Abstract
The initial crisis early warning index system of the supply chain quality has been built up according to crisis inducement, documentary research and corporate research. Then we evaluated the feasibility, importance, independence of the index system by using experts scoring methodology. The initial crisis early warning index system has been filtered by fuzzy inference system (FIS) and the final crisis early warning index system has been established.
Introduction
The quality of supply chain reflects the degree of supply chain system to meet customer demand in terms of product logistics, information flow and capital flow [1]. If problems occur in any part of logistics, information flow and capital flow, the supply chain can be broken. Therefore, it is of great theoretical and practical significance to study the crisis early warning management of supply chain quality. In the case of crisis early warning, the construction of crisis early warning index system is the first step. With a scientific and reasonable index system, we can lay a good foundation for future research. The supply chain is a large and complex system with many variables and complex features. For comprehensively, objectively, effectively and accurately assessing the crisis situation of supply chain quality, a scientific crisis warning index system which can fully reflect the characteristics of the supply chain quality crisis is necessary.
Establishe initial crisis early warning index system of supply chain quality
Indicators extracted from the literature
The literature on supply chain quality, supply chain robustness, supply chain crisis, supply chain risk and supply chain performance evaluation is searched through databases such as zhiwang, wanfang, Elsevier, Emeral, IEL. The relevant literature has demonstrated the influence factors such as supply chain operation and supply chain risk, and extracted the evaluation indexes related to the supply chain quality, and sorted out these indicators. As for the expression meaning, the combination of different statements, such as the same meaning of the profitability of investments and the return on assets, will be unified as the return on assets when the index is extracted.
The indicators obtained from enterprise research
This part obtain the corresponding evaluation index through the enterprise recognizes the crisis inducement of the supply chain quality. Limited conditions, this paper presents an in depth investigation and research on nice core enterprises of supply chain in fuzhou and xiamen, including 3 textile enterprises, 3 automobile manufacturing enterprises and 3 logistics enterprises. These supply chains, where the enterprises are, are known locally, and have set up supply chain crisis early warning system. The research adopt interview method, interview object is the management of the company and the main business manager.
Based on previous information on the crisis early warning of supply chain quality, an interview outline for the incentive identification of supply chain quality crisis is established. According to the outline, ask questions and learn more about the situation. These common indicators extracted from the nine supply chain quality crisis inducement information.
Establishe initial crisis early warning index system of the supply chain quality
The crisis early warning indexs extracted from the above two steps are sorted out, and the indexs that are obviously repeated or contained are merged. They are classified according to the supply chain quality crisis inducements, which is the initial index system, see Table 1.
The initial crisis early warning index system of the supply chain quality
The initial crisis early warning index system of the supply chain quality
Fuzzy Inference System (FIS) is also known as Fuzzy System, which is based on Fuzzy set theory and Fuzzy Inference method, and has the ability to deal with Fuzzy information. The description and characterization of fuzzy systems are based on natural language, and the rules of inference are more closely related to human thought habits. It doesn’t require an exact mathematical model, and not sensitive to the change of object parameters, it has strong robustness. In terms of function, FIS is mainly composed of fuzziness, fuzzy rule base, fuzzy reasoning and defuzzification. Its basic structure is shown in Fig. 1, and its working process is as follows.
Supply chain quality is a fuzzy concept, so the crisis early warning index system of supply chain quality can be regarded as a fuzzy set, the each indicator as one of the elements, its membership is analyzed. In this paper, the index is analyzed by Matlab fuzzy toolbox.

The basic structure of FIS.
To enhance the rationality and scientificity of the index, the expert selection method is adopted to perfect the index. Experts evaluate the feasibility, importance and independence of the indicators, then select the indexes according to the membership degree obtained by the fuzzy inference system.
Expert choice
In order to make the expert data more objective and comprehensive, this study invited 10 experts to investigate, of which five from the university supply chain management professional professor, the other five is selected from the previous research enterprises in the high-level managers. Surveyed by questionnaires, and received 10 ratings.
Questionnaire design and data collection
The purpose of this questionnaire is to allow the experts to score the initial warning index system of the supply chain quality, evaluate the feasibility, importance and independence of the indicators. “Feasibility” is the degree of feasibility of the indicators used to evaluate the supply chain quality and the degree of “universality” in different supply chains. “Importance” is whether the indicators are important in crisis early warning index system of supply chain quality, and the degree of effectiveness of its assessment and its warning purposes. “Independence” refers to the degree of duplication or inclusion between indicators, and the need for indicators to exist.
The questionnaire take the initial index system as the problem item. Experts evaluate the index in terms of “feasibility”, “importance” and “independence”. The single score has 1–5 grades, representing the degree of strength of each property. After the questionnaires collected, the arithmetic mean of each indicator in different properties is calculated, which is the score of the index on that attribute.
Index selection based on FIS
The determination of fuzzy rules
In order to use fuzzy reasoning to analyze the selection of indexes, a fuzzy inference system is needed first. The system input is the feasibility, independence and importance of the index. Output is the discard, discussion and retention of the index, and its corresponding output value is [1, 2], (2, 3.5], and (3.5, 5) respectively. FIS fuzzy rules are designed in this paper as Table 2.
FIS rules
FIS rules
This section uses the Matlab fuzzy toolbox graphical user interface (GUI) to establish the fuzzy inference system to obtain the membership of each index. The process is as follows.
I/O Settings
Enter “fuzzy” in Matlab environment, call FIS editor, and defines three input variables: “feasibility”, “independence”, “importance”, and define an output “output1”.
Edit membership functions
The membership function of this paper adopts gaussian type membership function. Gauss membership function method is a kind of fuzzy method. This kind of fuzzy method has good anti-interference ability, and the fuzzy results are more closely related to the cognitive characteristics of people.
Set the input range of the input variable “feasibility” to [1 5], the membership functions are named as “not feasibility”, “moderate feasibility”and “feasibility”, and set up corresponding function parameters: [0.68 1], [0.68 3] and [0.68 5]. The membership function of the variable feasibility is shown in Fig. 2. Similarly, we can set up the membership function of variable “independence” and “importance” by the above steps.

The membership function diagram of the variable feasibility.
The parameters of “discard”, “discussion” and “retention” of the variable “output1” are set as [0.68 1], [0.5 2.75] and [1 5].
Set the parameter according to the fuzzy rules designed in Table 2.
Screen initial index system based on FIS
According to the data of each index sorted out, the value of the membership degree of each indicator is calculated by using Matlab tool, as shown in Table 3.
As can be seen from Table 3, the membership of number 10, 11, 13, 14, 20, 28, 29, 33, 34, 35, 42, 47, 52, 54, 55, 56, 57, 60, 61, 65, 67, 69, 71, 73, 71, 71, 79, 81, 83, 96, 103, 105 is equal to 2 or less, which is eliminated directly. For the index of membership in (2, 3.5], we discuss the elimination, merger and retention.
Index membership
Index membership
The membership degree of “current asset turnover” is 3.45. The index is closely related to “turnover of inventory” and “turnover of account receivable”, which can be reflected by these two indexes. So after discussion decide to abandon it. “Information communication among node enterprises” and “information sharing level among node enterprises” can be reflected by “the timely and accurately information transfer rate among node enterprises” and “the frequency of informational communication among node enterprises”, delete them. “Transport delay rate among node enterprises”, ”the damage rate of the goods” and “risk in transit” is due to transportation control is lax, After discussion, the three indicators are merged into “transport control errors”. “The delivery error rate” and “return handling error rate” are all due to the problem of the order management, merged into “order disposal failure”. The “key suppliers’s bankruptcy and losses” is 3.50 but experts believe that the index has a huge impact on the crisis of supply chain quality, so keep it.
After the screening of the fuzzy inference system, the effect of the supply chain quality crisis warning index system and the warning index is shown in Table 4.
The crisis early warning index system of the supply chain quality
The crisis early warning index system of the supply chain quality
The competition of the new century is the competition between the supply chain, and the quality of the supply chain directly affects the success or failure of the members of the supply chain. Therefore, supply chain members must establish a supply chain quality crisis warning system to grasp the crisis situation of the supply chain quality in order to take timely measures to eliminate potential crises.
The key of the early warning model is the selection of indicators. Based on a large number of related literatures and enterprises, this paper establishes an initial warning index system. Questionnaire is designed according to these initial warning indicators. Let the experts evaluate the feasibility, importance and independence of these indicators. Then apply FIS for screening to get 37 indicators, constituting the index system of the crisis early warning model. The model does not determine the weight of each indicator and the range of safety values of each indicator, which is the focus of future research.
Footnotes
Acknowledgments
This paper is supported by Education and Scientific Research Base Special Project of the 13th Five-Year Plan of Fujian Provincial Science of Education (No.FJJKCG18-109).
