CPT-TODIMS Sort new energy automobile supplier classification method based on interval intuitionistic fuzziness

Abstract

The invention discloses a new energy automobile supplier classification method based on an interval intuitionistic fuzzy CPT-TODISort method, and the method specifically comprises the following steps: firstly, defining a problem target, an alternative scheme supplier set, an evaluation set of a decision maker, the definition of a supplier classification type, and the boundary of the classification type; thirdly, constructing a decision matrix; then, an original attribute weight vector is obtained through an IVIF entropy weight method, and the weight proportion of alternative scheme suppliers is calculated; secondly, calculating a global foreground dominance degree of a supplier; and finally, according to the overall dominance of each alternative scheme and the category boundary, allocating alternative scheme suppliers to different categories. According to the method, the subjectivity that a decision maker directly gives attribute weight information in a traditional TODIM method is eliminated, different influences of the psychological behaviors of the decision maker on the final classification result in the sorting process are visually displayed, and the importance of considering the psychological behaviors of the decision maker in the classification process is verified.

Description

Technical Field

[0001] The present invention relates to the field of classification of new energy vehicle parts suppliers, and in particular to a classification method of new energy vehicle suppliers based on interval intuitionistic fuzzy CPT-TODIMSort. Background Art

[0002] As global environmental pollution, ecological degradation and climate warming become increasingly serious, the Chinese government is vigorously supporting and promoting the development of new energy vehicles. Therefore, the rapid market share growth and huge market potential of domestic new energy vehicles in recent years have also driven the development of the new energy industry supply chain. Compared with the traditional automobile industry, the new energy vehicle industry has a lower threshold, and the new energy vehicle supply chain is highly vulnerable and the basic characteristics of terminal demand are difficult to predict. This makes new energy vehicle companies more dependent on upstream supply chain suppliers, and also tests the resilience of the supply chain, that is, the stability of supplier delivery. Therefore, the new energy vehicle industry urgently needs a more resilient supply chain system at the macro level, and needs to choose more resilient parts suppliers at the micro level.

[0003] The current application of multi-criteria decision-making methods in the selection of automotive parts suppliers ignores the psychological and logical behavior of decision makers in the decision-making process, and the psychological and logical behavior of decision makers can directly affect the final supplier selection results. The multi-criteria decision-making method used in this paper for the classification of new energy vehicle parts suppliers is CPT-TODIMSort, which is obtained by improving the traditional TODIM method. Due to the limitations of the traditional TODIM method, it does not provide an appropriate method to determine the weight of the indicators. Therefore, decision makers are still required to make subjective evaluations of the various indicators of automotive parts suppliers, which leads to the subjectivity of the final sorting or classification results. Therefore, the attribute weights of automotive parts suppliers are transformed in combination with the cumulative prospect theory, eliminating the adverse effects of the attribute weight information directly given by the decision maker in the traditional TODIM method on the final result, thereby making the decision in reality more reasonable. Moreover, this method retains the original advantages of the TODIM method. By considering the psychological and logical behavior of decision makers, it intuitively shows the different effects of the psychological behavior of decision makers on the final sorting and classification results in the sorting process of new energy vehicle parts suppliers. Summary of the invention

[0004] An improved TODIMSort method based on CPT is used to solve the supplier classification problem of multi-attribute decision-making in interval-valued intuitive fuzzy environment. The CPT-TODIMSort method consists of five stages, including problem definition, constructing decision matrix, obtaining weight ratio, obtaining alternative rankings, and assigning categories.

[0005] S1: Problem Definition

[0006] First, define the goal of the problem, and the set of suppliers of alternative solutions is A n ={A1,A2,...,A n}(n=1,2,...,n), the index set is C m ={C1,C2,...,C m}(m=1,2,...,m), and alternative solution supplier A k About Indicator C n Evaluation of kn The decision maker set is I l ={I1,I2,...,I h}(l=1,2,...,h), at the same time, we assume that the weight vector of these h decision makers is ω=(ω1,ω2,...,ω h ) T ,ω l ≥0, But the weights of the attributes are completely unknown.

[0007] Next, the supplier classification categories are defined. Categories z (z=1, 2, ..., Z) are defined along with their numbers, which are arranged in order from most preferred to least preferred, and are labeled (eg, excellent, very good, or good).

[0008] Finally, define the boundaries of each supplier classification category. Figure 1 As shown, the classes can be defined by the limit boundaries. For the limit boundaries, it consists of the local limit boundaries lp zn Composition, that is, each standard C belonging to class z n The minimum level required is Figure 1 It can be seen that if there are Z categories, then Z-1 limit bounds are required.

[0009] This stage also requires collecting preference information on category boundaries, which is the standard and basis for the final classification. The extreme category boundaries shown in Table 1 are the parameters provided by decision makers, and this information will be used in the next stage.

[0010] Table 1 Standards for limit category boundaries

[0011]

[0012] S2: Constructing a decision matrix

[0013] Suppose the decision matrix in the interval-valued intuitionistic fuzzy environment given by the lth decision maker is in and They represent the membership and non-membership of the lth decision maker to the g evaluation indicators of r suppliers. and satisfy and

[0014] S2-1: Use formula (1) to convert cost-based indicators into benefit-based indicators and form a matrix.

[0015]

[0016] S2-2: Use the IVIFWA operator to combine all matrices from different experts Integrate into a group decision matrix

[0017]

[0018] S3: Get weight ratio

[0019] S3-1: Apply IVIF entropy weight method to obtain the original attribute weight vector υ=(υ1,υ2,...,υ m ),υ g ≥0(g=1,2,...,m).

[0020]

[0021] S3-2: According to formula (5), for scheme A r and A k Based on the numerical comparison of (r, k∈n and r≠k), the modified potential energy weight is obtained, where δ and τ are parameters describing the curvature of the weighting function. Indicates supplier A r Relative to A k In C g The benefits on Indicates supplier A r Relative to A k In C g δ and τ represent the loss aversion coefficient, δ, τ>0. When δ and τ are smaller (0<δ, τ<1), the risk aversion is higher; when δ and τ are larger (δ, τ>1), the risk aversion is lower, and δ, τ=1 means risk neutrality.

[77] If δ and τ are large enough, it means that the decision maker is risk-seeking and all losses can be ignored.

[0022]

[0023] S3-3: Calculate supplier A according to formula (6) r With A k The weight ratio

[0024]

[0025] S4: Obtain ranking of alternatives

[0026] S4-1: Use formula (7) to determine supplier A r With A k In attribute C g Distance D under (ρ=2) rkg .

[0027]

[0028] S4-2: Determine supplier A using formula (8) r With A k Compared to the ratio of foreground advantages in the g-th attribute.

[0029]

[0030] S4-3: Calculate supplier A according to formulas (9) and (10) r The local foreground dominance Λ(A r )(r=1,2,...,n) and the global foreground dominance φ(A r )(r=1,2,...,n).

[0031]

[0032] S4-4: Calculate the limit category limit lp according to formula (11) zn The global prospect dominance φ of .

[0033]

[0034] S5: Category Classification

[0035] Based on the overall superiority of each alternative and the category boundaries obtained in the previous section, the alternative suppliers can be assigned to different categories, which can be sorted by using the extreme category boundary diagram, such as Figure 2 shown.

[0036] In the extreme case, supplier A k Will be assigned to class z according to the following rules:

[0037] If φ(lp1)≤φ(A k )≤1, then A k Belongs to Category 1;

[0038] If φ(lp z+1 )≤φ(A k )<φ(lp z), then A k Belongs to the z+1th class, where z=1,2,.,Z-2;

[0039] If 0≤φ(A k )<φ(lp z-1 ), then A k Belongs to category Z.

[0040] Therefore, the above rules can be used to classify the suppliers of alternative solutions into corresponding categories using limit boundaries. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0042] Figure 1 Define examples for categories of limit boundaries;

[0043] Figure 2 is the limit category boundary diagram;

[0044] Figure 3 It is a flow chart of the method in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. The technical solution of the present invention is further introduced below in conjunction with the specific embodiments and the accompanying drawings.

[0046] Step 1: Problem Definition

[0047] In this case, the goal of the parts supplier classification problem is to classify candidate suppliers into different levels based on the evaluation indicators of new energy vehicle parts suppliers. We have determined 6 evaluation indicators based on a large number of relevant supplier classification literature and the current status of the new energy industry supply chain, and evaluated the 10 candidate parts suppliers selected by this new energy vehicle company. These 6 evaluation indicators are C1 (quality assurance level), C2 (cost control efficiency), C3 (delivery punctuality), C4 (supply security), C5 (information sharing level), and C6 (development and comprehensive capabilities). Among them, all six indicators are benefit-oriented indicators. Secondly, we invited 5 experts in the new energy vehicle field to evaluate these 10 suppliers from the above 6 dimensions. The weight vectors of these 5 decision makers are ω = (0.2, 0.2, 0.2, 0.2, 0.2) T .

[0048] Regarding the classification of these 10 suppliers, we defined four categories, and the corresponding categories are as follows: the first category is high-quality and low-risk suppliers, the second category is high-quality medium-risk suppliers, the third category is medium-quality medium-risk suppliers, and the fourth category is low-quality and high-risk suppliers.

[0049] The evaluation results of the first expert are shown in Table 2, and the evaluation results of the second to fifth experts are shown in Tables 3 to 6.

[0050] Table 2 Decision matrix obtained by the first expert

[0051]

[0052] Table 3 Decision matrix obtained by the second expert

[0053]

[0054] Table 4 Decision matrix obtained by the third expert

[0055]

[0056] Table 5 Decision matrix obtained by the fourth expert

[0057]

[0058]

[0059] Table 6 Decision matrix obtained by the fifth expert

[0060]

[0061] Secondly, the preference information provided by decision makers was collected. The decision weights of the five experts were all 0.2, and the category boundaries of each category are shown in Table 7.

[0062] Table 7 Limits of each category

[0063]

[0064] Step 2: Construct the decision matrix

[0065] Step 2-1: C1 (quality assurance level), C2 (cost control efficiency), C3 (delivery punctuality), C4 (supply security), C5 (information sharing level), C6 (development and comprehensive capabilities) are all benefit-oriented indicators;

[0066] Step 2-2: According to formula (2), use the IVIFWA operator to convert all matrices from different experts Integrate into a group decision matrix

[0067]

[0068] Step 3: Get the weight ratio

[0069] Step 3-1: Using formula (3) and formula (4), we can get the original index weight vector υ=(υ1,υ2,υ3,υ4,υ5,υ6) T .

[0070] Table 8 Original index weight vector

[0071]

[0072] Step 3-2: Use formula (5) to calculate alternative A r and A k Based on the numerical comparison, the modified potential energy weight Ψ is obtained rkg ( g ). (δ is set to 0.61, τ is set to 0.69), as shown in Table 9.

[0073] Table 9 Modified potential energy weight Ψ rkg ( g )

[0074]

[0075]

[0076] Step 3-3: Use formula (6) to calculate solution A r With A k The proportional weight Ψ * rkg ( g )(r,k=1,2,3,…,10), as shown in Table 10.

[0077] Table 10 Proportional weight Ψ * rkg ( g )

[0078]

[0079]

[0080] Step 4: Get the ranking of alternatives

[0081] Step 4-1: Use formula (7) (Note: D rk =D kr ,r≠k;D rk =0, r=k) Determine alternative plan A r With A k In attribute C g The distance D rkg , as shown in Table 11.

[0082] Table 11 The distance between each two alternatives

[0083]

[0084]

[0085] Step 4-2: Determine solution A using formula (8) r With A k Prospect Advantage Ratio Compared The results are shown in Table 1 to Table 12 ( ξ=0.88,σ=2.25).

[0086] Table 12 Prospect advantage ratios between alternatives

[0087]

[0088]

[0089] Step 4-3: Use formulas (9)-(10) to calculate alternative plan A r The local foreground dominance Λ(A r ) and the global prospect dominance φ(A r )(r=1,2,3,…,10), local foreground dominance Λ(A r ) See Table 13, the global foreground dominance φ(A r )See Table 14.

[0090] Table 13 Alternative A r Partial prospect dominance

[0091]

[0092] Table 14 Alternative A r The global prospect dominance

[0093]

[0094]

[0095] For Alternative A rThe global prospect dominance φ(A r )(r=1,2,3,…,10) to sort and obtain the corresponding sorting results of alternative options.

[0096] φ(A6)>φ(A7)>φ(A9)>φ(A8)>φ(A4)>φ(A 10 )>φ(A3)>φ(A1)>φ(A2)>φ(A5)

[0097] Therefore, A6>A7>A9>A8>A4>A 10 >A3>A1>A2>A5.

[0098] Step 4-4: Calculate the limit class limit lp according to formula (24) zn The global prospect dominance φ(lp zn ), see Table 15.

[0099] Table 15 Limit category boundaries lp zn The global prospect dominance

[0100]

[0101] Step 5: Category Assignment

[0102] according to Figure 2 and the classification rule of step 5 of the invention, φ(lp1)≤φ(A r )≤1, since lp1=0.886246771, the global prospect advantage of the alternative φ(A r ) meets this inequality, the corresponding supplier will be classified into Class 1. Similarly, according to φ(lp z+1 )≤φ(A k )<φ(lp z ) and 0≤φ(A k )<φ(lp z-1 ), and the remaining suppliers can be classified into Category 2, Category 3 and Category 4 respectively. Table 16 lists the final classification results.

[0103] Table 16 Final classification results

[0104]

[0105] It can be seen that suppliers A6 and A7 are assigned to the category of high-quality and low-risk suppliers, while suppliers A9, A8, A4 and A 10 They were assigned to the high-quality medium-risk supplier category, A3 and A1 suppliers were assigned to the medium-quality medium-risk supplier category, and the fourth category of low-quality high-risk suppliers included A2 and A5.

Claims

1. The CPT-TODIMSort new energy vehicle supplier classification method based on interval intuitionistic fuzzy includes the following steps: S1: Problem Definition First, define the goal of the problem, and the set of suppliers of alternative solutions is A n ={A1,A2,...,A n }(n=1,2,...,n), the index set is C m ={C1,C2,...,C m }(m=1,2,...,m), and alternative solution supplier A k About Indicator C n Evaluation of kn The decision maker set is I l ={I1,I2,...,I h }(l=1,2,...,h), at the same time, we assume that the weight vector of these h decision makers is ω=(ω1,ω2,...,ω h ) T ,ω l ≥0, But the weights of the attributes are completely unknown. Next, the supplier classification categories are defined. Categories z (z=1, 2, ..., Z) are defined along with their numbers, which are arranged in order from most preferred to least preferred, and are labeled (eg, excellent, very good, or good). Finally, define the boundaries of each supplier classification category. As shown in Figure 1, the category can be defined by the limit boundaries. For the limit boundary, it consists of the local limit boundary lp zn Composition, that is, each standard C belonging to class z n The minimum level required, as can be seen from Figure 1, is that if there are Z categories, then Z-1 limit bounds are required. S2: Constructing a decision matrix Suppose the decision matrix in the interval-valued intuitionistic fuzzy environment given by the lth decision maker is in and They represent the membership and non-membership of the lth decision maker to the g evaluation indicators of r suppliers. and satisfy and S2-1: Use formula (1) to convert cost-based indicators into benefit-based indicators and form a matrix. S2-2: Use the IVIFWA operator to combine all matrices from different experts Integrate into a group decision matrix S3: Get weight ratio S3-1: Apply IVIF entropy weight method to obtain the original attribute weight vector υ=(υ1,υ2,...,υ m ),υ g ≥0(g=1,2,...,m). S3-2: According to formula (5), for scheme A r and A k Based on the numerical comparison of (r, k∈n and r≠k), the modified potential energy weight is obtained, where δ and τ are parameters describing the curvature of the weighting function. Indicates supplier A r Relative to A k In C g The benefits on Indicates supplier A r Relative to A k In C g δ and τ represent the loss aversion coefficient, δ, τ>0. When δ and τ are smaller (0<δ, τ<1), the risk aversion is higher; when δ and τ are larger (δ, τ>1), the risk aversion is lower, and δ, τ=1 represents risk neutrality. If δ and τ are large enough, it means that the decision maker is risk-seeking and all losses can be ignored. S3-3: Calculate supplier A according to formula (6) r With A k The weight ratio S4: Obtain ranking of alternatives S4-1: Use formula (7) to determine supplier A r With A k In attribute C g Distance D under (ρ=2) rkg . S4-2: Determine supplier A using formula (8) r With A k Compared to the ratio of foreground advantages in the g-th attribute. S4-3: Calculate supplier A according to formulas (9) and (10) r The local foreground dominance Λ(A r )(r=1,2,...,n) and the global foreground dominance φ(A r )(r=1,2,...,n). S4-4: Calculate the limit category limit lp according to formula (11) zn The global prospect dominance φ of . S5: Category Classification According to the overall superiority of each alternative and category boundaries obtained in S4, the alternative suppliers can be assigned to different categories, which can be sorted by using the extreme category boundary diagram, as shown in Figure 2. In the extreme case (Figure 3), supplier A k Will be assigned to class z according to the following rules: If φ(lp1)≤φ(A k )≤1, then A k Belongs to Category 1; If φ(lp z+1 )≤φ(A k )<φ(lp z ), then A k Belongs to the z+1th class, where z=1,2,.,Z-2; If 0≤φ(A k )<φ(lp z-1 ), then A k Belongs to category Z. Therefore, the above rules can be used to classify the suppliers of alternative solutions into corresponding categories using limit boundaries.

2. The interval-valued intuitionistic fuzzy CPT-TODIMSort new energy vehicle supplier classification method according to claim 1 is characterized in that: In step S3, the IVIF entropy weight method is used to obtain the original attribute weight vector υ, which eliminates the adverse effect of the attribute weight information directly given by the decision maker in the traditional TODIM method on the final result, thereby making the decision in reality more reasonable.

3. The CPT-TODIMSort new energy vehicle supplier classification method based on interval intuitionistic fuzzy according to claim 1 is characterized in that: In step S4, the original advantages of the TODIM method are retained. By considering the psychological and logical behavior of decision makers, the different impacts of decision makers' psychological behavior on the final sorting and classification results in the process of sorting new energy vehicle parts suppliers are intuitively demonstrated.

4. The CPT-TODIMSort new energy vehicle supplier classification method based on interval intuitionistic fuzzy according to claim 1 is characterized in that: In step S5, the alternative solution suppliers can be allocated to different categories according to the overall advantage of each alternative solution and the category boundary, so as to make the result more intuitive.

Images