”. Since then, green innovation has been rapidly promoted across China [2]. Particularly, enterprises carry out green innovation to cope with environmental protection pressure and achieve sustainable development [3,4,5]. Therefore, green innovation ability evaluation is important for enterprises to recognize their strengths and weaknesses in green innovation, thus helping them improve their market competitiveness [6,7,8].

Notably, personnel quality [6] and education and training level of R&D personnel [7] are important indicators in evaluating green innovation ability of an enterprise. Thus, different from evaluating green innovation ability of enterprises, how to evaluate a person’s green innovation ability is a new direction. Those persons can be employees working in enterprises or college students who will work for enterprises when they graduate. There are two motivations for evaluating college students’ green innovation ability. One is to better select talents with green innovation ability for enterprises. The other is to improve the education quality of green innovation for universities. The two folds help improve the overall level of green innovation for the society.

College students’ green innovation ability is a little different from the green innovation ability of enterprises including green product, green process, and green service as main elements [4]. Furthermore, it differs from green consumption ability and behavior which concern personal behavior and consciousness in green development. College students’ green innovation ability can indirectly promote the enhancement of green innovation ability for enterprises or areas. It emphasizes a person’s innovation ability on green development, including green technology innovation ability, green management innovation ability, and so on [4,5,6,7,8,9,10].

As far as we know, college students’ green innovation ability has received little attention. Some studies have been done in green innovation ability from the perspective of enterprises [4,5,6,7,8,9,10] or areas [11]. Few studies [12] consider green innovation ability from the perspective of green innovation ability cultivation, i.e., green innovation education in universities. Particularly, how to evaluate college students’ green innovation ability still requires exploration.

Therefore, this paper addresses how to evaluate the green innovation ability of college students, including the following content: (1) to determine the criteria of college students’ green innovation ability; (2) to determine weights of the criteria of college students’ green innovation ability; and (3) to propose a novel method to evaluate students’ green innovation ability.

This paper aims at finding some implications of three aspects: for enterprises to select college students with excellent innovative ability; for universities to cultivate talents with green innovation ability; and for policymakers to establish policies to improve green innovation of the society.

The contributions of this paper are as follows:

• A set of criteria (including four main criteria and thirteen sub criteria) to evaluate college students’ green innovation ability is proposed.
• A novel three-phase framework to evaluate college students’ green innovation ability is formulated from the perspective of open innovation. The weights of the criteria are calculated using the best worst method (BWM), which requires less comparison data and leads to more consistent comparisons, compared with other multiple-criteria decision-making (MCDM) methods. Modified fuzzy technique for order of preference by similarity to ideal solution technique (TOPSIS) is adopted to rank the alternatives considering the relative importance of the two separations.
• Implications are summarized from three aspects, including the academic, industry, and policymakers.

The rest of this paper is organized as follows. Section 2 briefly reviews literature regarding green innovation and MCDM methods. Section 3 introduces a three-phase method which combines BWM and modified fuzzy TOPSIS technique. In Section 4, a case study is conducted. Results are discussed in Section 5 and conclusions are summarized in Section 6.

2. Literature Review

In this paper, we introduce the literature review from three folds: green innovation and open innovation, related MCDM methods, and green innovation ability evaluation methods.

2.1. Green Innovation and Open Innovation

“Green innovation”, “eco-innovation”, “ecological innovation”, and “environmental innovation” are similar terminologies [13]. This paper uses the same definition of green innovation described in [14]. Today, considerable research has been done on green innovation; see the recent survey [15]. The research themes/topics are multi-perspective, covering management and strategic dimension of green innovation (e.g., green innovation integration [16,17] and adoption [18,19,20] strategy; collaboration [21,22,23] and networking [24,25] in green innovation; green innovation management systems [26], green supply chain management [27,28], etc.); and performance (financial [29,30,31], non-financial [32,33] and both [34]).

Of note, open innovation plays a mediating role in enhancing green innovation [35]. Open innovation represents a high-potential scheme to share knowledge and collaborate [36]. Many firms find in open innovation the means to create innovation through the commercialization of scientific knowledge, such as collaborating with universities [37]. Open innovation through university-industry collaboration (UIC) is one of the promising mechanisms to develop new technologies [38]. UIC is a fundamental part of innovation systems [39] and helps a firm achieve superior innovation outcomes [40]. Several studies show that UIC has a positive impact on innovation [41,42]. Moreover, mega projects are an ideal setting for open innovation [43]. Research shows that projects used to recruit and train employees or teach students in universities have a distinct influence on innovation [44].

2.2. Related MCDM Methods

The MCDM methods have received considerable attention and achieved great success in many areas. Specifically, there are two types of MCDM methods. For more details of state-of-the-art surveys, please refer to [45,46,47]. Below, we briefly introduce the two types of methods.

(1)

MCDM methods based on multi-attribute utility and value theories. First, this type of approach requires building a decision matrix of alternatives. Then, a score for each alternative over each criterion is given by experts. Finally, combined with weights of criteria, the rating of each alternative can be obtained using some aggregation functions. Several methods belongs to this approach, such as TOPSIS [48], VIKOR (VIse Kriterijumska Optimizacija kompromisno Resenje in Serbian, multiple criteria optimization compromise solution) [49,50], MULTIMOORA (multiplicative multi-objective optimization by ratio analysis) [51], MACBETH (measuring attractiveness by a categorical based evaluation technique) [52], and UTA (utilities additives) [53].

Among the above methods, TOPSIS is the most widely used. Since it was proposed, many efforts have been made to modify [54,55,56] (or extend [57,58,59,60,61]) TOPSIS, and to compare [62,63] (or combine [64,65,66]) it with other MCDM methods.

(2)

MCDM methods based on outranking methods. This method is pairwise-based, which compares two alternatives regarding each criterion to obtain dominance degrees. Then, the out ranking is calculated using an aggregate function of dominance degrees. There are several common outranking methods, for instance, BWM [67], AHP (analytic hierarchy process) [68,69], PROMETHEE (preference ranking organization method for enrichment evaluations) [70], ELECTRE (ELimination Et Choix Traduisant la REalité in French, elimination and choice expressing the reality) [71], and GLDS (gained and lost dominance score) [72].

Among the above outranking methods, BWM has been demonstrated as better than AHP in consistency ratio, which yields more reliable results. Furthermore, BWM needs fewer pairwise comparisons than AHP, which means an easier and more efficient process. BWM has been widely used in areas such as supplier selection [10,73], airline industry [74,75], etc., either alone or combined with other MCDM methods.

2.3. Green Innovation Ability Evaluation Methods

Green innovation ability reflects the green innovation development level and potential of an enterprise, a region, or a person. There are considerable studies on the evaluation of the similar term “green innovation efficiency” [76,77,78]. However, there are a few studies on the evaluation of the green innovation ability of manufacturing enterprises. Methods include the cloud model combining decision-making trial and evaluation laboratory (DEMATEL) method [7], AHP combining entropy weight method [6], and AHP combining osculating value process (OVP) method [8]. Some researchers have studied the selection of suppliers in small and medium enterprises based on green innovation ability [9,10]. Furthermore, evaluation of regional innovation ability from a green and low-carbon perspective is studied using the fuzzy AHP method [11]. As far as we know, there is no research investigating the evaluation of the green innovation ability of a person, especially for college students from the open innovation perspective.

Therefore, this paper studies the evaluation of college students’ green innovation ability from the open innovation perspective based on UIC through a project. The evaluation method is BWM combining modified fuzzy TOPSIS, which can be better determine the weights of the criteria and rank the alternatives.

3. Methodology

This paper introduces a three-phase methodology (see Figure 1) that combines BWM and modified fuzzy TOPSIS. The first phase decides the selection criteria for evaluation of college students’ green innovation ability by experts from different areas. The second phase calculates criteria weights of main and sub criteria using BWM. Note, many methods are used to determine criteria weights; BWM requires fewer pairwise comparisons and performs better with respect to the consistency ratio compared to those methods. The third phase uses modified fuzzy TOPSIS method to rank college students’ green innovation ability. More details of each phase are presented below:

3.1. Identification of Criteria

Criteria encompass main and sub ones. First, literature regarding green innovation is reviewed and main and sub criteria are selected as indicators for evaluation. Then, experts from different areas discuss and achieve consensus on the criteria.

3.2. Calculation of Criteria Weights Using BWM

BWM is a MCDM method to derive criteria weights using pair comparisons. The steps of this method can be described as follows:

Step 1. Determine criteria for evaluation.

Main and sub criteria are identified. We denote the n main criteria as $\{c_1,c_2,\dots ,c_n\}$.

Step 2. Select the best and worst criteria of main and sub criteria.

The best (i.e., most preferred, or most important) criterion and worst (i.e., least preferred, or least important) criterion are selected by the experts from main as well as sub criteria, respectively.

Step 3. Obtain best-to-others (BO) and others-to-worst (OW) vectors.

The preferences of the best criterion over the other criteria and the other criteria over the worst criterion are given by the experts using a number between 1 and 9. The larger the number, the more important the criterion is. BO and OW vectors are defined as follows:

$$\begin{gathered} V_B=(v_{B1},v_{B2},\dots ,v_{Bn}) \\ {V}_W={(v_{1W},v_{2W},\dots ,v_{nW})}^T \end{gathered}$$

where $v_{Bj}$ is the preference of the best criterion B over criterion j and $v_{jW}$ is the preference of criterion j over the worst criterion W. Note that $v_{BB}=v_{WW}=1$.

Step 4. Optimize the weights of all the criteria $(w_1^{*},w_2^{*},\dots ,w_n^{*})$.

The objective is to minimize error distances, which can be represented by minimizing the maximum absolute differences $\left|w_B-v_{Bj}{w}_j\right|$ and $\left|w_j-v_{jW}{w}_W\right|$. Thus, the optimal weights can be obtained through a min-max model below:

$${\min }_{}\max \{\left|w_B-v_{Bj}{w}_j\right|,\left|w_j-v_{jW}{w}_W\right|\}$$

(1)

$$\mathrm{s}.\mathrm{t}.{\displaystyle \sum _{j=1}^n{w}_j}=1$$

(2)

$$w_j\ge 0j=1,2,\dots ,n$$

(3)

Equations (1)–(3) can be converted to a linear programming formulation as follows:

$$\begin{gathered} \min \xi \\ \mathrm{s}.\mathrm{t}.{\displaystyle \sum _{j=1}^n{w}_j}=1 \\ {w}_j\ge 0j=1,2,\dots ,n \end{gathered}$$

(4)

$$\left|w_B-v_{Bj}{w}_j\right|\le \xi j=1,2,\dots ,n$$

(5)

$$\left|w_j-v_{jW}{w}_W\right|\le \xi j=1,2,\dots ,n$$

(6)

A unique solution (i.e., the optimal weights) can be obtained by solving the above linear programming model. The optimal value ${\xi }^{*}$ indicates the consistency of comparisons.

3.3. Ranking the Students’ Green Innovation Ability Using Modified Fuzzy TOPSIS

The TOPSIS method can rank the alternatives in the MCDM. Considering that crisp values cannot describe human judgements precisely, the fuzzy TOPSIS [79,80] method uses a linguistic scale for comparison of alternatives. For flaws in the TOPSIS method, studies have been done to modify the method. In this paper, we use the modified TOPSIS method proposed by Kuo [54]. The steps can be summarized as follows:

Step 1. Build the initial fuzzy matrix $\tilde{D}$.

Let each expert rate each student with respect to each criterion using a linguistic variable (see

Table 1. Linguistic variables and their corresponding fuzzy numbers.

Linguistic Variables	Corresponding Fuzzy Numbers
Very Low (VL)	(0, 0.1, 0.3)
Low (L)	(0.1, 0.3, 0.5)
Medium (M)	(0.3, 0.5, 0.7)
High (H)	(0.5, 0.7, 0.9)
Very High (VH)	(0.7, 0.9, 1.0)

), which can be described as a triangular fuzzy number ${\tilde{x}}_{ij}=(a_{ij},b_{ij},c_{ij})$. ${\tilde{x}}_{ijk}$ indicates the fuzzy performance score of student i evaluated in criteria j by expert k and ${\tilde{x}}_{ijk}=(a_{ijk},b_{ijk},c_{ijk})$. Let K denote the number of experts. The method of aggregating the experts’ fuzzy ratings is as follows:

$$a_{ij}=\underset{k}{\min }\left\{a_{ijk}\right\},b_{ij}=\frac{1}{K}{\displaystyle \sum _{k=1}^K{b}_{ijk}},c_{ij}=\underset{k}{\max }\left\{c_{ijk}\right\}.$$

(7)

The initial fuzzy matrix can be defined as Equation (8).

$$\tilde{D}=\left[\begin{array}{l}{\tilde{x}}_{11}{\tilde{x}}_{12}{\dots }_{}{\tilde{x}}_{1n}\\ {\tilde{x}}_{21}{\tilde{x}}_{22}{\dots }_{}{\tilde{x}}_{2n}\\ ⋮⋮\ddots ⋮\\ {\tilde{x}}_{m1}{\tilde{x}}_{m2}\dots {\tilde{x}}_{mn}\end{array}\right]$$

(8)

Step 2. Normalize the fuzzy decision matrix.

To ensure the fuzzy number belongs to [0, 1], let $\tilde{R}$ be the normalized fuzzy matrix below:

$$\tilde{R}=\left[\begin{array}{l}{\tilde{r}}_{11}{\tilde{r}}_{12}\dots {\tilde{r}}_{1n}\\ {\tilde{r}}_{21}{\tilde{r}}_{22}\dots {\tilde{r}}_{2n}\\ ⋮⋮\ddots ⋮\\ {\tilde{r}}_{m1}{\tilde{r}}_{m2}\dots {\tilde{r}}_{mn}\end{array}\right]$$

(9)

where

$${\tilde{r}}_{ij}=(\frac{a_{ij}}{c_j^{+}},\frac{b_{ij}}{c_j^{+}},\frac{c_{ij}}{c_j^{+}}),i=1,2,\dots ,m,j=1,2,\dots ,n$$

(10)

$$c_j^{+}=\underset{i}{\max }\left\{c_{ij}\right\},i=1,2,\dots ,m,j=1,2,\dots ,n$$

(11)

Step 3. Construct weighted normalized fuzzy decision matrix.

Let $\tilde{V}$ be the weighted normalized fuzzy decision matrix. It is defined as follows:

$$\tilde{V}=\left[\begin{array}{l}{\tilde{v}}_{11}{\tilde{v}}_{12}\dots {\tilde{v}}_{1n}\\ {\tilde{v}}_{21}{\tilde{r}}_{22}\dots {\tilde{v}}_{2n}\\ ⋮⋮\ddots ⋮\\ {\tilde{v}}_{m1}{\tilde{v}}_{m2}\dots {\tilde{v}}_{mn}\end{array}\right]$$

(12)

where

$${\tilde{v}}_{ij}={\tilde{r}}_{ij}(·)w_{j}i=1,2,\dots ,m,j=1,2,\dots ,n$$

(13)

Note, $w_j$ is obtained in Step 4 of Phase 2 through a linear programming formulation.

Step 4. Determine the fuzzy positive ideal reference point (FPIRP) and the fuzzy negative ideal reference point (FNIRP).

Let $A^{+}$ and $A^{-}$ represent FPIRP and FNIRP, respectively. They are defined as follows:

$$A^{+}=({\tilde{v}}_1^{+},{\tilde{v}}_2^{+},\dots ,{\tilde{v}}_n^{+})$$

(14)

$$A^{-}=({\tilde{v}}_1^{-},{\tilde{v}}_2^{-},\dots ,{\tilde{v}}_n^{-})$$

(15)

where ${\tilde{v}}_j^{+}=\underset{i}{\max }\left\{{\tilde{v}}_{ij}\right\}$ and $v_j^{-}=\underset{i}{\min }\left\{{\tilde{v}}_{ij}\right\}$, $j=1,2,\dots ,n$.

Step 5. Calculate the distances of each alternative to FPIRP and FNIRP.

Let $d_i^{+}$ and $d_i^{-}$ be the distances of each alternative to FPIRP and FNIRP, respectively. They are defined as follows:

$$d_i^{+}={\displaystyle \sum _{j=1}^{n}d({\tilde{v}}_{ij},{\tilde{v}}_{{}_j}^{+})},i=1,2,\dots ,m$$

(16)

$$d_i^{-}={\displaystyle \sum _{j=1}^{n}d({\tilde{v}}_{ij},{\tilde{v}}_{{}_j}^{-})},i=1,2,\dots ,m$$

(17)

where $d({\tilde{v}}_{ij},{\tilde{v}}_{{}_j}^{+})$ denotes the distance between ${\tilde{v}}_{ij}$ and ${\tilde{v}}_{{}_j}^{+}$ while $d({\tilde{v}}_{ij},{\tilde{v}}_{{}_j}^{-})$ denotes the distance between ${\tilde{v}}_{ij}$ and ${\tilde{v}}_{{}_j}^{-}$.

Step 6. Calculate the closeness coefficient and rank the order of alternatives.

The closeness coefficient (CC_i) is calculated as that proposed by [54], see Equation (18).

$$C{C}_i=w^{+}(d_i^{-}/{\displaystyle \sum _{i=1}^m{d}_i^{-}})-w^{-}(d_i^{+}/{\displaystyle \sum _{i=1}^m{d}_i^{+}})i=1,2,\dots ,m$$

(18)

where $-1\le {CC}_i\le 1$, $0\le w^{+}\le 1$, $0\le w^{-}\le 1$, and $w^{+}+w^{-}=1$. $w^{+}$ and $w^{-}$ indicate the importance levels of NIS and PIS, respectively.

4. Case Study

In this section, we illustrate the proposed method by applying it in a real case. Following the case’s background, the detailed process of the proposed method is presented, including identification of criteria, determination of criteria weights, and ranking students’ green innovation ability.

4.1. Case Background

Under the background of integration of higher education with industry in China, Company ‘A’ and University ‘B’ have a long-term cooperative relationship in education. Company ‘A’ is in charge of the green development of port logistics for Port ‘C’. Considering the importance of green development, Company ‘A’ and University ‘B’ cooperatively cultivate green innovation ability of students who major in logistics management. For green development practice, students are supervised by tutors from Company ‘A’. University ‘B’ is in charge of theoretical knowledge education on green development.

Therefore, how to evaluate students’ green innovation ability has become a concerning issue. In this study, the framework of the evaluation method is proposed. An expert team was composed of four experts from Company ‘A’ and University ‘B’. All the experts have rich experience in green development, and they are all 45 to 55 years old. Specifically, two experts are managers in the planning department of Company ‘A’, who have drawn the green port guidelines for Port ‘C’. The other two experts are professors from the logistics management department of University ‘B’, who teach supply chain management and have done some research on green logistics. Below, we use the panel consensus approach to get a common consensus of the experts on the criteria and their pair comparison. In addition, five students were selected as the evaluation objects, which were S1, S2, S3, S4, and S5. All of them are senior students from the logistics management department of University ‘B’, who are about twenty years old. They have learned a lot of knowledge at school, such as logistics management, port management, environmental management, etc. Most importantly, they have participated in a project in improving green development of Port ‘C’ when they were interning in Company ‘A’. The four experts rated the five students independently.

4.2. Identification of Criteria

Green innovation is divided into two major categories: green technology innovation and green management innovation [81]. Green technology innovation includes green process innovation and green product innovation, both of which can improve the environmental and economic performance of enterprises [31]. Green management innovation refers to adopting new environmental management measures and practices to improve production and management processes thus reducing negative environmental impacts [82,83]. Therefore, we define green technology innovation ability and green management innovation ability as two folds of college students’ green innovation ability. Moreover, because knowledge accumulation and achievements are important indicators for evaluating students’ innovation ability [12], we added green innovation knowledge accumulation and achievements as evaluation indicators of the college students’ green innovation ability.

Below, we go through related literature to select some criteria for evaluation. Then, experts were invited to give some suggestions on the criteria. Finally, four main criteria and thirteen sub criteria were selected, see

Table 2. Main and sub criteria for evaluation.

Main Criteria	Sub Criteria	Description	References
Green innovation knowledge accumulation (C1)	Basic knowledge accumulation related to green development (C11)	Knowledge about environmental issues and the concept and evolution of green development	[84,85]
	Professional frontier knowledge accumulation related to green innovation (C12)	Such as green logistics and green supply chain, etc.	[12]
	Interdisciplinary knowledge accumulation related to green innovation (C13)	Such as green product design, green materials, green equipment, green recycling, green packaging, digital technology, etc.	[12]
Green technology innovation ability (C2)	Green thinking (C21)	Integrate environment concerns and frontier technologies in port development	[86]
	Green product innovation ability (C22)	Skills in improving ecological maintenance, environmental protection, and green initiatives	[4,5]
	Green process innovation ability (C23)	Skills in improving energy saving and emission reduction, cleaner production, and process upgrading	[4,5,6,7,8]
Green management innovation ability (C3)	Systems thinking (C31)	Develop an effective strategy and a broad collection of analytical skills	[87,88]
	Green management mechanism innovation ability (C32)	Mechanism design for improving environment management, energy management, quality management, etc.	[4,5,6,7,8]
	Green management practice innovation ability (C33)	Practices in improving environment management, energy management, quality management, etc.	[4,5,6,7,8]
Green Innovation Achievements (C4)	Papers related to green innovation (C41)	The number and quality of published papers related to green innovation	[12]
	Patents related to green innovation (C42)	The number of patents related to green innovation	[12]
	Scientific and technological works related to green innovation (C43)	The number of award-winning works related to green innovation	[12]
	Promotion and application of green innovation-related achievements (C44)	Promotion and application of the above achievements	[12]

. Specifically, scores of sub criteria on green innovation knowledge accumulation (C1) and green innovation achievements (C4) were rated by experts from University ‘B’. Scores of sub criteria on green technology innovation ability (C2) and green management innovation ability (C3) were rated by experts from Company ‘A’, through a project in improving green development of Port ‘C’.

4.3. Determination of Criteria Weights

First, the experts determine the best and worst criteria, which are C3 and C1, respectively. Then, preferences of the best criterion over the other criteria (BO) and the other criteria over the worst criterion (OW) are given by the experts using a number between 1 and 9, see

Table 3. Main criteria comparisons.

BO	C1	C2	C3	C4
Best criteria: C3	9	2	1	2
OW	Worst criteria: C1
C1	1
C2	5
C3	9
C4	4

. In addition, using the same method, pairwise comparisons of the sub criteria in each main criterion are presented in

Table 4. Pairwise comparisons of sub criteria in green innovation knowledge accumulation (C1) sub criteria.

BO	C11	C12	C13
Best criteria: C13	9	2	1
OW	Worst criteria: C11
C11	1
C12	4
C13	9

,

Table 5. Pairwise comparisons of sub criteria in green technology innovation ability (C2) sub criteria.

BO	C21	C22	C23
Best criteria: C23	9	3	1
OW	Worst criteria: C21
C21	1
C22	4
C23	9

,

Table 6. Pairwise comparisons of sub criteria in green management innovation ability (C3) sub criteria.

BO	C31	C32	C33
Best criteria: C33	9	2	1
OW	Worst criteria: C31
C31	1
C32	3
C33	9

and

Table 7. Pairwise comparisons of sub criteria in green innovation achievements (C4) sub criteria.

BO	C41	C42	C43	C44
Best criteria: C44	9	4	3	1
OW	Worst criteria: C41
C41	1
C42	2
C43	4
C44	9

.

Weights of main and sub criteria are shown in

Table 8. Weights of main and sub criteria.

Main Criteria		Sub Criteria
Criteria	Weights	Criteria	Local Weights	Global Weights	Ranking
C1	0.053	C11	0.071	0.004	13
		C12	0.304	0.016	11
		C13	0.625	0.033	9
C2	0.245	C21	0.071	0.017	10
		C22	0.243	0.060	5
		C23	0.686	0.168	2
C3	0.474	C31	0.077	0.036	7
		C32	0.288	0.137	3
		C33	0.635	0.301	1
C4	0.228	C41	0.060	0.014	12
		C42	0.154	0.035	8
		C43	0.206	0.047	6
		C44	0.580	0.132	4

. First, weights of main criteria were calculated using Equations (2)–(6). The consistency value $\xi$ was 0.018. It was close to zero, which means high consistency of pair comparisons. Then, we calculated local weights of sub criteria in each criteria using the same method. The consistency values were 0.018, 0.043, 0.058, and 0.036. They were all less than 0.1. Finally, the global weight of each sub criteria was obtained by multiplying its local weight and the weight of its corresponding main criteria.

4.4. Ranking the Students’ Green Innovation Ability

First, the four experts rate the five students from the aspect of each sub criterion using linguistic variables described in Table 1. The aggregating rating of each student by the experts was obtained using Equation (7). The initial fuzzy decision matrix is shown in

Table 9. Initial fuzzy decision matrix ($\tilde{D}$).

Students	C11	C12	C13	C21	C22	C23	C31	C32	C33	C41	C42	C43	C44
S1	(0.3, 0.6, 0.9)	(0.5, 0.7, 0.9)	(0.1, 0.4, 0.7)	(0.3, 0.6, 0.9)	(0.3, 0.5, 0.7)	(0.5, 0.7, 0.9)	(0.1, 0.3, 0.5)	(0.3, 0.5, 0.7)	(0.1, 0.3, 0.5)	(0.3, 0.5, 0.7)	(0.1, 0.3, 0.5)	(0.3, 0.5, 0.7)	(0.5, 0.7, 0.9)
S2	(0.3, 0.5, 0.7)	(0.1, 0.45, 0.7)	(0.5, 0.7, 0.9)	(0.1, 0.35, 0.7)	(0.3, 0.65, 0.9)	(0.3, 0.6, 0.9)	(0.3, 0.6, 0.9)	(0.1, 0.3, 0.5)	(0.5, 0.7, 0.9)	(0.5, 0.7, 0.9)	(0.1, 0.4, 0.7)	(0.5, 0.7, 0.9)	(0.3, 0.65, 0.9)
S3	(0.3, 0.5, 0.7)	(0.3, 0.5, 0.7)	(0.3, 0.6, 0.9)	(0.5, 0.7, 0.9)	(0.3, 0.5, 0.7)	(0.3, 0.6, 0.9)	(0.1, 0.5, 0.9)	(0.3, 0.5, 0.7)	(0.3, 0.5, 0.7)	(0.3, 0.6, 0.9)	(0.5, 0.7, 0.9)	(0.3, 0.5, 0.7)	(0.1, 0.35, 0.7)
S4	(0.3, 0.65, 0.9)	(0.1, 0.3, 0.5)	(0.3, 0.6, 0.9)	(0.5, 0.7, 0.9)	(0.3, 0.6, 0.9)	(0.3, 0.55, 0.9)	(0.5, 0.7, 0.9)	(0.5, 0.7, 0.9)	(0.3, 0.65, 0.9)	(0.1, 0.5, 0.9)	(0.3, 0.5, 0.7)	(0.3, 0.6, 0.9)	(0.1, 0.3, 0.5)
S5	(0.5, 0.7, 0.9)	(0.1, 0.3, 0.5)	(0.3, 0.6, 0.9)	(0.1, 0.3, 0.5)	(0.5, 0.7, 0.9)	(0.3, 0.5, 0.7)	(0.3, 0.55, 0.9)	(0.3, 0.5, 0.7)	(0.1, 0.3, 0.5)	(0.3, 0.5, 0.7)	(0.3, 0.5, 0.7)	(0.1, 0.3, 0.5)	(0.3, 0.6, 0.9)

. Then, normalized and weighted normalized fuzzy decision matrixes were obtained by Equations (10) and (13), respectively, see

Table 10. Normalized fuzzy decision matrix ($\tilde{R}$ ).

Students	C11	C12	C13	C21	C22	C23	C31	C32	C33	C41	C42	C43	C44
S1	(0.33, 0.67, 1.0)	(0.56, 0.78, 1.0)	(0.11, 0.44, 0.78)	(0.33, 0.67, 1.0)	(0.33, 0.56, 0.78)	(0.56, 0.78, 1.0)	(0.11, 0.33, 0.56)	(0.33, 0.56, 0.78)	(0.11, 0.33, 0.56)	(0.33, 0.56, 0.78)	(0.11, 0.33, 0.56)	(0.33, 0.56, 0.78)	(0.56, 0.78, 1.0)
S2	(0.33, 0.56, 0.78)	(0.11, 0.5, 0.78)	(0.56, 0.78, 1.0)	(0.11, 0.39, 0.78)	(0.33, 0.72, 1.0)	(0.33, 0.67, 1.0)	(0.33, 0.67, 1.0)	(0.11, 0.33, 0.56)	(0.56, 0.78, 1.0)	(0.56, 0.78, 1.0)	(0.11, 0.44, 0.78)	(0.56, 0.78, 1.0)	(0.33, 0.72, 1.0)
S3	(0.33, 0.56, 0.78)	(0.33, 0.56, 0.78)	(0.33, 0.67, 1.0)	(0.56, 0.78, 1.0)	(0.33, 0.56, 0.78)	(0.33, 0.67, 1.0)	(0.11, 0.56, 1.0)	(0.33, 0.56, 0.78)	(0.33, 0.56, 0.78)	(0.33, 0.67, 1.0)	(0.56, 0.78, 1.0)	(0.33, 0.56, 0.78)	(0.11, 0.39, 0.78)
S4	(0.33, 0.72, 1.0)	(0.11, 0.33, 0.56)	(0.33, 0.67, 1.0)	(0.56, 0.78, 1.0)	(0.33, 0.67, 1.0)	(0.33, 0.61, 1.0)	(0.56, 0.78, 1.0)	(0.56, 0.78, 1.0)	(0.33, 0.72, 1.0)	(0.11, 0.56, 1.0)	(0.33, 0.56, 0.78)	(0.33, 0.67, 1.0)	(0.11, 0.33, 0.56)
S5	(0.56, 0.78, 1.0)	(0.11, 0.33, 0.56)	(0.33, 0.67, 1.0)	(0.11, 0.33, 0.56)	(0.56, 0.78, 1.0)	(0.33, 0.56, 0.78)	(0.33, 0.61, 1.0)	(0.33, 0.56, 0.78)	(0.11, 0.33, 0.56)	(0.33, 0.56, 0.78)	(0.33, 0.56, 0.78)	(0.11, 0.33, 0.56)	(0.33, 0.67, 1.0)

and

Table 11. Weighted normalized fuzzy decision matrix ($\tilde{V}$ ).

Students	C11	C12	C13	C21	C22	C23	C31	C32	C33	C41	C42	C43	C44
S1	(0.001, 0.003, 0.004)	(0.009, 0.012, 0.016)	(0.004, 0.015, 0.026)	(0.006, 0.011, 0.017)	(0.02, 0.033, 0.047)	(0.093, 0.131, 0.168)	(0.004, 0.012, 0.02)	(0.046, 0.076, 0.107)	(0.033, 0.1, 0.167)	(0.005, 0.008, 0.011)	(0.004, 0.012, 0.019)	(0.016, 0.026, 0.037)	(0.073, 0.103, 0.132)
S2	(0.001, 0.002, 0.003)	(0.002, 0.008, 0.012)	(0.018, 0.026, 0.033)	(0.002, 0.007, 0.013)	(0.02, 0.043, 0.06)	(0.056, 0.112, 0.168)	(0.012, 0.024, 0.036)	(0.015, 0.046, 0.076)	(0.167, 0.234, 0.301)	(0.008, 0.011, 0.014)	(0.004, 0.016, 0.027)	(0.026, 0.037, 0.047)	(0.044, 0.095, 0.132)
S3	(0.001, 0.002, 0.003)	(0.005, 0.009, 0.012)	(0.011, 0.022, 0.033)	(0.009, 0.013, 0.017)	(0.02, 0.033, 0.047)	(0.056, 0.112, 0.168)	(0.004, 0.02, 0.036)	(0.046, 0.076, 0.107)	(0.1, 0.167, 0.234)	(0.005, 0.009, 0.014)	(0.019, 0.027, 0.035)	(0.016, 0.026, 0.037)	(0.015, 0.051, 0.103)
S4	(0.001, 0.003, 0.004)	(0.002, 0.005, 0.009)	(0.011, 0.022, 0.033)	(0.009, 0.013, 0.017)	(0.02, 0.04, 0.06)	(0.056, 0.103, 0.168)	(0.02, 0.028, 0.036)	(0.076, 0.107, 0.137)	(0.1, 0.217, 0.301)	(0.002, 0.008, 0.014)	(0.012, 0.019, 0.027)	(0.016, 0.031, 0.047)	(0.015, 0.044, 0.073)
S5	(0.002, 0.003, 0.004)	(0.002, 0.005, 0.009)	(0.011, 0.022, 0.033)	(0.002, 0.006, 0.009)	(0.033, 0.047, 0.06)	(0.056, 0.093, 0.131)	(0.012, 0.022, 0.036)	(0.046, 0.076, 0.107)	(0.033, 0.1, 0.167)	(0.005, 0.008, 0.011)	(0.012, 0.019, 0.027)	(0.005, 0.016, 0.026)	(0.044, 0.088, 0.132)

. Next, according to the maximum target in green innovation ability, we defined FPIRP and FNIRP as $v_j^{+}=(1,1,1)$ and $v_j^{-}=(0,0,0)$, respectively. Finally, the distances of each alternative to FPIRP and FNIRP were calculated using Equations (16) and (17), respectively. The closeness coefficient CC was calculated by Equation (18). Like [54], we let w⁺ = 0.4 and w⁻ = 0.6. In addition, we analyze the sensitivity of w⁺ and w⁻ in the next section. Ranking of students is shown in

Table 12. Calculation results and ranking of students.

Students	${\mathit{d}}_{\mathit{i}}^{+}$	${\mathit{d}}_{\mathit{i}}^{-}$	CCi	Ranking
S1	12.461	0.577	0.031	4
S2	12.352	0.694	0.054	1
S3	12.43	0.617	0.039	3
S4	12.371	0.682	0.052	2
S5	12.497	0.547	0.025	5

.

5. Results and Discussion

5.1. Results

The weights of main and sub criteria presented in Table 8 show that the global weights ranking is C33 > C23 > C32 > C44 > C22 > C43 > C31 > C42 > C13 > C21 > C12 > C41 > C11. Green management practice innovation ability (C33) is ranked as first. Green process innovation ability (C23) and green management mechanism innovation ability (C32) are ranked as second and third, respectively. The reason is that logistics management is a management science. Experts pay more attention in students’ management ability and students tends to be good at management. Furthermore, some other criteria regarding technology are important, such as promotion and application of green innovation-related achievements (C44), green product innovation ability (C22), and scientific and technological works related to green innovation (C43). Thinking, papers, and knowledge are not so preferred, since experts can evaluate students through a project to rate their application ability.

Table 12 shows that the ranking of the students is S2 > S4 > S3 > S1 > S5. Although Student 2 does not get the highest score in most of criteria, he apparently gets a higher score in C13, C33, C41, and C43 than the other students, see Table 11. Most importantly, he gets a high score in the most preferred criterion C33. Student 5 does not get the lowest score in most of criteria, but he gets a low score in the preferred criteria, such as C23, C43, and C21. Among them, his score in C43 is much lower than the other students. Thus, results show that scores do not determine ranking. However, scores have an impact on the distances of each alternative to FPIRP and FNIRP, which has a further impact on the closeness coefficient CC.

Table 13. CC of students with different w⁺ and w⁻.

Students	w+ = 0.2, w− = 0.8	w+ =0.4, w− = 0.6	w+ = 0.5, w− = 0.5	w+ = 0.6, w− = 0.4	w+ = 0.8, w− = 0.2
1	−0.123	−0.046	−0.008	0.031	0.108
2	−0.115	−0.03	0.012	0.054	0.138
3	−0.12	−0.041	−0.001	0.039	0.118
4	−0.116	−0.032	0.01	0.052	0.135
5	−0.126	−0.051	−0.013	0.025	0.1

shows that CC does not change with the value of w⁺ and w⁻. It represents that w⁺ and w⁻ are not sensitive parameters for the results. However, when we changed the value of the weight of C3 (the most preferred criterion), the results were influenced. We let the weight of C3 be 0.2, 0.4, 0.6, and 0.8, while the other criteria changed in proportion, see

Table 14. Weights of the main criteria with different values of the weight of C3.

Main Criteria	w3 = 0.474	w3 = 0.2	w3 = 0.4	w3 = 0.6	w3 = 0.8
C1	0.053	0.0806	0.060	0.040	0.020
C2	0.245	0.3726	0.279	0.186	0.093
C3	0.474	0.2000	0.4	0.6	0.8
C4	0.228	0.3468	0.260	0.173	0.087

. The responding ranking results are shown in

Table 15. Ranking of students with different weights of the main criteria C3.

Students	w3 = 0.474	w3 = 0.2	w3 = 0.4	w3 = 0.6	w3 = 0.8
1	4	2	2	4	5
2	1	1	1	2	2
3	3	4	4	3	3
4	2	3	3	1	1
5	5	5	5	5	4

. In general, weights of criteria have a bigger impact on ranking than w⁺ and w⁻.

5.2. Discussions

We conclude some implications from the following aspects:

(1)

Academic implication

Firstly, through the evaluation of college students’ green innovation ability, universities can find the inadequacies in culturing students’ green innovation ability. That is to say, teaching quality can be indirectly reflected. For example, criteria with low average score should be paid more attention. In the case study, students had a lower average score in criteria C12, C33, and C44 than the other criteria. This indicates that university ‘B’ should improve in teaching professional frontier knowledge, culturing green management practice innovation ability, and promoting the green innovation-related achievements.

Secondly, universities can strengthen education in the aspects corresponding to the criteria with high global weights. In the case study, the criteria with top three global weights were C23, C32 and C33. They represent green process innovation ability, green management mechanism innovation ability, and green management practice innovation ability, respectively. Those abilities need to be cultured through practices. Thus, universities should create more training opportunities when culturing students.

(2)

Industry implication

Firstly, through the evaluation, Company ‘A’ can select the talented internship students with green innovation ability as employees. In the case study, student S2 ranks first. This means that he/she has the greatest green innovation ability. Likewise, a company can use the evaluation method to select job applicants when recruiting. The difference is that companies should have an extra project test for the job applicants because they do not practice or participate in any project in the company. Thus, a project is needed for companies to rate the job applicants regarding the sub criteria of C2 and C3.

Secondly, companies should realize that it is good for them to cultivate college students with universities. A company can cooperate with several universities to conduct some projects, thus creating opportunities to select and cultivate students with green innovation abilities. Further, those projects can have a direct positive impact on the innovation ability of the company.

(3)

Policymakers implication

Firstly, policymakers of education department can trace the quality of education using this method to evaluate college senior students’ green innovation ability every year. They first identify the weakness of students through the rating in the evaluation. Then they can adjust the policy related to the weakness presented in the evaluation. In the case study, students achieved a low score in C33, which means that students need more green management practices to improve their innovation ability. Thus, policies on the integration of higher education with industry should be promoted to encourage industries to provide opportunities for college students to have more practice.

Secondly, policymakers of the education department should pay more attention to establishing policies regarding those criteria with high global weights. For example, incentive and support policies to improve green process innovation ability, green management mechanism innovation ability, and green management practice innovation ability should be established.

6. Conclusions and Future Research

Currently, improving green innovation ability for an enterprise is important, and so is cultivating students’ green innovation ability for high school. This paper studies how to evaluate college students’ green innovation ability. First, a three-phase method is proposed. The first phase is to identify criteria for evaluation based on the panel consensus approach. The second phase uses BWM to determine weights of criteria, which needs less pair comparison and has a higher consistency ratio compared to similar methods. In the third phase, modified fuzzy TOPSIS method is used to rank college students’ green innovation ability. Then, a case study is conducted to verify the validity of the proposed method. In all, this study contributes in two ways: one is putting forward criteria for evaluating college students’ green innovation ability; the other is proposing a new framework for the evaluation.

However, this study still has some limitations. First, this study can use other combinations of methods such as PROMETHEE, ELECTRE, GLDS and VIKOR, MULTIMOORA, and MACBETH. Comparison between methods should be thoroughly analyzed. Secondly, this study conducts a case study focused on one region. For better application to the whole of China, surveying of more universities and more students is needed. Lastly, the problem of some indicators influencing some other indicators should be solved by methods such as AHP-OVP, etc.