A multi-criteria decision method for ship design selection method and system

Abstract

The application discloses a kind of ship design multi-criteria decision-making method selection method and system. By generalizing the characteristics of general MCDM problem and defining the characteristics of common MCDM method, and establishing the matching mapping relationship between the characteristics of general MCDM problem in problem domain and the characteristics of common MCDM method in method domain, a generalized MCDM method preliminary selection rule is generated, and a preliminary applicable MCDM method / set under a specific decision situation is obtained. On this basis, a generalized MCDM method applicability criterion and the corresponding evaluation model are established. Following the principle of combining qualitative and quantitative analysis, the actual application effect of the selected MCDM method is compared comprehensively, and the optimal MCDM method under the specific decision situation is selected accordingly. The two-step ship design MCDM method selection strategy and system of preliminary selection and optimization can effectively ensure the applicability of the preliminary selected MCDM method and the superiority of the optimized MCDM method.

Description

Technical Field

[0001] This invention belongs to the field of ship design, specifically relating to a strategy and system for selecting a multi-criteria decision-making (MCDM) method in ship design. Background Technology

[0002] Ship design is a complex, iterative, and multidisciplinary systems engineering project. In this process, a systematic and scientific comprehensive evaluation of design schemes can identify and analyze the strengths and weaknesses of various performance indicators, thereby providing designers with guidance for decision-making and optimization. Therefore, scientific and effective evaluation and decision-making methods are of great significance for improving the quality of ship design.

[0003] Multi-Criteria Decision-Making (MCDM) is an effective method for solving evaluation and decision problems. With the rapid development of fundamental theories such as game theory, information theory, matrix theory, fuzzy set theory, rough set theory, and grey number theory, classic MCDM methods and their improved versions have been proposed, including the Analytic Hierarchy Process (AHP), the Top-Solution Approximation Method (TOPSIS), the Fuzzy Comprehensive Evaluation Method (FCE), the VIKOR Multi-Criteria Compromise Solution Ranking Method, the PROMETHEE Preference Order Structure Evaluation Method, and the Grey Relational Analysis Method (GRA). According to incomplete statistics, there are dozens or even hundreds of existing MCDM methods. These methods have been widely applied to evaluation, ranking, classification, and selection problems in fields such as engineering, economics, management, agriculture, and medicine.

[0004] One of the challenges in applying the MCDM method lies in selecting a scientifically effective MCDM method for a given practical MCDM problem—in other words, the applicability of the MCDM method. Often, ship designers choose one or more MCDM methods from their existing knowledge based on their experience and subjective preferences, and then use them directly or with slight modifications to solve the actual MCDM problem. This approach to selecting MCDM methods usually makes it difficult to guarantee the applicability of the chosen method and the reasonableness of the results. Research shows that choosing an inappropriate MCDM method may lead to erroneous evaluation results, resulting in serious consequences such as flawed decision-making. Summary of the Invention

[0005] The technical problem this invention aims to solve is that, although the MCDM method has been proven to be an effective method for solving ship design scheme evaluation and decision-making problems, there are numerous existing MCDM methods, each with significant differences in their basic theories, mathematical models, main problems addressed, applicable conditions, advantages, and disadvantages. This results in ship designers lacking effective references when selecting MCDM methods for practical evaluation and decision-making problems. Therefore, to improve the rationality and effectiveness of ship design scheme evaluation and decision-making, it is urgent to establish a strategy and system for selecting MCDM methods in ship design, providing methodological and tool support for selecting MCDM methods in different decision-making scenarios.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is to provide a multi-criteria decision-making method for ship design, characterized by comprising:

[0007] The initial selection strategy for MCDM methods in ship design is used to initially select applicable MCDM methods / sets based on the characteristics of specific evaluation and decision-making problems.

[0008] The MCDM method optimization strategy for ship design is used to further optimize the MCDM method with the best application effect for actual evaluation and decision-making problems.

[0009] Preferably, the initial selection strategy of the MCDM method for ship design includes:

[0010] Step 101: Organize common MCDM problems and frequently used MCDM methods, and establish an MCDM problem library and an MCDM method library respectively;

[0011] Step 102: Analyze the characteristics of common MCDM problems, summarize the characteristics of general MCDM problems, and construct a generalized MCDM problem characteristic description framework;

[0012] Step 103: Based on the generalized MCDM problem characteristic description framework constructed in Step 102, analyze the characteristics of general MCDM problems and establish a formal description of the characteristics of general MCDM problems.

[0013] Step 104: Analyze the mathematical models of commonly used MCDM methods. Based on the formal description of the general MCDM problem characteristics established in Step 103, identify the corresponding characteristics of commonly used MCDM methods.

[0014] Step 105: Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain, generate generalized preliminary selection rules for MCDM methods, and establish a generalized preliminary selection rule base for MCDM methods.

[0015] Step 106: Analyze the characteristics of the actual MCDM problem under a specific decision-making scenario, and traverse the rules that match the generalized MCDM method preliminary selection rule base established in step 105 to obtain the preliminary applicable MCDM method / set under the specific decision-making scenario.

[0016] Step 107: If no generalized MCDM method initial selection rule matching the characteristics of the actual MCDM problem in a specific decision-making context is found in step 106, update the characteristics of the actual MCDM problem in the specific decision-making context analyzed in step 106 to the generalized MCDM problem characteristic description framework constructed in step 102, and repeat steps 102 to 106.

[0017] Step 108: If more than one MCDM method is initially applicable under a specific decision-making scenario obtained in step 106, the MCDM method optimization strategy for ship design is further adopted to select the MCDM method with the best application effect.

[0018] Preferably, the preferred strategy of the MCDM method for ship design includes:

[0019] Step 201: Based on the concept of systems engineering and following the principle of combining qualitative and quantitative methods, establish a generalized MCDM method applicability criterion and corresponding evaluation model;

[0020] Step 202: Based on the mathematical model of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method;

[0021] Step 203: Solve the actual MCDM problem using each of the initially selected MCDM methods;

[0022] Step 204: Based on the generalized MCDM method applicability criteria and corresponding evaluation model established in Step 201, comprehensively compare and analyze the actual application effects of each initially selected MCDM method, and finally obtain the optimal MCDM method under the specific decision-making scenario.

[0023] Another technical solution of the present invention is to provide a multi-criteria decision-making method selection system for ship design, characterized in that it includes:

[0024] The MCDM problem characteristic definition module is used for:

[0025] Analyze the characteristics of common MCDM problems;

[0026] Based on the characteristics of common MCDM problems, the characteristics of general MCDM problems can be summarized.

[0027] Based on the characteristics of general MCDM problems, a generalized MCDM problem characteristic description framework is constructed, which includes the following:

[0028] The characteristics of general MCDM problems are described from different dimensions through a series of descriptors, thereby constructing a generalized MCDM problem characteristic description framework;

[0029] Based on the generalized MCDM problem characteristic description framework, the characteristics of general MCDM problems are analyzed.

[0030] Based on the characteristics of general MCDM problems, a formal description of the characteristics of general MCDM problems is established.

[0031] The MCDM method characteristic analysis module is used for:

[0032] Analyze the mathematical models of commonly used MCDM methods;

[0033] Based on the mathematical models of commonly used MCDM methods and the formal description of the characteristics of general MCDM problems, identify the corresponding characteristics of commonly used MCDM methods;

[0034] The MCDM method initial selection rule generation module is used for:

[0035] Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain;

[0036] Preliminary selection rules for generating generalized MCDM methods;

[0037] The MCDM method applicability criterion definition module is used for:

[0038] Define a generalized MCDM method applicability criterion;

[0039] Establish a generalized evaluation model for the applicability of the MCDM method;

[0040] The MCDM method's initial selection and optimization module is used for:

[0041] Analyze the characteristics of real-world MCDM problems in specific decision-making scenarios;

[0042] In the generalized MCDM method preliminary rule base, rules that match the characteristics of the actual MCDM problem in the specific decision-making context are traversed to obtain the MCDM method / set that is initially applicable to the specific decision-making context;

[0043] Based on the mathematical models of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method;

[0044] Each of the initially selected MCDM methods was used to calculate and solve the actual MCDM problem;

[0045] Based on the generalized MCDM method applicability criteria and corresponding evaluation models, the actual application effects of each initially selected MCDM method are comprehensively compared and analyzed to obtain the optimal MCDM method under a specific decision-making scenario.

[0046] The calculation results and application effects of each initially selected MCDM method are visualized.

[0047] The basic library maintenance and management module is used for:

[0048] Common MCDM maintenance and management issues and their characteristics;

[0049] Maintenance and management of general MCDM problem characteristics and their formal description;

[0050] Maintain and manage a generalized MCDM problem characteristic description framework;

[0051] Commonly used MCDM methods for maintenance and management and their characteristics;

[0052] Preliminary selection rules for generalized MCDM methods for maintenance and management;

[0053] Applicability criteria and evaluation model of generalized MCDM method for maintenance management.

[0054] Preferably, the construction of a generalized MCDM problem characteristic description framework includes:

[0055] Decision-making objectives;

[0056] Scale for comparing the performance indicators of the proposed solutions;

[0057] Whether or not indicator weights are considered, the general MCDM problem can be further divided into two categories: those without indicator weights and those with indicator weights.

[0058] Based on whether there are uncertainties, the general MCDM problem can be subdivided into: no uncertainties and uncertainties.

[0059] Based on the distribution of qualitative and quantitative indicators, general MCDM problems can be further subdivided into: those containing only qualitative indicators; those containing only quantitative indicators; and those containing both qualitative and quantitative indicators.

[0060] Based on whether the number of solutions is stable, the general MCDM problem can be further divided into: stable and unstable.

[0061] Preferably, the decision-making objective includes selection, classification, sorting + selection, and classification + selection;

[0062] The methods for comparing the performance indicators of the schemes include: qualitative comparison, quantitative comparison, and relative comparison.

[0063] Preferably, the inclusion of both qualitative and quantitative indicators is further subdivided into: primarily qualitative indicators and primarily quantitative indicators.

[0064] The number of schemes is stable and can be further subdivided into: number of schemes ≥ 2, number of schemes = 1.

[0065] Preferably, the sorting problem is further subdivided into local sorting and complete sorting.

[0066] Preferably, when considering indicator weights in the MCDM problem, the weighting method is further subdivided into: qualitative weighting, quantitative weighting, and relative weighting.

[0067] When uncertainties exist, the general MCDM problem can be further subdivided into: uncertainty of input data, uncertainty of decision preferences, and dual uncertainty of input data and decision preferences.

[0068] Preferably, the uncertainty of the input data is further subdivided into: uncertainty of indicator weights, uncertainty of scheme performance, and dual uncertainty of indicator weights and scheme performance;

[0069] The uncertainty of the decision preference is further subdivided into: uncertainty of the threshold for determining equal importance (superiority or inferiority), uncertainty of the threshold for determining more important (excellent) factors, and dual uncertainty of the thresholds for determining both equal importance (superiority or inferiority) and more important (excellent) factors.

[0070] To address the technical challenge of the large variety and number of MCDM methods available today, which leaves ship designers without effective references when selecting MCDM methods for specific evaluation and decision-making problems, this invention provides a method and system for selecting MCDM methods in ship design. By generalizing the characteristics of general MCDM problems and defining the characteristics of commonly used MCDM methods, and establishing a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain, a generalized preliminary selection rule for MCDM methods is generated, resulting in a preliminary set of applicable MCDM methods / methods for specific decision-making scenarios. Based on this, a generalized MCDM method applicability criterion and corresponding evaluation model are established. Following the principle of combining qualitative and quantitative analysis, the actual application effects of the selected MCDM methods are comprehensively compared, and the optimal MCDM method for specific decision-making scenarios is selected accordingly. The established two-step strategy and system for selecting MCDM methods in ship design, involving preliminary selection and optimal selection, effectively ensures the applicability of the initially selected MCDM methods and the superiority of the selected MCDM methods.

[0071] This invention provides reliable guidance for selecting the MCDM method in evaluating ship design schemes under different decision-making scenarios, thereby enabling a systematic and scientific comprehensive evaluation of ship design schemes, accurately identifying the strengths and weaknesses of various performance indicators, and ultimately providing designers with guidance and optimization basis for scheme decisions. Therefore, this invention is of great significance for improving the quality of ship design. Attached Figure Description

[0072] Figure 1 This illustrates the selection strategy for multi-criteria decision-making methods in ship design;

[0073] Figure 2 This illustrates a multi-criteria decision-making system for ship design.

[0074] Figure 3 This illustrates a generalized framework for describing the characteristics of MCDM problems;

[0075] Figure 4 This illustrates the overall evaluation index system for a certain research vessel.

[0076] Figure 5 The calculation results of each preliminary MCDM method are illustrated;

[0077] Figure 6 The sorting results of the calculations from each of the initial MCDM methods are shown.

[0078] Figure 7 The standard deviation of the calculation results for each initial MCDM method is shown. Detailed Implementation

[0079] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0080] One aspect of this invention discloses a method for selecting an MCDM method in ship design, such as... Figure 1 As shown, it specifically includes two parts:

[0081] (I) Preliminary selection strategy for MCDM method in ship design;

[0082] (II) Optimization strategy of MCDM method for ship design.

[0083] The "Initial Selection Strategy for MCDM Methods in Ship Design" selects applicable MCDM methods / sets based on the characteristics of specific evaluation and decision-making problems. The "Optimization Strategy for MCDM Methods in Ship Design" further optimizes the MCDM methods with the best application performance based on the actual evaluation and decision-making problems themselves.

[0084] The two-step MCDM method of preliminary selection and optimization established in this invention can effectively ensure the applicability of the preliminary MCDM method and the superiority of the optimization MCDM method, and provide reliable guidance for the selection of MCDM method for ship design scheme evaluation and decision-making under different decision-making scenarios.

[0085] (I) Preliminary Selection Strategy for MCDM Methodology in Ship Design

[0086] The initial selection strategy of the MCDM method for ship design has the following overall technical solution:

[0087] Step 101: Organize common MCDM problems and frequently used MCDM methods, and establish an MCDM problem library and an MCDM method library respectively;

[0088] Step 102: Analyze the characteristics of common MCDM problems, summarize the characteristics of general MCDM problems, and construct a generalized MCDM problem characteristic description framework;

[0089] Step 103: Based on the generalized MCDM problem characteristic description framework constructed in Step 102, analyze the characteristics of general MCDM problems and establish a formal description of the characteristics of general MCDM problems.

[0090] Step 104: Analyze the mathematical models of commonly used MCDM methods. Based on the formal description of the general MCDM problem characteristics established in Step 103, identify the corresponding characteristics of commonly used MCDM methods.

[0091] Step 105: Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain, generate generalized preliminary selection rules for MCDM methods, and establish a generalized preliminary selection rule base for MCDM methods.

[0092] Step 106: Analyze the characteristics of the actual MCDM problem under a specific decision-making scenario, and traverse the rules that match the generalized MCDM method preliminary selection rule base established in step 105 to obtain the preliminary applicable MCDM method (or MCDM method set) under the specific decision-making scenario.

[0093] Step 107: If no generalized MCDM method initial selection rule matching the characteristics of the actual MCDM problem in a specific decision-making context is found in step 106, update the characteristics of the actual MCDM problem in the specific decision-making context analyzed in step 106 to the generalized MCDM problem characteristic description framework constructed in step 102, and repeat steps 102 to 106.

[0094] Step 108: If more than one MCDM method is initially applicable under a specific decision-making scenario obtained in step 106, the MCDM method optimization strategy for ship design is further adopted to select the MCDM method with the best application effect.

[0095] (II) Optimization Strategy of MCDM Method in Ship Design

[0096] The overall technical solution of the MCDM method for ship design optimization strategy is as follows:

[0097] Step 201: Based on the concept of systems engineering and following the principle of combining qualitative and quantitative methods, establish a generalized MCDM method applicability criterion and corresponding evaluation model;

[0098] Step 202: Based on the mathematical model of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method;

[0099] Step 203: Solve the actual MCDM problem using each of the initially selected MCDM methods;

[0100] Step 204: Based on the generalized MCDM method applicability criteria and corresponding evaluation model established in Step 201, comprehensively compare and analyze the actual application effects of each initially selected MCDM method, and finally obtain the optimal MCDM method under the specific decision-making scenario.

[0101] Another aspect of this invention is to provide a system for selecting MCDM methods in ship design, such as... Figure 2 As shown, it mainly includes:

[0102] (a) The MCDM problem characteristic definition module is used for:

[0103] (1) Analyze the characteristics of common MCDM problems.

[0104] (2) Based on the characteristics of common MCDM problems, summarize the characteristics of general MCDM problems.

[0105] (3) Based on the characteristics of general MCDM problems, a generalized MCDM problem characteristic description framework is constructed, which specifically includes the following:

[0106] By describing the characteristics of general MCDM problems from different dimensions through a series of descriptors, a generalized framework for describing the characteristics of MCDM problems is constructed, such as... Figure 3 As shown, it further includes the following:

[0107] 1) Decision-making objectives

[0108] In general MCDM problems, the decision objectives typically include:

[0109] (a) Choose;

[0110] (b) Classification;

[0111] (c) Sort + Select;

[0112] (d) Classification + Selection.

[0113] The sorting problem can be further subdivided into:

[0114] (a) Local sorting;

[0115] (b) Fully sorted.

[0116] 2) Scale for comparing the performance indicators of the proposed solutions

[0117] In general MCDM problems, the main methods for comparing the performance metrics of different solutions are:

[0118] (a) Qualitative comparison;

[0119] (b) Quantitative comparison;

[0120] (c) Relative comparison.

[0121] 3) Whether to consider indicator weights

[0122] Depending on whether indicator weights are considered, the general MCDM problem can be subdivided into:

[0123] (a) Without considering indicator weights;

[0124] (b) Consider the weight of the indicators.

[0125] When considering indicator weights in the MCDM problem, it can be further subdivided according to the method of indicator weighting:

[0126] (a) Qualitative empowerment;

[0127] (b) Quantitative weighting;

[0128] (c) Relative weighting.

[0129] 4) Are there any uncertainties?

[0130] Based on the presence or absence of uncertainties, the MCDM problem can generally be subdivided into:

[0131] (a) Does not exist;

[0132] (b) There are uncertainties.

[0133] When uncertainties exist, the general MCDM problem can be further subdivided into:

[0134] (a) Uncertainty of input data;

[0135] (b) Uncertainty in decision preferences;

[0136] (c) The dual uncertainty of input data and decision preferences.

[0137] The uncertainty of the input data can be further subdivided into:

[0138] (a) Uncertainty regarding indicator weights;

[0139] (b) Uncertainty regarding the performance of the proposed solution;

[0140] (c) The dual uncertainty of indicator weights and scheme performance.

[0141] The uncertainty of decision preferences can be further subdivided into:

[0142] (a) Uncertainty in determining the threshold for equal importance (superiority / inferiority);

[0143] (b) Uncertainty regarding the threshold for determining what is more important (excellent);

[0144] (c) The dual uncertainty of the threshold for determining equal importance (good or bad) and more important (excellent).

[0145] 5) Distribution of qualitative / quantitative indicators

[0146] Based on the distribution of qualitative / quantitative indicators, MCDM problems can generally be subdivided into:

[0147] (a) Includes only qualitative indicators;

[0148] (b) Includes only quantitative indicators;

[0149] (c) It includes both qualitative and quantitative indicators.

[0150] Among them, "(c) includes both qualitative and quantitative indicators" can be further subdivided into:

[0151] (a) Primarily qualitative indicators;

[0152] (b) Primarily quantitative indicators.

[0153] 6) Number of options

[0154] First, based on whether the number of solutions is stable, the MCDM problem can generally be subdivided into:

[0155] (a) Stable;

[0156] (b) Unstable.

[0157] The instability of the number of schemes means that new schemes may be added to the evaluation in the future. In this case, the evaluation results of some MCDM methods will change accordingly due to the addition of new schemes, which is not conducive to the analysis and application of the evaluation results.

[0158] The number of solutions is stable and can be further subdivided into:

[0159] (a) Number of options ≥ 2;

[0160] (b) Number of options = 1.

[0161] (4) Based on the generalized MCDM problem characteristic description framework, analyze the characteristics of general MCDM problems.

[0162] (5) Based on the characteristics of general MCDM problems, establish a formal description of the characteristics of general MCDM problems, which specifically includes the following:

[0163] Further analysis of the characteristics of general MCDM problems and establishment of a formal description of the characteristics of general MCDM problems will lay the foundation for subsequent identification of the corresponding characteristics of commonly used MCDM methods.

[0164] The formal description of the characteristics of a typical MCDM problem is as follows:

[0165] 1) The decision-making objective is as follows:

[0166]

[0167] In the formula, DP represents decision problems, C1(DP) is the value of DP on problem characteristic C1, u is the selected design scheme, dim is the number, q is the design scheme corresponding to the selection threshold, v is the design scheme that was not selected, k is the ranked design scheme, and S D () represents the satisfaction level of decision-making experts with the design scheme, η() is the vector norm, and k R For alternative design schemes, a subset of equally good and bad options is considered; approximately, for partial or complete ranking, u B The design schemes are categorized.

[0168]

[0169] In equation (2), C1.1 (DP) represents DP in problem characteristic C. 1.1 The values ​​on the above are: E() is the comprehensive evaluation value of the design scheme, R is the incomparability relationship, and a i Let i be the design scheme, and G(A) be the overall performance of design scheme A.

[0170] 2) Scale for comparing the performance indicators of the proposed solutions

[0171]

[0172] In equation (3), C2(DP) is the value of DP on problem characteristic C2, g i (a) represents the performance of scheme a for index i, r represents the quantitative difference in performance between schemes for index i, and e jk For the target indicator g, scheme a j With a k Relative comparison of performance, E |A|×|A| This is a pairwise comparison matrix, with the matrix dimension representing the number of alternative design options.

[0173] 3) Whether to consider indicator weights

[0174]

[0175] In equation (4), C3(DP) is the value of DP on problem characteristic C3, r is the quantitative difference between the weights of indicator i and indicator j, and w ij W represents the relative weights between indicator i and indicator j. |G|×|G| This is a pairwise comparison matrix, with the matrix dimension being the number of indicators.

[0176]

[0177] In equation (5), C 3.1 (DP) represents DP in problem characteristic C. 3.1 The value that can be taken on.

[0178] 4) Are there any uncertainties?

[0179]

[0180] In equation (6), C4(DP) is the value of DP on problem characteristic C4, and N fuzzy It is a triangular fuzzy number or a trapezoidal fuzzy number, (n l ;n u ;α F ;β F ) represents the membership function parameter of the fuzzy number, q represents the equivalence threshold, and p represents the better threshold.

[0181]

[0182]

[0183]

[0184] In the formula, C 4.1 (DP) represents DP in problem characteristic C. 4.1 The value of C 4.1.1 (DP) represents DP in problem characteristic C. 4.1.1 The value of C 4.1.2 (DP) represents DP in problem characteristic C. 4.1.2 The value that can be taken on.

[0185] 5) Distribution of qualitative / quantitative indicators

[0186]

[0187] In equation (10), g° represents the number of qualitative indicators, g * Let n be the number of quantitative indicators, and C5(DP) be the value of DP on problem characteristic C5.

[0188]

[0189] In equation (10), C 5.1 (DP) represents DP in problem characteristic C. 5.1 The value that can be taken on.

[0190] 6) Number of options

[0191]

[0192] In equation (12), C6(DP) is the value of DP on problem characteristic C6, and n a To evaluate the number of alternative solutions before starting, n′ a This represents the number of alternative solutions at any given moment during the evaluation process.

[0193]

[0194] In equation (13), C 6.1 (DP) represents DP in problem characteristic C. 6.1 The value that can be taken on.

[0195] (II) The MCDM method characteristic analysis module is used for:

[0196] (1) Analyze the mathematical model of commonly used MCDM methods.

[0197] (2) Based on the mathematical models of commonly used MCDM methods and the formal description of the characteristics of general MCDM problems, identify the corresponding characteristics of commonly used MCDM methods, specifically including the following:

[0198] A preliminary overview of commonly used MCDM methods is as follows: 1) SAW (Simple Weighted Average), 2) TOPSIS, 3) VIKOR, 4) COPRAS, 5) PROMETHEE II, 6) AHP, 7) WGA (Weighted Geometric Average), 8) GRA (Dunner's Relationship), 9) MAUT, 10) MAVT, 11) UTA, 12) WASPAS, 13) FCE, 14) FAHP, 15) FTOPSIS, 16) FSAW, 17) PROMETHEE II, 18) FVIKOR.

[0199] Based on the formal description of the general MCDM problem characteristics and the mathematical models of the commonly used MCDM methods mentioned above, the corresponding characteristics of the commonly used MCDM methods are identified, as shown in Table 1.

[0200] Table 1. Characteristics of commonly used MCDM methods

[0201] Serial Number MCDM method m1 m1.1 m2 m3 m3.1 m4 m4.1 m4.1.1 m4.1.2 m5 m5.1 m6 m6.1 1 SAW 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 2 TOPSIS 3 2 2 1 2 0 0 0 0 3 2 1，2 1，2 3 VIKOR 3 2 2 1 2 0 0 0 0 3 2 1，2 1，2 4 COPRAS 3 2 2 1 2 0 0 0 0 3 2 1，2 1，2 5 PROMETHEE II 3 2 2 1 2 1 2 0 3 3 2 1，2 1 6 AHP 3 2 3 1 3 0 0 0 0 3 1，2 1 1 7 WGA 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 8 GRA 3 2 2 1 2 0 0 0 0 3 2 1，2 1，2 9 MAUT 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 10 MAVT 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 11 UTA 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 12 WASPAS 3 2 2 1 2 0 0 0 0 3 1，2 1，2 1，2 13 FCE 3 2 2 1 2 1 1 2 0 3 1 1，2 1，2 14 FAHP 3 2 3 1 3 1 1 3 0 3 1，2 1 1 15 FTOPSIS 3 2 2 1 2 1 1 3 0 3 2 1，2 1，2 16 FSAW 3 2 2 1 2 1 1 3 0 3 1，2 1，2 1，2 17 FPROMETHEEII 3 2 2 1 2 1 3 3 3 3 2 1，2 1 18 FVIKOR 3 2 2 1 2 1 1 3 0 3 2 1，2 1，2

[0202] It is worth noting that the present invention aims to establish a strategy and system for selecting MCDM methods in ship design, and to verify its feasibility and effectiveness. Therefore, the initial number of MCDM methods is not the main consideration. The MCDM method library can be further enriched in the future. The established strategy and system for selecting MCDM methods in ship design can be used to provide better support for the selection of MCDM methods for evaluating ship design schemes under different decision-making scenarios.

[0203] (III) The MCDM method initial selection rule generation module is used for:

[0204] (1) Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain.

[0205] (2) Preliminary selection rules for generating generalized MCDM methods, specifically including the following:

[0206] Assume that the dynamic programming problem in a general MCDM problem has the problem property c = [c1, c2, ..., c r For practical MCDM problems, DP * Problematic features s / r represents the number of characteristics of a typical / practical MCDM problem, where s ≤ r. Assume a commonly used MCDM method is m = [m1, m2, ..., m...]. k ], MCDM method m kThere is a method property p = [p1, p2, ..., p t When the actual MCD M problem DP * If a solution exists, then s≤r≤t.

[0207] Construct the characteristic matrix F of the MCDM method = (f ij ) k×t For practical MCDM problems, DP * Select the initially applicable MCDM methods according to the following formula.

[0208]

[0209] By establishing a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain, a preliminary selection rule for generalized MCDM methods is generated, as shown in Table 2.

[0210] Table 2 Preliminary Selection Rules for Generalized MCDM Methods

[0211]

[0212] (iv) The MCDM method applicability criterion definition module is used for:

[0213] (1) Define the generalized MCDM method applicability criteria.

[0214] (2) Establish a generalized evaluation model for the applicability of the MCDM method, which includes the following:

[0215] Based on systems engineering principles and following the principle of combining qualitative and quantitative methods, a generalized MCDM method applicability criterion and corresponding evaluation model are established to compare and analyze the application effects of each initially selected MCDM method on practical MCDM problems.

[0216] 1) In terms of qualitative analysis, we mainly consider the ease of use of the methods, the difficulty of data preparation, and the availability of supporting tools.

[0217] 2) In terms of quantitative analysis, the analysis mainly focuses on the consistency of the ranking and the dispersion of the distribution.

[0218] (a) Order consistency

[0219] Ranking consistency refers to the degree of consistency in the ranking results of different MCDM methods; higher consistency indicates higher accuracy. The correlation between the calculation results of different MCDM methods can be tested using the Spearman correlation coefficient.

[0220]

[0221] In the formula, ρ ijLet x be the correlation coefficient between the results calculated by method i and method j. ik With x jk These are the calculation results for scheme k under method i and method j, respectively.

[0222] According to the Spearman correlation coefficient ρ ij The calculation results are used to construct the Spearman correlation coefficient matrix ρ:

[0223]

[0224] The average correlation ρ between method i and other decision-making methods is calculated using the following formula. i , ρ i The larger the value, the better the effect of method i.

[0225]

[0226] (b) Dispersion of distribution

[0227] The dispersion of decision results refers to the degree of dispersion of the calculation results of each scheme using different decision-making methods. It is generally believed that the greater the dispersion, the better the decision effect.

[0228] Measured by standard deviation σ:

[0229]

[0230] In the formula, σ i Let σ be the standard deviation of method i. i The larger x is, the better method i becomes. ij For scheme j under method i Calculation results Let represent the expected results of each scheme under method i.

[0231] (V) The MCDM method preliminary selection and optimization module is used for:

[0232] (1) Analyze the characteristics of actual MCDM problems in specific decision-making scenarios;

[0233] (2) Traverse the rules in the generalized MCDM method preliminary selection rule base that match the characteristics of the actual MCDM problem in the specific decision-making situation to obtain the MCDM method (or MCDM method set) that is initially applicable in the specific decision-making situation.

[0234] (3) Based on the mathematical model of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method;

[0235] (4) The actual MCDM problem is solved by each of the preliminary MCDM methods;

[0236] (5) Based on the generalized MCDM method applicability criteria and the corresponding evaluation model, the actual application effects of each initially selected MCDM method are comprehensively compared and analyzed to obtain the optimal MCDM method under a specific decision-making scenario.

[0237] (6) Visualize the calculation results and application effects of each preliminary MCDM method.

[0238] (vi) Basic library maintenance and management module, used for:

[0239] (1) Common MCDM maintenance and management issues and their characteristics;

[0240] (2) Maintenance and management of general MCDM problem characteristics and their formal description;

[0241] (3) Maintain and manage a generalized MCDM problem characteristic description framework;

[0242] (4) Commonly used MCDM methods for maintenance and management and their characteristics;

[0243] (5) Preliminary selection rules for generalized MCDM methods for maintenance and management;

[0244] (6) Applicability criteria and evaluation model of generalized MCDM method for maintenance management.

[0245] The following specific example will be used to illustrate the technical solution disclosed in this invention.

[0246] 1. Preliminary selection of MCDM method for ship design

[0247] Taking into account the actual situation in the field of ship design, the scope of ship design scheme evaluation and decision-making issues is clarified. Based on the established generalized MCDM method preliminary selection rules, a preliminary applicable MCDM method (or MCDM method set) is obtained, as shown in Table 3.

[0248] Table 3. Scope of Ship Design Scheme Evaluation and Decision-Making Problems and Preliminary Applicable MCDM Method

[0249]

[0250] As shown in Table 3, the ship design scheme evaluation and decision-making problem can be broadly divided into two categories: deterministic decision-making scenarios and fuzzy decision-making scenarios. These mainly involve the selection of multiple schemes during the ship design phase and the comprehensive evaluation after the initial selection of a scheme. Taking the selection of the MCDM method for the overall scheme evaluation and decision-making of a research vessel under a deterministic decision-making scenario as an example, this study verifies the established MCDM method selection strategy for ship design.

[0251] For the overall scheme evaluation decision problem of a research vessel under a given decision-making scenario, based on the general MCDM method preliminary selection rules, the following MCDM methods are preliminarily applicable: SAW, WGA, MAUT, MAVT, UTA, WASPAS, TOPSIS, VIKOR, COPRAS, GRA, and AHP. Considering the actual application of MCDM methods in the field of ship design, SAW, WGA, TOPSIS, GRA, and several classic improved methods are selected for comparative analysis of their practical application effects.

[0252] 2. Optimal Selection of MCDM Method for Ship Design

[0253] For the overall scheme evaluation decision problem of a research vessel under a given decision-making scenario, given the evaluation index system ( Figure 4 Given the indicator weights and standardized indicator performance values ​​(Table 4), each of the preliminary MCDM methods was used to optimize the five candidate overall schemes. The application effects of each preliminary MCDM method were comprehensively compared and analyzed from the aspects of sorting consistency, distribution dispersion, ease of use, data preparation difficulty, and tool support.

[0254] Table 4. Weights and Performance of Evaluation Indicators for a Scientific Research Vessel's Overall Design

[0255] Evaluation indicators Indicator weights Option 1 Option 2 Option 3 Option 4 Option 5 Total life cycle cost 0.1864 0.9074 0.9117 0.9204 0.8010 0.9265 Technological risks 0.0977 0.9529 0.8474 0.6271 0.6315 0.8110 Cost risk 0.0154 0.5635 0.6585 0.9071 0.8270 0.6755 Schedule risk 0.0394 0.9567 0.9751 0.6218 0.8446 0.7566 Discharge reserve 0.0282 0.8162 0.5172 0.9646 0.8741 0.7009 Space reserves 0.0105 0.5488 0.7194 0.6750 0.7253 0.5380 Stable reserves 0.0114 0.6392 0.6908 0.5983 0.5419 0.6200 Main unit power reserve 0.0123 0.7734 0.8828 0.6255 0.6145 0.5617 Power plant power reserves 0.0045 0.9788 0.8976 0.8080 0.9567 0.5920 reliability 0.0251 0.9824 0.5934 0.7366 0.5762 0.6200 Maintainability 0.0869 0.5788 0.7449 0.6758 0.9129 0.7086 protection 0.0426 0.9853 0.7228 0.9154 0.7692 0.5248 compatibility 0.0162 0.9786 0.8232 0.7926 0.9981 0.9514 stability 0.0029 0.7427 0.8547 0.7749 0.5391 0.9724 speed 0.0084 0.9001 0.8773 0.9586 0.7213 0.7454 Wave resistance 0.0084 0.5709 0.6380 0.6429 0.5533 0.7446 range 0.0014 0.7109 0.8399 0.8786 0.9809 0.6689 Self-sustaining 0.0012 0.9579 0.8275 0.8769 0.5023 0.9500 Maneuverability 0.0068 0.8961 0.5813 0.6902 0.8875 0.6846 Protective capabilities 0.0154 0.9797 0.5595 0.7839 0.9087 0.5556 Damage resistance 0.0054 0.8279 0.7492 0.5379 0.9343 0.8901 Damage control capability 0.0434 0.5179 0.9799 0.5270 0.5422 0.6949 Lifesaving ability 0.0154 0.9246 0.6702 0.7654 0.6999 0.6208 Marine biological research capabilities 0.0154 0.9670 0.7926 0.8896 0.6299 0.7020 Marine geological research capabilities 0.0057 0.8394 0.6119 0.9670 0.9000 0.5482 Marine water body scientific research capabilities 0.0441 0.8789 0.8756 0.5650 0.7157 0.5660 Marine meteorological scientific research capabilities 0.0145 0.8716 0.6275 0.7844 0.9553 0.9710 Navigation capabilities 0.1069 0.6961 0.7530 0.7347 0.5909 0.9781 Capacity to Charge 0.0169 0.8277 0.8495 0.5060 0.6319 0.7876 communication capability 0.0432 0.5856 0.9455 0.6686 0.5728 0.5299 Instrument support capability 0.0287 0.8530 0.9796 0.5811 0.5680 0.6174 Resupply Reception Capability 0.0107 0.5159 0.7736 0.8971 0.9346 0.6766 food security capability 0.0116 0.6385 0.5693 0.6556 0.7899 0.9106 Medical security capabilities 0.0126 0.5231 0.5746 0.7643 0.7749 0.5077 Entertainment security capabilities 0.0045 0.5486 0.6288 0.5828 0.5725 0.5215

[0256] Based on the above input data, the calculation results and ranking results of each initially selected MCDM method are shown in Table 5.

[0257] Table 5. Calculation results and ranking results of each preliminary MCDM method.

[0258]

[0259] Calculate the Spearman correlation coefficient matrix ρ:

[0260]

[0261] Further calculate the average correlation ρ between method i and other decision-making methods. i :

[0262] ρ i =[0.94,0.94,0.88,0.88,0.82,0.94]

[0263] Figure 5 and Figure 6 The calculation results and ranking results of each MCDM method under the deterministic decision-making scenario are presented. Through analysis of... Figure 5 , Figure 6 And a comparative analysis of Spearman correlation coefficients revealed that:

[0264] (1) The calculation results of each preliminary MCDM method under the decision-making scenario are basically consistent in distribution and ranking, and have a high degree of credibility overall.

[0265] (2) When used alone, GRA-Dun correlation, improved GRA-area correlation and improved GRA-slope correlation all produce reverse order phenomenon and have certain limitations, while the improved GRA-comprehensive correlation has better application effect.

[0266] (3) SAW, WGA and improved GRA-comprehensive correlation degree performed best in terms of ranking consistency, followed by TOPSIS and improved TOPSIS.

[0267] To further analyze the applicability of each preliminary MCDM method, the discreteness of the obtained calculation results is analyzed. Figure 7 The standard deviations of the calculation results obtained from each initial MCDM method are given. Figure 7 As can be seen, the dispersion of calculation results varies considerably among different MCDM methods, with TOPSIS, GRA-Dun correlation, and improved GRA-comprehensive correlation showing better dispersion. Furthermore, TOPSIS and GRA algorithms are relatively mature, well-supported by software tools, require less data preparation, and are easy to use.

[0268] Taking all factors into consideration, the improved GRA-comprehensive correlation degree is the most effective application method for evaluating the overall scheme of a research vessel under a given decision-making scenario.

Claims

1. A method for selecting a multi-criteria decision-making method for ship design, characterized in that, include: The initial selection strategy for MCDM methods in ship design is used to initially select applicable MCDM methods based on the characteristics of specific evaluation and decision-making problems. The MCDM method optimization strategy for ship design is used to further optimize the MCDM method with the best application effect for practical evaluation and decision-making problems. The initial selection strategy for the MCDM method in ship design includes: Step 101: Organize common MCDM problems and frequently used MCDM methods, and establish an MCDM problem library and an MCDM method library respectively; Step 102: Analyze the characteristics of common MCDM problems, summarize the characteristics of general MCDM problems, and construct a generalized MCDM problem characteristic description framework; Step 103: Based on the generalized MCDM problem characteristic description framework constructed in Step 102, analyze the characteristics of general MCDM problems and establish a formal description of the characteristics of general MCDM problems. Step 104: Analyze the mathematical models of commonly used MCDM methods. Based on the formal description of the general MCDM problem characteristics established in Step 103, identify the corresponding characteristics of commonly used MCDM methods. Step 105: Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain, generate generalized preliminary selection rules for MCDM methods, and establish a generalized preliminary selection rule base for MCDM methods. Step 106: Analyze the characteristics of the actual MCDM problem under a specific decision-making scenario, and traverse the rules that match the generalized MCDM method preliminary selection rule base established in step 105 to obtain the preliminary applicable MCDM method / set under the specific decision-making scenario. Step 107: If no generalized MCDM method initial selection rule matching the characteristics of the actual MCDM problem in a specific decision-making context is found in step 106, update the characteristics of the actual MCDM problem in the specific decision-making context analyzed in step 106 to the generalized MCDM problem characteristic description framework constructed in step 102, and repeat steps 102 to 106. Step 108: If more than one MCDM method is initially applicable under a specific decision-making scenario obtained in step 106, the MCDM method optimization strategy for ship design is further adopted to select the MCDM method with the best application effect.

2. The method for selecting a multi-criteria decision-making method for ship design as described in claim 1, characterized in that, The preferred strategies for the MCDM method in ship design include: Step 201: Based on the concept of systems engineering and following the principle of combining qualitative and quantitative methods, establish a generalized MCDM method applicability criterion and corresponding evaluation model; Step 202: Based on the mathematical model of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method; Step 203: Solve the actual MCDM problem using each of the initially selected MCDM methods; Step 204: Based on the generalized MCDM method applicability criteria and corresponding evaluation model established in Step 201, comprehensively compare and analyze the actual application effects of each initially selected MCDM method, and finally obtain the optimal MCDM method under the specific decision-making scenario.

3. A multi-criteria decision-making system for ship design, characterized in that, include: The MCDM problem characteristic definition module is used for: Analyze the characteristics of common MCDM problems; Based on the characteristics of common MCDM problems, the characteristics of general MCDM problems can be summarized. Based on the characteristics of general MCDM problems, a generalized MCDM problem characteristic description framework is constructed, which includes the following: The characteristics of general MCDM problems are described from different dimensions through a series of descriptors, thereby constructing a generalized MCDM problem characteristic description framework; Based on the generalized MCDM problem characteristic description framework, the characteristics of general MCDM problems are analyzed. Based on the characteristics of general MCDM problems, a formal description of the characteristics of general MCDM problems is established. The MCDM method characteristic analysis module is used for: Analyze the mathematical models of commonly used MCDM methods; Based on the mathematical models of commonly used MCDM methods and the formal description of the characteristics of general MCDM problems, identify the corresponding characteristics of commonly used MCDM methods; The MCDM method initial selection rule generation module is used for: Establish a matching mapping relationship between the characteristics of general MCDM problems in the problem domain and the characteristics of commonly used MCDM methods in the method domain; Preliminary selection rules for generating generalized MCDM methods; The MCDM method applicability criterion definition module is used for: Define a generalized MCDM method applicability criterion; Establish a generalized evaluation model for the applicability of the MCDM method; The MCDM method's initial selection and optimization module is used for: Analyze the characteristics of real-world MCDM problems in specific decision-making scenarios; In the generalized MCDM method preliminary rule base, rules that match the characteristics of the actual MCDM problem in the specific decision-making context are traversed to obtain the MCDM method / set that is initially applicable to the specific decision-making context; Based on the mathematical models of each preliminary MCDM method, prepare the input data required for the application of each preliminary MCDM method; Each of the initially selected MCDM methods was used to calculate and solve the actual MCDM problem; Based on the generalized MCDM method applicability criteria and corresponding evaluation models, the actual application effects of each initially selected MCDM method are comprehensively compared and analyzed to obtain the optimal MCDM method under a specific decision-making scenario. The calculation results and application effects of each initially selected MCDM method are visualized. The basic library maintenance and management module is used for: Common MCDM maintenance and management issues and their characteristics; Maintenance and management of general MCDM problem characteristics and their formal description; Maintain and manage a generalized MCDM problem characteristic description framework; Commonly used MCDM methods for maintenance and management and their characteristics; Preliminary selection rules for generalized MCDM methods for maintenance and management; Applicability criteria and evaluation model of generalized MCDM method for maintenance management.

4. The multi-criteria decision-making system for ship design as described in claim 3, characterized in that, The construction of the generalized MCDM problem characteristic description framework includes: Decision-making objectives; Methods for comparing the performance indicators of the proposed solutions; Whether or not indicator weights are considered, the general MCDM problem can be further divided into two categories: those without indicator weights and those with indicator weights. Based on whether there are uncertainties, the general MCDM problem can be subdivided into: no uncertainties and uncertainties. Based on the distribution of qualitative and quantitative indicators, general MCDM problems can be further subdivided into: those containing only qualitative indicators; those containing only quantitative indicators; and those containing both qualitative and quantitative indicators. Based on whether the number of solutions is stable, the general MCDM problem can be further divided into: stable and unstable.

5. The multi-criteria decision-making system for ship design as described in claim 4, characterized in that, The decision-making objectives include selection, classification, sorting + selection, and classification + selection; The methods for comparing the performance indicators of the schemes include: qualitative comparison, quantitative comparison, and relative comparison.

6. The multi-criteria decision-making system for ship design as described in claim 4, characterized in that: The aforementioned indicators include both qualitative and quantitative indicators, which can be further subdivided into: primarily qualitative indicators and primarily quantitative indicators. The number of schemes is stable and can be further subdivided into: number of schemes ≥ 2, number of schemes = 1.

7. The multi-criteria decision-making system for ship design as described in claim 5, characterized in that, The sorting is further subdivided into local sorting and complete sorting.

8. The multi-criteria decision-making system for ship design as described in claim 4, characterized in that, When considering indicator weights in MCDM problems, the weighting method can be further subdivided into: qualitative weighting, quantitative weighting, and relative weighting. When uncertainties exist, the general MCDM problem can be further subdivided into: uncertainty of input data, uncertainty of decision preferences, and dual uncertainty of input data and decision preferences.

9. The multi-criteria decision-making system for ship design as described in claim 8, characterized in that, The uncertainty of the input data is further subdivided into: uncertainty of indicator weights, uncertainty of scheme performance, and dual uncertainty of indicator weights and scheme performance. The uncertainty of decision preferences is further subdivided into: uncertainty in determining equally important thresholds, uncertainty in determining more important thresholds, and dual uncertainty in determining both equally important and more important thresholds.

Images