1. Introduction
Extreme changes in climate and frequent natural disasters have forced governments to pay more attention to environmental protection and have enacted many environmental protection regulations and penalties. In recent years, international governments have developed a number of environmental policies for energy-intensive industries (EIIs), hoping to reduce anthropogenic greenhouse gases emissions through legislation. EIIs include industries such as electronics, chemicals, machinery, petroleum, automotive, and biotechnology. They emit more than 45% of all industries and public activities [
1]. For global sustainable development, the World Trade Organization (WTO), the World Health Organization (WHO), the European Union (EU), and other international organizations have enacted many environmental protection monitoring legislations and agreements [
2].
The biopharmaceutical industry is a relatively new energy-intensive industry that is recognized as one of the most promising industries in the 21st century, and its development is critical to the technological advancement of global healthcare [
1,
3]. Emerging biotech pharmaceuticals are made up of complex biomolecules that provide solutions for chronic and debilitating diseases [
4]. Biotech products have been approved for marketing in Europe, and the market value of these products is expected to reach US
$35 billion in 2020 [
5]. The biopharmaceutical industry uses bio-based products to make commercially valuable drugs, including hormones, fusion proteins, cytokines, blood factors, vaccines, and redox molecules [
6]. The biopharmaceutical industry strives to meet the rigorous standards required for the production of therapeutic drugs through a series of complicated manufacturing processes and costly clinical trials. To minimize investment costs, many companies look for partners or ways to outsource. Because of the requirements of advanced technology, high investment, and long-term R&D cycles, the pharmaceuticals industry is classified as a high-risk industry [
7]. Therefore, it is especially advantageous for biopharmaceutical companies to form strategic alliances with other companies upstream and downstream the supply chain to enhance competitiveness, including shortening product development time, reducing development costs and risks, and increasing product diversity.
The goal of a strategic alliance is to integrate two or more companies, and the joint management of the overall supply chain can achieve resource sharing and market diversification. In general, strategic alliances involve formal legal or private informal partnerships, and partners can complement their strengths and weaknesses to reduce business risk [
8]. The biggest advantage of strategic alliances in the biopharmaceutical industry is the ability to jointly develop more valuable biotech products and promote the development of human health care [
9]. Due to the rise of environmental awareness, governments in different countries have established environmental regulations for the biopharmaceutical industry through legislative units.
At present, the most common method of strategic alliance partner evaluation in the biopharmaceutical industry is financial cost benefit analysis, which focuses on the profit and loss balance of business operations, that is, financial and cost indicators, ignoring the goal of environmental protection [
10,
11]. According to the literature review, the strategy for partner selection for biopharmaceutical production still focuses on financial performance [
12,
13]. In addition, according to Ramasamy et al. [
1], environmental standards are not talked about in the research related to the biopharmaceutical industry. Green biopharmaceutical production is a modern production model that takes into account environmental impacts and resource efficiency. The goal is to minimize the negative impact of pharmaceutical production on the environment. The evaluation criteria for green biopharmaceutical production should include procedures along the supply chain from product design and manufacturing to transportation and scrapping. Compared to other industries, there has been relatively little research on the evaluation of green biopharmaceutical strategic alliance partners. The first multi-criteria decision-making (MCDM) framework for the bio-manufacturing industry was developed by George et al. [
3], whose evaluation criteria included earnings capacity, asset utilization, long-term solvency, productivity, and manufacturing knowledge. Shakeri and Radfar [
14] presented a comprehensive model for performance evaluation of the biopharmaceutical industry strategic alliance. They mainly surveyed the strategic performance of alliances between manufacturers and exporters of medical biotechnology products in Iran between 2000 and 2012. The model explores the relationship between several factors, including partner fit, alliance ability, capital amount, and learning ability. In recent years, strategic alliances of biopharmaceutical multinationals have also received much attention, especially in the context of cultural diversity in research and development and innovation [
15]. Unfortunately, there is still no research to establish a complete strategic alliance partner evaluation framework for the green biopharmaceutical industry.
Some advanced countries have listed the biopharmaceutical industry as one of the key development projects. Therefore, evaluating strategic alliance partners in the green biotechnology industry is an important task. The MCDM method has excellent evaluation performance in complex environments. It does not require the basic assumptions of traditional statistics, and only requires a small sample of expert interview data. The MCDM’s goal is to integrate objective survey data with expert subjective judgments and provide effective management information to support decision-makers in developing best strategies [
16]. Common methods for determining weights are the analytic hierarchy process (AHP) [
17], analytic network process (ANP) [
18], best-worst method (BWM) [
19], decision making trial and evaluation laboratory (DEMATEL) [
20], and entropy [
21]. Methods for performance integration and evaluation include technique for order preference by similarity to an ideal solution (TOPSIS) [
22], Visekriterijumska Optimizacija i Kompromisno Resenje (VIKOR) [
23], ELECTRE [
18], preference ranking organization method for enrichment evaluation (PROMETHEE) [
24], and data envelopment analysis (DEA) [
25]. MCDM methods have been widely used in the assessment and selection of various industries. Büyüközkan et al. [
26] used AHP and TOPSIS to determine the ranking of partners in the logistics value chain. The criteria for evaluation are mainly divided into the individual ability of the partner and the organizational cooperation ability of the alliance. Wang et al. [
25] proposed a selection framework for aerospace and defense alliance partners with the use of the DEA approach to predict the future operational performance of viable alliance partners. There are also some studies that use MCDM to evaluate strategic alliance partners, such as collaborative development of communities [
27], cocreation strategies for telecom operators [
28], and innovation and entrepreneurship of clean technologies [
29].
This paper proposes a strategic alliance partner evaluation framework for the green biotechnology industry, using a hybrid MCDM approach to evaluate partners’ performance. First, based on the relevant literature, and the discussion with the decision-makers of the target company is made to establish a complete evaluation criteria system, especially the environmental protection criteria. Second, the BWM method is used to obtain the weight of the criteria. The BWM method is one of the most popular weight calculation methods in the past five years. It overcomes the two shortcomings of AHP, that is, the large number of pairwise comparisons and the poor consistency. Finally, modified fuzzy TOPSIS is used to calculate the total evaluation scores of each partner. The addition of fuzzy theory overcomes the problem of information uncertainty. In addition, this paper improves the TOPSIS technique proposed by Kuo [
30] and introduces the concept of the aspiration level into the calculation process of TOPSIS, thereby avoiding having to select the best apple from a barrel of rotten apples [
31,
32,
33]. The improved TOPSIS can be used to obtain the room for improvement for each partner based on their distance from the aspiration level, so that more management information can be obtained in practical applications. Finally, this study applies data obtained from the survey of a multinational green biopharmaceutical company in Taiwan as a case study. The method can help decision-makers be more systematic in the decision-making process and the results provide more reliable suggestions for improvement of their partners.
The rest of the paper is organized as follows.
Section 2 presents the criteria for evaluation of green biopharmaceutical industry strategic alliance partners.
Section 3 describes the proposed hybrid MCDM model approach and its basic concepts.
Section 4 demonstrates the feasibility and practicality of the proposed model in a real-world application.
Section 5 summarizes the discussion of the whole study and provides future research directions.
3. The Proposed Model for Strategic Alliance Partner Evaluation
This section describes the MCDM methods used and their detailed calculation processes, including best worst method, fuzzy set theory, and fuzzy modified TOPSIS-AL technique.
3.1. The Best-Worst Method
BWM was proposed by Rezaei [
19]. Compared with AHP, the BWM questionnaire is more concise and achieves better consistency. BWM has been widely used in decision making in various industries. Rezaei et al. [
48] proposed the service quality (SERVQUAL) model to assess the service quality of the aerospace baggage handling system and investigated passengers from different nationalities. Through the analysis of BWM, it was determined that “reliability” is the most important indicator. Omrani et al. [
49] combined the BWM and MULTIMOROA methods to assess the human development index. This study demonstrates that BWM is a more efficient method of calculating weights than AHP. There are other practical applications, such as site selection [
50], supplier evaluation [
51], company performance evaluation [
52], key factors analysis for sustainable building [
53], and so on. The detailed processes of BWM obtaining weights are described as follows:
Step 1. Determine the evaluation criteria set of the decision system
Decision-makers develop a set of criteria for evaluating the strategic alliance partners.
Step 2. Select the best and worst criteria
Based on the n criteria developed in Step 1, decision-makers pick the best (i.e., most satisfied, most preferred, or most important) and the worst (i.e., least satisfied, least preferred, or least important) criteria. The best and worst criteria chosen are key factors influencing the results of the BWM analysis.
Step 3. Compare the best criterion with other criteria to generate BO (Best-to-Others) vector
Decision-makers assess the relative importance level of the best criteria and other criteria, as shown in
Table 1. The evaluation scale ranges from 1 to 9 and the BO vector is generated. Scale 1 is considered to be equally important, and scale 9 is absolutely important and belongs to the highest level of scale. It is expressed as:
where
indicates the importance level of the best criterion
B relative to the criterion
j, and the comparison between the best criterion and itself must be 1 (i.e.,
).
Step 4. Compare the worst criterion with the other criteria and generate OW (Others-to-Worst) vector
Similar to Step 3, the decision-makers evaluate the relative importance level of other criteria to the worst criterion, and then produces an OW vector, which is expressed as:
where
indicates the importance level of the remaining criterion
j relative to the worst criterion
W, and the comparison between the worst criterion and itself must be 1 (i.e.,
).
Step 5. Calculate the optimal weightsfor each criterion
The best criterion weight value is obtained by the linear programming (LP) model. The input data is BO and OW vectors (the weight ratio of the best criterion to the remaining criteria and the weight ratio of the remaining criteria to the worst criterion). We should find a solution where the maximum absolute differences
and
for all
j is minimized. Considering the non-negativity and sum condition for the weights. The complete model is expressed as follows:
In Equation (1), the objective function of the minimized maximum can be converted to a minimized objective function for calculation. The minimized objective function after conversion can be presented by the following model:
Equation (2) has the possibility to generate multiple optimal solutions. Therefore, Rezaei [
54] proposed a linear BWM model and modified the minimized objective function as:
Equation (3) is a linear problem, it only gets a single optimal solution, and the best weight value is obtained. can be regarded as a consistency indicator, and when it is close to 0, it means that it has a high degree of consistency.
3.2. The Fuzzy Modified TOPSIS-AL Technique
TOPSIS is currently one of the most effective MCDM methods for integrating performance values. The method mainly finds positive and negative ideal solutions in the alternative combinations, and determines the relative position of each alternative by calculating the distances between each alternative and the positive and negative ideal solutions. The best alternative is to be closest to the positive ideal solution and farthest away from the negative ideal solution. TOPSIS is easy to understand and operate and has been used in many problems [
55,
56,
57,
58]. Furthermore, when performing decision-making processes in an uncertain context, they are often influenced by subjective and vague judgments. Therefore, Zadeh [
59] first introduced fuzzy set theory as a soft computing method for decision ambiguity. According to the definition of fuzzy sets, expert opinions are usually described by linguistic variables. In practice, linguistic variables can be represented by fuzzy numbers, forming a set of ambiguities. The most common linguistic variable is the triangular fuzzy number (TFN), proposed by Pedrycz [
60]. This paper combines TOPSIS with fuzzy theory to reflect the inaccuracy of the practice assessment environment and to replace the relatively better solution in the existing solutions with the aspiration level. The detailed TOPSIS operation steps are described as follows.
Step 1. Define the symbol
Suppose there are
m alternatives
,
n criteria
, and the weight of the criteria is defined as
. Each expert
Dk (
k = 1, 2,…,
p) evaluates the performance of the alternative
Ai (
i = 1, 2,…,
m) according to the criterion
cj (
j = 1, 2,…,
n).
Table 2 shows the scales of performance evaluation.
Step 2. Construct an initial fuzzy decision matrix
Expert
Dk evaluates all alternatives according to the scales of
Table 2. This paper uses the arithmetic mean to summarize the evaluation values of all experts and obtains the initial evaluation fuzzy decision matrix, expressed as
where
,
and
,
k = 1, 2,…,
p.Step 3. Construct a normalized fuzzy decision matrix
The purpose of normalization is to unify the units of all evaluation criteria and to make the values within the matrix bound to 0 to 1. The normalized fuzzy matrix is
. The conventional normalization method is to take the best performance value in the alternatives as the denominator, i.e.,
This article introduces the concept of the aspiration level into this step. The modified formula is
where
xasprie = 10 (the highest level of the evaluation scales).
Step 4. Construct the weighted normalized fuzzy decision matrix
Considering the importance of each criterion, the weighted value (
wj) of the criterion evaluation is multiplied by the normalized fuzzy decision matrix
to obtain the weighted normalized fuzzy decision matrix. The calculation method is as follows.
Step 5. Define positive ideal solutions and negative ideal solutions (PIS and NIS)
Based on the concept of aspiration level, the positive and negative ideal solutions of the alternatives should be 1 and 0 after normalization. Therefore, the positive ideal solution and the negative ideal solution (
Aasprie and
Aworst) of the alternatives are calculated as follows
Step 6. Calculate the distances between each alternative and the PIS and NIS
Based on the definition of the Euclidean distance square, Equations (10) and (11) are used to calculate the separation distances between the alternative
i and the PIS and NIS. At this step, the fuzzy values have been defuzzified to be converted into crisp values.
Step 7. Calculate the closeness coefficient (CCi)
The
CCi is a reliable ranking index. According to Lo et al. [
22], the ranking index considers the distance between all alternatives and PIS and NIS, overcoming the shortcomings of the traditional TOPSIS ranking index. The formula is as follows:
Here, w+ and w− represent the weights that reflect the relative importance of the PIS and NIS in the consciousness of a decision-maker, respectively. In general, both w+ and w− are set to 0.5. The closer the CCi is to 1, the closer it is to the aspiration level. Conversely, when it is very close to −1, it means that the performance is extremely poor.
5. Discussion and Conclusions
The biopharmaceutical industry is one of the emerging high-tech industries, and advanced countries have invested huge sums of money to promote the industry. In order to improve the level of healthcare, biopharmaceutical-related products are constantly being developed. Compared to other manufacturing industries, the biopharmaceutical industry has a relatively high technical threshold, and most biopharmaceutical companies use strategic alliances to increase the competitiveness of the supply chain. According to the literature review, previous studies have rarely explored the evaluation framework of biopharmaceutical strategic alliance partners; in particular, the environmental protection criteria have not been established. This paper proposes an evaluation model for a green biopharmaceutical strategic alliance partner to bring a more complete evaluation framework and analysis method to the industry. First, through a large number of literature reviews and expert interviews, five dimensions and 25 criteria were established to establish an evaluation framework. Secondly, this paper uses BWM to obtain the criterion weight, which is an effective and reliable method for determining the weight in the MCDM problem, because it requires less pairwise comparisons and easy to obtain high consistency results. Finally, we improved TOPSIS, proposed by Kuo [
30], by introducing the concepts of fuzzy theory and aspiration level to optimize the shortcomings of TOPSIS.
According to the BWM results of
Table 7, the innovation capability (
D2) is the most important dimension based on the dimension level, with a weight value of up to 0.387. Product innovation capability is the most important competitiveness criterion for the biopharmaceutical industry. The government often uses patents as a basis for assessing the company’s potential [
43]. Zhang et al. [
7] believe that R&D capabilities and patented technologies are key factors in the survival of the biopharmaceutical industry. Because of the long cycle of drug development and high investment costs, products can easily be imitated or even replaced without the support of patented technology. Their research echoes the results of our analysis. The two most important criteria are R&D capability (
C24) and core technical patent (
C21). Organizational management (
D3) is the second most important dimension. The organizational cooperation of strategic alliances can be divided into organic coalitions, bureaucratic foundations, coalitions of intense interdependency, and reciprocal foundations. Regardless of the organizational approach, effective management mechanisms are needed to create mutually beneficial effects. Most of the strategic alliance partners hope to jointly create high-value brands and build customer loyalty through brand image to increase market share and revenue. We shared and fed back the results of the BWM analysis to all the experts, who say that this information can assist them in decision making in strategic alliances.
The proposed model provides a systematic analysis process that can completely evaluate and prioritize the partners. This study has confirmed that combining BWM with TOPSIS to analyze the strategic alliance partner problem should be an effective model. The calculation procedure proposed in this study optimizes TOPSIS. The results show that
A3 is currently the best performing strategic alliance partner.
A3 is a multinational food company with a turnover of NT
$399.861 billion in 2017. The partner has a large sales channel, and in recent years, it wants to develop a pathway for the pharmaceutical industry, but lacks biopharmaceutical technology. Therefore, it is one of the members of the research case.
Figure 2 shows that
A3 has relatively good performance compared to other partners. Based on the results of this evaluation, all partners can develop relevant improvement strategies to reach the aspiration level.
Although environmental protection (
D5) is not the most important dimension, it still has significant impact on the overall evaluation system. Since the green biopharmaceutical industry must pay attention to environmental protection, we explored whether the weight change of environmental protection (
D5) would affect the results of the overall evaluation system. Sensitivity analysis was used to verify that the partner’s prioritization wold be changed significantly. The weight value of the environmental protection (
D5) was changed from 0.1 to 0.9, and the other criteria were weighted proportionally.
Table 12 shows the ranking results of the nine test runs. Obviously, run five’s partner ranking has changed. According to
Figure 3, when environmental awareness becomes more and more important (
D5’s weight is getting higher and higher), the ranking of
A1 is getting higher and higher, indicating that
A1 environmental protection awareness is better than other partners. On the contrary,
A5 is a company that pays less attention to environmental protection. However, it is worth noting that when the weight of
D5 changes, it still does not affect the rankings of
A2 and
A3.
In summary, the proposed evaluation model provides a systematic approach, a selection and evaluation tool for strategic alliance partners in the biotech pharmaceutical production industry. This effective soft computing method can reduce the subjectivity of management decisions. To the best of our knowledge, there has been no academic study exploring strategic alliances in the green biopharmaceutical industry. Our model integrates several state-of-the-art methods and considers various real-world situations, including the consideration of message uncertainty and the introduction of the concept of the aspiration level. Our results demonstrate the validity and reliability of the proposed model. Such a model would bring several benefits to the case company. It would: (i) make it easier to identify the most appropriate strategic alliance partner; (ii) provide a basis for improvement of the strategic partners; and (iii) help decision-makers be more systematic in the decision-making process.
In addition, the results led to several new findings: (i) R&D capacity remains the most important condition for manufacturers; (ii) a biopharmaceutical production company must be supported by multiple invention patents to avoid being imitated by their competitors; (iii) sensitivity analysis reveals which partners are environmentally conscious, which will strongly influence sustainable development of the strategic alliances; (iv) organizational management is the second most important dimension for evaluation, with emphasis on the mutual assistance and mutual trust of partners in strategic alliances to jointly create an excellent brand image.
In the future, researchers can use different MCDM methods to evaluate partner performance, such as VIKOR, PROMETHEE, GRA, and DEA, etc. In addition, the quantitative data of actual enterprises can be further investigated to make the evaluation results more accurate.