Dynamic fuzzy game decision about supply chain resource integration in online shopping

Abstract

This paper proposes a method for determining the judging factors for evaluating the resource individuals based on the analysis about the supply chain resource integration (SCRI) in online shopping and from the perspective of the dynamic coordination and equilibrium between the supply and demand service capacities. Moreover, it demonstrates that this SCRI is a dynamic and coalitional game process. During this process, the changes in both the subjective game strategies of resource individuals and in the objective factors can facilitate the dynamism of SCRI. This paper puts forward a decision method by building a comprehensive evaluation function based on the rational division of the game process, and subsequently uses fuzzy membership functions to describe the changes in the subjective and objective conditions of both game sides, as well as the changes in their game psychology for profit-taking and risk aversion at different game stages. Finally, the paper presents an application case to verify the effectiveness of the SCRI decision.

Keywords

Online shopping supply chain resource integration dynamic game decision analysis

1 Introduction

Online shopping has become a necessary means of shopping conveniently [1 –3]. The essence of online shopping is to meet different customer needs by building an integrated trading platform for information, funds, and logistics [2]. In online shopping, different customized services (i.e., customized payment, shipping, and after-sale service modes) require online shopping companies to provide the corresponding service capacities by building a rational supply chain network. This supply chain is led by the online shopping company (core business) and primarily consists of suppliers, third-party logistics (3PL) companies, and other companies or entities related to capital flow (i.e., banks and third-party payment platforms). These companies are all resources of the online shopping supply chain. Therefore, the integration and allocation process of these resources based on different customized demands directly determines the service capacities of the online shopping company [1, 3].

Along with the rapid development of Internet technology and mobile Internet technology, a large number of new ideas and modes have emerged in the online shopping industry. Such phenomena present both opportunities and challenges for online shopping companies in the context of the intensified competition among companies, fierce price wars, complex and volatile industry environment, and forthcoming reshuffle or integration. An increasing number of online shopping companies have realized that competing based on extremely low prices is no longer an effective means of gaining competitive advantage; instead, meeting different customer needs should be the strategic focus. However, the growths of product categories and increase in customer scale have spurred problems caused by unreasonable and non-coordinated resource integration. For example, problems such as “warehouse explosion,” “outsoaring business integrity,” and “manipulated trading credibility,” which are all induced by unreasonable supply chain resource integration (SCRI), cannot be overlooked. Therefore, new problems will challenge the online shopping company, such as how to rationally select supply chain resources, how to improve SCRI flexibility, how to improve the collaboration benefits and reduce the risks of SCRI, and how to realize SCRI under circumstances of uncertain demand.

Generally speaking, in SCRI, how to select the supply chain co-operators (resource individuals) reasonably are a key problem for the core enterprise to consider. In SCRI under online shopping circumstance, the online shopping company should also reasonably select its supply chain co-operators. It is the key of the success or failure of the resource integration. Especially, it is most important for the online shopping company to design and construct a reasonable and effective SCRI decision method to cope with several new features of the SCRI in online shopping discussed above. Therefore, in order to fully manifest the strategic, dynamic, and flexible demand characteristics of SCRI in the online shopping mode and construct a reasonable and effective SCRI decision method, this paper starts from the capacity equilibrium between the supply and demand of online shopping and from the dominant factors for evaluating and integrating the supply chain resource individuals, introduces the game method [4, 5] into the systematic analyses on the SCRI decision problem in online shopping. This paper regards the SCRI co-operator selection and evaluation process as a coalitional and dynamic game, then introduces the characteristics of different capacity demands into the SCRI decision and uses fuzzy membership function to respectively describe the characteristics of game sub-processes and the dynamic integration parameters of the resource individuals. On the basis of these aspects, this paper develops a comprehensive integration decision method and tests the feasibility and effectiveness of the decision method using an application case.

2 Literature review

At present, online shopping has become a new research focus [1 –3]. Studies indicate that an increasing number of people gradually pay attention to the supply chain operation in the online shopping environment [2]. Different from the traditional offline supply chain operation [6], online shopping (especially in the context of the variety of products or services and customer demands) is characterized by customized service that accordingly determines the dynamism and uncertainty of customer demands. This feature highlights the critical importance of dealing with SCRI uncertainty, which is also an essential approach for avoiding and solving the excess capacities and warehouse explosion that exist in online shopping practices. In this aspect, SCRI and supply chain scheduling optimization problems in the dynamic and uncertain demand environment are important research issues, as shown in Table 1. For example, Yao [7] investigates SCRI from the perspective of the fourth-party logistics and constructs a decision-making optimization model and algorithm. Yao and Liu [8] examine dynamic optimization problem in an uncertain demand environment (mass customization environment) and develop allocation optimization model and algorithm. However, these studies principally focus on manufacturing enterprises and seldom deal with online shopping service enterprises.

At the same time, supply chain co-operators represent an important aspect of SCRI; several studies have explored the different selection methods of supply chain co-operators and their performance evaluations, as shown in Table 1. Because these methods are always limited in their use of static evaluation indexes to evaluate and select the supply chain members (co-operators), it is difficult use them to reflect the complex and dynamic cooperation and competition relationships of SCRI in the online shopping environment. Therefore, we should systematically analyze the SCRI problem in online shopping and probe a new co-operator selection method.

3 SCRI in online shopping and its demand character

Generally, the resources to be integrated by the online shopping company include supplier resource individuals, logistics resource individuals (e.g., self-support logistics or 3PL companies), and resource entities (e.g., banks or third-party payment institutions). To provide customized services to different customers through an online shopping platform (including network interaction and supply chain support platforms), the online shopping company should recognize the characteristics of different customized service demands (including predictable and uncertain demands), identify the essential requirements for different customized services (including different service modes and their combination service modes) to the supply chain service capacities, and make SCRI decisions at various times with diverse levels of dynamic equilibrium between resource supply and customer demand to effectively utilize supply chain resources, improve the integration benefits, and reduce supply chain risks.

Compared with the offline channel, online shopping has an obvious advantage in terms of meeting the diversified and individualized shopping demands of customers. On the one hand, this advantage emanates from more extensive customer coverage; additionally, it not only reflects the planning demands of multi-level customers at different times, but also effectively meets the dynamic demands of customers. On the other hand, online shopping platforms have dynamic and multi-interface characteristics that are based on information flexibility, thus facilitating the integration of different supply chain resources. In SCRI, the online shopping company should consider the capacity requirements of the general services of customers and strategically plan for the unpredictable and sudden capacity demands. In this manner, the SCRI scheme becomes more feasible and flexible in responding to “warehouse explosion” and other problems that are unique to online shopping.

In this context, the characteristics of the customized demands in online shopping should be determined. In investigating the current mainstream online shopping companies, we provide a division example of the customized service modes of the current mainstream B2C online shopping (see Table 2).

4 Service capacity analysis of online shopping

The service capacities required by different online shopping service modes and their combination service modes must be provided by various combinations of supply chain resources, as shown in Fig. 1. At different times, the online shopping company should carefully discriminate and filter the supply chain resources and seek the best integration method for different service modes. The reason is that the characteristics and strengths of the resources required by every service mode are different; at the same time, the relationships among these resources vary as well.

As pointed out in the introduction, the characteristics of online shopping increase the dynamism and uncertainty of customer demands because of the complex varieties of shopping goods, as well as the wide region of clients and their complex demands. Such aspect distinguishes online shopping from the offline shopping mode. All of these characteristics will induce several problems in online shopping and subsequently disrupt the original normal operations of shopping activities. For example, the order surges caused by peak demands during holidays or special days (i.e., anniversary, Valentine’s Day, and Singles Day) exceed the processing capacities of the supply chain and logistics, thus delaying shopping deliveries or damaging the service quality. Moreover, the insufficient performance or integrity of 3PL cooperation will adversely affect customer service satisfaction in online shopping.

Therefore, in addition to providing general service capacities to meet customer demands at different times, the online shopping company should prepare the emergent service capacities for special periods to effectively deal with the uncertainties of customer service demands and supply chain cooperation relationships. Moreover, the online shopping company should develop its strategic potential capacities based on its long-term development considerations, as illustrated in Fig. 2.

From the perspective of strategic development, the general service capacity is the essential capacity for the online shopping company to satisfy the customized demands. Hence, to provide this capacity, the online shopping company should select the size and quality of the capacity as primary indexes to integrate the supply chain resources. Meanwhile, the emergent service capacity should deal with the unexpected events in online shopping (i.e., warehouse explosion event) through advanced preparation. In addition to the capacity quality, the accurate time window of this capacity is critically important for evaluating the supply chain resources. Strategic potential capacity is principally based on the future service orientation and demand forecasting, and its uncertainty is strong; thus, the online shopping company should ensure the elasticity of this capacity when conducting SCRI (see Table 3).

The different capacity requirements should prompt the online shopping company to focus on the various capacity attributes when planning and implementing the SCRI to reduce the cooperation risks and simultaneously realize the maximum benefits.

5 Judging factor analysis of SCRI

Supply chain resources for online shopping are characterized by complex categories, wide distribution, and dynamic changes. Resource integration involves two processes. First, the main body (online shopping company) must reasonably evaluate and select the higher quality resources by determining the factors for judging whether the resource individual is fit for integration. Second, as the integration objects, the supply chain resource individuals (co-operators) will also evaluate the integration fitness through these factors, which is a typical interactive evaluation and selection process. These judging factors are indexes for evaluating and selecting the supply chain resource individuals.

Effective evaluation indexes are significant for the supply chain co-operator selection both in qualitative search and computational optimization. Although scholars systematically propose the basic principle and evaluation indexes [11 , 14–16] for co-operator selection, the characteristics of different industries and enterprises as well as the strategic positions and operational modes of enterprises will require various supply chain co-operators.

As shown in Table 2, the functions and roles provided by various service modes differ in the shopping process. For example, in the logistics service mode, Attribute 1 (door-to-door delivery), Attribute 2 (self-pickup service), and Attribute 3 (self-help mailbox in community) focus on the delivery service, and the supply chain resources that correspond with this service mode are the 3PL companies or the self-operating logistics department of the online shopping company. When integrating these logistics resources, the online shopping company will select them based on their specific features. For example, a 3PL company with a wider distribution network, a larger size of delivery staff, a better service quality, a higher efficiency, and a better information technology utilization (e.g., enhanced tracking positioning capacity) than other 3PL firms would become the preferred resource individual for the online shopping company in terms of evaluating and selecting door-to-door delivery resources.

These judging factors (i.e., distribution network, size of delivery staff, quality and efficiency of delivery, and IT utilization level) represent the objective characteristic factors of the supply chain resources individuals, and such factors are relatively stable in a particular period. Moreover, these factors do not arbitrarily change with the negotiation process between the main body and the object body of the integration. However, other factors, such as the collaborative service price (determining the integration outsourcing cost of the online shopping service), collaborative service speed and time window (identifying the outsourcing delivery time of the online shopping service), compensation level of collaborative risks (ascertaining the integration risks of the main body), and processing quality of collaborative service activities (determining the shopping service quality), can signify the game chips for bargaining in SCRI.

Therefore, we can classify the judging factors (evaluation indexes) into two categories based on their roles in changing the game psychology of rivals in the SCRI consultation process, as shown in Table 4. One type of factors [denoted by f_xx(Y) in Table 4] has no role in influencing the psychology of the game opponent. Such factors include business environment, corporate culture, organizational structure, personnel quality, management level, service capacity, and IT level. These factors are relatively stable during a certain period, and they cannot change along with the subjective game strategies. By contrast, other types of factors can often be used as game chips in SCRI, such as service quality, price, delivery time, risk compensation, and others [f_xx(N) in Table 4].

At the same time, the online shopping company will distinguish the supply chain resources aimed at providing different service capacities. Therefore, even to the same service mode attribute, the online shopping company will give different levels of attention to mining the judging factors based on different service capacities. For example, the online shopping company will give more attention to reducing the logistics service cost and improving the logistics service punctuality aimed at the resource integration for general service capacity when selecting a 3PL company to realize the door-to-door delivery service (attribute) of the logistics service mode. By contrast, the online shopping company will give more attention to the usability and flexibility of the resources during the period of emergent service time aimed at its emergent service capacity integration.

Therefore, in the game process of SCRI, the selection of evaluation indexes and their weights set aimed at different service capacities is the most important undertaking for the online shopping company, which must be reflected in the negotiation process of SCRI. In the actual operation, mining the judging factors shown in Table 4 must be based on detailed market surveys and studies on the various situations of supply chain co-operators, together with their strategic positions and operations. The rationality of mining judging factors and setting their weights will directly determine the success or failure of SCRI, and this issue is a critical strategic problem.

6 Dynamic game process of SCRI

Based on the process of mining judging factors, the online shopping company can evaluate and select resource individuals aimed at different service capacity demands. However, SCRI in the online shopping mode is more complex compared with the static selection process that is based on the same judging factors. Two principles underlie this phenomenon. First, this SCRI decision involves not only the evaluation by the integration main body (online shopping company) of its future integration benefits and risks, but also the assessment of each resource individual on the integration benefits and SCRI participation risks that will ultimately determine the participation decision. Second, from a dialectical perspective, any process can be divided into several different stages (i.e., beginning, intermediate, and ending stages). The action psychology of participants constantly differs at every stage. In the SCRI negotiation with the main body, each individual may employ different strategies at various game stages (i.e., adjusting the information on service price, time window, service quality, service capacity, and other factors) and adopt them as game behaviors to guide and contain the SCRI main body. This process forms the dynamic characteristic of the SCRI game in online shopping and fully reflects the real profit-seeking psychology of both collaborative sides.

The game strategies of the integration main body and other resource individuals curtail the complete and adequate information sharing between these two sides at different game stages. Therefore, this process becomes inconsistent with the reality if the main body only conducts integration based on a single static evaluation function. To resolve this problem, this paper introduces dynamic game theory into the SCRI decision in the online shopping mode to reflect the subjective profit-seeking psychology changes of both game sides at different game stages. The paper subsequently uses a fuzzy membership function to describe the changes in these game strategies with different factors at various game stages to quantitatively analyze the problem.

For example, Fig. 3 illustrates the changes in the unit service price strategy of some resource individuals and their service time window strategy. The higher the membership degree is, the lower the unit price is and the higher the satisfaction of the clients to the service time window becomes. Thus, in the intermediate game stage, the resource individual offers an advantageous service price but a contrary strategy in service time window, and then explores the real attention of the main body to both judging factors in SCRI. Finally, in the ending stage, the resource individual adopts a strategy of slightly increasing the price under the conditions of the best service time window.

7 Dynamic game decision of SCRI

We classify the supply chain resources for integration based on the customized service mode attribute that is shown in Table 2. We denote a_nm as each type of resource individual, L_nm as the number of individuals in one type, and s_nml (l = 1, 2, …, L_nm) as the index of each individual.

Moreover,

t = number of dynamic game stages

F_1u(Y)(nml) (t)= fuzzy membership degree of index u for evaluating individual s_nml to realize the main body’s general service capacities demanded for its service attributes a_nm

F_2u(Y)(nml) (t)= fuzzy membership degree of index u for evaluating individual s_nml to realize the main body’s emergent service capacities demanded for its service attributes a_nm

F_3u(Y)(nml) (t)= fuzzy membership degree of index u for evaluating individual s_nml to realize the main body’s strategic potential capacities demanded for its service attributes a_nm

w_1u(nm)= weight of evaluation index u set by the main body aimed at integrating the general service capacities demanded for its service attributes a_nm

w_2u(nm)= weight of evaluation index u set by the main body aimed at integrating the emergent service capacities demanded for its service attributes a_nm

w_3u(nm)= weight of evaluation index u set by the main body aimed at integrating the strategic potential capacities demanded for its service attributes a_nm

D = total sub-process number of SCRI games divided by the integration main body, and we denote d (d = 1, 2, …, D) as its index

μ_d (t)= fuzzy membership degree of stage t in sub-process d divided by the main body

w_1(nm), w_2(nm), and w_3(nm)= respectively the weights of general service capacity, emergent service capacity, and strategic potential capacity demanded for the main body under its different a_nm.

The following two situations are relevant in this case:

(1) Without consideration for the sub-process strategy, we denote the comprehensive evaluation function as Formula (1).

$\begin{matrix} E_{s_{nml}} (t) & = & \sum_{u = 1}^{U_{nm}} [w_{1 (nm)} w_{1 u (nm)} F_{1 u (Y) (nml)} (t) \\ + w_{2 (nm)} w_{2 u (nm)} F_{2 u (Y) (nml)} (t) \\ + w_{3 (nm)} w_{3 u (nm)} F_{3 u (Y) (nml)} (t)] \end{matrix}$ (1)

(2) With consideration for the game sub-process strategy, we denote the comprehensive evaluation function as Formula (2).

$\begin{matrix} E_{s_{nml}}^{*} (t) & = & \sum_{d = 1}^{D} {μ_{d} (t) \sum_{u = 1}^{U_{nm}} [w_{1 (nm)} w_{1 u (nm)} F_{1 u (Y) (nml)} (t) \\ + w_{2 (nm)} w_{2 u (nm)} F_{2 u (Y) (nml)} (t) \\ + w_{3 (nm)} w_{3 u (nm)} F_{3 u (Y) (nml)} (t)]} \\ = & \sum_{d = 1}^{D} [μ_{d} (t) E_{s_{nml}} (t)] \end{matrix}$ (2)

To obtain the final evaluation results, we calculate the comprehensive evaluation value of different resource individuals using Formula (1) or (2) at each game stage.

To ensure that the integration results are close to the actual situation, we use Formula (3) in calculating the comprehensive evaluation value of resource individual s_nml in a dynamic SCRI game decision.

$\begin{matrix} G_{s_{nml}} & = & \sum_{t = 1}^{T} E_{s_{nml}} (t) / T \\ or G_{s_{nml}}^{*} & = & \sum_{t = 1}^{T} E_{s_{nml}}^{*} (t) / T \end{matrix}$ (3)

8 Case analysis

Online shopping enterprise DD seeks to expand its logistics service mode of door-to-door delivery through supply chain resource integration on the basis of its strategic positioning. Assume that four 3PL companies (3PL1, 3PL2, 3PL3, and 3PL4) meet the basic demands of the integration of DD. The rationality of SCRI will directly determine the online shopping experience value of customers; hence, DD should make a better integration game decision. DD ultimately adopts the dynamic game negotiation strategy in integrating these 3PL resources to enhance the expected benefits and reduce the collaboration risks.

Prior to integration, DD should identify the evaluation indexes (judging factors) for resource individual selection. Integration aims to meet both the normal distribution service requirements and the emergent delivery requirements on the coming holidays or the days of sales promotion. It also seeks to fulfill the additional capacities enhancement required for future extended delivery services based on the strategic planning of the company. Thus, DD determines the evaluation indexes and their related weights in Table 5 based on the analysis of its internal and external environments, market situations, and strategic locations.

Assume that DD determines that the dynamic SCRI game process includes five stages based on its historical game negotiation experiences. The fuzzy membership function relationships of the game strategies of every 3PL company that is aimed at different service capacities at various game stages are shown in Fig. 4. The real game data of every 3PL company at different game stages are presented in Table 6.

Moreover, assume that the 3PL companies often truly express their desires in the intermediate and ending game stages according to the characteristics of resource integration, subjective and objective factors as well as the historical experiences in their integrations. Thus, the sub-process game strategies that DD adopts are depicted in Fig. 5. We can obtain the comprehensive evaluation data in Table 7 using Formulas (1) to (3) to calculate the data of Table 6 and Fig. 5.

When we do not consider the sub-process strategy of DD, according to the data in Table 7, we can obtain G_3PL1= 0.815, G_3PL2= 0.817, G_3PL3= 0.816, and G_3PL4= 0.829. Thus, DD should select 3PL4 to achieve integration. Practice shows that this result is consistent with the actual situation; the 3PL company strategies in Fig. 5 indicate that the performances of the related indexes of 3PL4 are more advantageous at each stage in the SCRI game process.

Meanwhile, when we consider the sub-process strategy of the integration main body DD, according to the data in Table 7, we can calculate that G*_3PL1 = 0.553, G*_3PL2 = 0.551, G*_3PL3 = 0.561, and G*_3PL4 = 0.559; thus DD should select 3PL3 to achieve integration. DD considers that 3PL companies often truly express their actual situations in the intermediate and ending game stages; hence, the beginning stage of the game is unimportant, and 3PL3 exhibits better performances of the indexes at these two game stages.

We can learn from the above decision results that in the preceding case analysis, every index has a reasonable sensitivity in the integration process, which fully reflects the subjective and objective decision situations of both integration sides, as well as their mastery of the degree of the potential benefits and risks to make a more comprehensive and systematic integration decision. Compared with this method, the other general methods of the supply chain co-operators selection (e.g. the references [12 –16]) always take the selection main body (core enterprise) as the dominant and realize the co-operators selection through constructing the static or dynamic evaluation index system as the basis for measuring and judging the co-operators. For these methods do not consider the subjective benefits and risk preferences of different co-operators (selection objects) under different collaboration environments, it is hard to reflect the realistic game situations of the two selection sides from the objective point of view so that the results of these co-operators selection methods will have certain discounts.

9 Conclusions

To provide a systematic and quantified research on SCRI in online shopping based on the analysis of its features, this paper indicates that the integration process is a coalitional game process between two collaborative sides. During this process, the changes both in the subjective game strategies and in the objective factors of the resource individuals tend to induce dynamic characteristics. To address this problem, through the rational division of the game process, this paper describes the characteristics of game sub-processes and the changes in the integration indexes of individuals using fuzzy membership functions. This paper proposes a comprehensive evaluation decision method using dynamic game stages as variables, which truly reflects the subjective and objective situations of both sides and their negotiation psychology for profit-seeking and risk aversion in the flexible integration decision process.

Game stages can be determined on the basis of the nature and characteristics of the customized service demands by the online shopping company. Thus, such game stages are not open to the resource individuals beforehand, thus ensuring the objectivity of the integration decision. At the same time, the sub-process division of the game has considerable flexibility that can be mastered freely by the integration main body, which is of substantial strategic significance in the SCRI process in online shopping.

In short, the major differences between the dynamic fuzzy game decision method about SCRI in online shopping discussed in this paper and the other methods of the supply chain co-operators selection mainly reflect in the following three aspects.

The method discussed in this paper can obtain different decision-making results according to the different game preferences of the resource integration main body (supply chain core enterprise). It is different from the other co-operators selection methods which always give non changeable decision results under a single hypothesis.

The method discussed in this paper can easily reflect the subjective benefits and risk preferences of different co-operators (selection objects) under different collaboration environments. It is more comprehensive than other methods which always make the co-operators selection decisions standing in the perspective of the main body.

Especially for the SCRI decision problem in online shopping mode, the integration requirements of the supply chain co-operators have obviously strategic, flexible and dynamic features. Therefore, the method discussed in this paper has obvious benefits to match this special decision problem.

Footnotes

Acknowledgments

This research is financially supported by the National Nature Science Foundation of China (Project Nos. 71472183). The author would also like to thank the anonymous reviewers and editors.

References

Boyer

and Hult

, Extending the supply chain: Integrating operations and marketing in the online grocery industry, Journal of Operations Management23 (2005), 642–661.

and Wang

, Research on inventory and transportation integrated optimization model of supply chain on online shopping based on the revenue sharing contract, Journal of Chemical & Pharmaceutical Research6 (2014), 131–136.

Çelik

, Influence of social norms, perceived playfulness and online shopping anxiety on customers’ adoption of online retail shopping, International Journal of Retail & Distribution Management39 (2011), 390–413.

Mikhail

and Erwin

, A game mechanism for single machine sequencing with zero risk, Omega44 (2014), 104–110.

Martínez

and Simchi

, Supplier-buyer negotiation games: Equilibrium conditions and supply chain efficiency, Production & Operations Management22 (2013), 397–409.

Chen

, Fang

and Wena

, Pricing policies for substitutable products in a supply chain with internet and traditional channels, European Journal of Operational Research224 (2013), 542–551.

Yao

, Decision optimization analysis on supply chain resources integration in the fourth party logistics, Journal of Manufacturing Systems29 (2010), 121–129.

Yao

and Liu

, Optimization analysis of supply chain scheduling in mass customization, International Journal of Production Economics117 (2009), 197–211.

Yao

, Scheduling optimization of co-operator selection and task allocation in mass customization supply chain based on collaborative benefits and risks, International Journal of Production Research51 (2013), 2219–2239.

10.

Yao

, Supply chain scheduling optimization in mass customization based on dynamic profit preference and application case study, Production Planning & Control22 (2011), 690–707.

11.

Şenyiğit

, Supplier selection and purchase problem for multi-echelon defective supply chain system with stochastic demand, Neural Computing & Applications22 (2013), 403–415.

12.

Chan

, Kumar

and Tiwari

, Global supplier selection: A fuzzy-AHP approach, International Journal of Production Research46 (2008), 3825–3857.

13.

Samvedi

, Jain

and Chan

, Quantifying risks in a supply chain through integration of fuzzy AHP and fuzzy TOPSIS, International Journal of Production Research51 (2013), 2433–2442.

14.

, Liu

and Chen

, Supplier selection using axiomatic fuzzy set and TOPSIS methodology in supply chain management, Fuzzy Optimization & Decision Making11 (2012), 147–176.

15.

Liao

and Kao

, An integrated fuzzy TOPSIS and MCGP approach to supplier selection in supply chain management, Expert Systems with Applications38 (2011), 10803–10811.

16.

and Blackhurst

, Supplier evaluation and selection: An augmented DEA approach, International Journal of Production Research47 (2009), 4593–4608.

17.

Chen

and Yan

, Network DEA model for supply chain performance evaluation, European Journal of Operational Research213 (2011), 147–155.

18.

Sharif

A.M.

, Alshawi

and Kamal

M.M.

, Exploring the role of supplier relationship management for sustainable operations: An OR perspective, Journal of the Operational Research Society65 (2014), 963–978.