System architecture scheme evaluation method based on improved intuitionistic fuzzy analytic hierarchy process
Patent Information
- Application Number
- CN202610966931.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-01
- Publication Date
- 2026-10-13
AI Technical Summary
但是上述方法直接应用于复杂信息系统架构方案评价时仍存在不足：例如层次分析法在系统架构方案评价中具有结构化、多目标权衡的优势，但其主观性、静态性与扩展性不足限制了在动态复杂信息系统中的应用；ATAM在需求冲突分析和非功能性需求评估中广泛应用，主要通过专家经验识别架构风险，主观性强；定量性能模型依赖精确的输入参数，难以应对动态变化需求
（1）系统多维度指标的科学权衡：通过构建直觉模糊相对重要性判断矩阵，更便于进行多方案比较，并且通过直觉模糊数的隶属度、非隶属度和犹豫度，能够从多个维度验证评价结果的合理性，避免单一指标主导决策；同时将复杂系统架构方案评价指标进行层次化结构分解，分解为一级指标、二级指标、三级指标等等，通过逐级比较和权重分配，可减少因指标耦合导致的评价偏差。
Smart Images
Figure CN122903785A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of system architecture evaluation technology, specifically relating to a system architecture scheme evaluation method based on an improved intuitionistic fuzzy hierarchical analysis method. Background Technology
[0002] System architecture evaluation, conducted during the system design phase, compares multiple candidate architectures to select the optimal solution, making it a core step in the design of complex information systems. Through evaluation, it's possible to scientifically determine which architecture best meets business requirements such as system performance, scalability, and security. This is the primary goal of aligning system design with business needs, preventing subjective judgment biases that could lead to a system failing to meet requirements after construction. Secondly, optimizing cost and resource allocation is another core value of evaluation; a well-designed architecture balances development costs, operational costs, and business benefits. In summary, system architecture evaluation is a crucial guarantee for ensuring that complex information system designs meet business requirements. Scientific evaluation methods allow for problem identification during the design phase and the selection of the most suitable solution, avoiding the reactive situation of discovering deficiencies only after deployment and laying the foundation for successful system implementation.
[0003] Commonly used methods for evaluating system architecture solutions include the Analytic Hierarchy Process (AHP), model-based architecture analysis (such as Architecture Trade-off Analysis and Model-Analog-Model-Modular), and quantitative performance modeling (such as queuing network models and discrete event simulations). These methods have proven effective in specific scenarios. However, when directly applied to the evaluation of complex information system architecture solutions, these methods still have shortcomings: for example, while the AHP has the advantages of being structured and allowing for multi-objective trade-offs in system architecture evaluation, its subjectivity, static nature, and lack of scalability limit its application in dynamic and complex information systems; ATAM is widely used in requirement conflict analysis and non-functional requirement assessment, but it mainly relies on expert experience to identify architectural risks, making it highly subjective; and quantitative performance models depend on precise input parameters, making it difficult to cope with dynamically changing requirements.
[0004] Specifically, the existing technology mainly has the following two technical problems: (1) The trade-off between multi-dimensional indicators of the system.
[0005] The evaluation of complex information system architecture solutions needs to consider multiple dimensions such as performance, scalability, reliability, security, and cost, but these indicators often conflict. How to quantify the weight of each indicator and how to balance different objectives are easily influenced by perception, and subjective weight allocation may lead to biased results.
[0006] (2) The problem of dynamic adaptability evaluation of complex information systems.
[0007] In system evaluation, it is common for the demand for a certain capability of a system to suddenly increase due to external needs, technological changes, internal state changes, etc. How to make system evaluation have dynamic and adaptive characteristics is the key to enabling the widespread application of system evaluation methods. Summary of the Invention
[0008] To address the challenges of conflicting and balancing multi-dimensional indicators and the difficulty of dynamic adaptive evaluation in system architecture scheme evaluation, this invention proposes a system architecture scheme evaluation method based on an improved intuitionistic fuzzy hierarchical analysis (AHP). This method, building upon AHP, introduces intuitionistic fuzzy numbers and relative importance scales to construct an intuitionistic fuzzy relative importance judgment matrix. It then determines the initial weights of this matrix and, combined with reinforcement learning, designs a state space, action space, and reward function for training a reinforcement learning model. The weights of the intuitionistic fuzzy judgment matrix are dynamically adjusted, and finally, a weighted average is calculated to compare the comprehensive scores of each system architecture scheme, thereby improving the scientific rigor and rationality of selecting complex information system architecture schemes.
[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a system architecture evaluation method based on an improved intuitionistic fuzzy hierarchical analysis method, the method comprising: Determine the top-level indicators for evaluating system architecture solutions, decompose the top-level indicators into a hierarchical structure, and construct a multi-level evaluation indicator system for system architecture solutions. Based on the relative importance of indicators at the same membership level, an index relative importance scaling table based on intuitionistic fuzzy numbers is constructed. Based on the index relative importance scaling table based on intuitionistic fuzzy numbers, the relative importance between indices at the same membership level is determined, so as to construct the intuitionistic fuzzy relative importance judgment matrix for each membership level. The intuitionistic fuzzy relative importance judgment matrix is converted into a fuzzy priority relation matrix, and the intuitionistic fuzzy weights of each level index are determined based on the fuzzy priority relation matrix of each membership level. Reinforcement learning is used to dynamically adjust intuitionistic fuzzy weights based on the needs of changing business operations. Using the intuitive fuzzy weight of the top-level indicator as the comprehensive fuzzy weight, the comprehensive fuzzy weight of the upper-level indicator is multiplied by the intuitive fuzzy weight of the lower-level indicator layer by layer to obtain the comprehensive fuzzy weight of the lower-level indicator, until the comprehensive fuzzy weight of the bottom-level indicator is obtained. Obtain the index values of each underlying indicator for the system architecture scheme to be evaluated, and sum the index values with the corresponding comprehensive fuzzy weights to obtain the comprehensive score of the system architecture scheme to be evaluated.
[0010] Furthermore, the method also includes: Obtain the comprehensive score of all system architecture schemes to be evaluated, and select the optimal system architecture scheme based on the comprehensive score.
[0011] Furthermore, the relative importance scale table of indicators based on intuitionistic fuzzy numbers includes membership degree μ and non-membership degree ν, which are used to represent the relative importance between two indicators: the more important the former indicator is relative to the latter indicator, the larger the value of membership degree μ, and the more important the latter indicator is relative to the former indicator, the larger the value of non-membership degree ν.
[0012] Furthermore, each element in the intuitionistic fuzzy relative importance judgment matrix is represented by a triple (μ, ν, π), where μ is the membership degree, representing the degree of recognition of the importance of the former indicator relative to the latter; ν is the non-membership degree, representing the degree of rejection; and π is the degree of hesitation, satisfying π=1-μ-ν, representing the degree of uncertainty.
[0013] Furthermore, the membership degree μ and the non-membership degree ν range from 0.5 to 0.9, and the hesitation degree π ranges from 0 to 0.1.
[0014] Furthermore, the feature is that the intuitionistic fuzzy relative importance judgment matrix is converted into a fuzzy priority relation matrix, and the intuitionistic fuzzy weights of each level index are determined based on the fuzzy priority relation matrix of each membership level, including: The intuitive fuzzy relative importance judgment matrix is converted into a fuzzy priority relation matrix using the formula r=μ / (μ+ν), where r is the fuzzy priority relation value; The fuzzy priority relation matrix is normalized to obtain the intuitive fuzzy weights of each level of indicators.
[0015] Furthermore, reinforcement learning is used to dynamically adjust the intuitionistic fuzzy weights based on the needs of changing business conditions, including: The relative importance judgment matrix of intuitionistic fuzziness is converted into a real matrix using a scoring function. The real matrix, the current intuitionistic fuzziness weight allocation, and the demand information of business changes are used as the state input of the reinforcement learning model. Define a discrete action space, where actions are used to increase or decrease the intuitive fuzzy weights of the indicators, and satisfy non-negativity and normalization constraints. Design a reward function that includes a positive reward for improved target metric scores and a penalty for decreased scores on non-target metric scores. The reinforcement learning model is trained, and the action is adjusted by selecting intuitionistic fuzzy weights according to the current policy. The intuitionistic fuzzy weights are updated, the adjusted comprehensive score and reward value are calculated, and the reinforcement learning model is updated. Training stops when the reward value stabilizes or reaches the preset number of iterations, resulting in an adjusted intuitive fuzzy weight that adapts to changing business needs.
[0016] Furthermore, the scoring function is: S = (μ-ν) / (μ+ν+π); The reward function is: R = R_target - R_penalty, where R_target is the positive reward, i.e., the improvement in the target metric score; and R_penalty is the penalty, R_penalty = λ × Σ. 非目标指标 (Old score of indicator - New score of indicator), where λ is the penalty coefficient.
[0017] In a second aspect, the present invention provides a computer device / apparatus / system, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.
[0018] Thirdly, the present invention provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the method described in the first aspect.
[0019] The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method provided by this invention can significantly improve the scientificity and rationality of complex system architecture selection. Compared with existing evaluation methods, it can achieve the following beneficial effects: (1) Scientific weighing of multi-dimensional indicators of the system: By constructing an intuitive fuzzy relative importance judgment matrix, it is easier to compare multiple schemes. Furthermore, by using the membership degree, non-membership degree, and hesitation degree of the intuitive fuzzy number, the rationality of the evaluation results can be verified from multiple dimensions, avoiding the dominance of a single indicator in decision-making. At the same time, the evaluation indicators of complex system architecture schemes are decomposed into hierarchical structures, such as first-level indicators, second-level indicators, third-level indicators, etc. Through hierarchical comparison and weight allocation, the evaluation bias caused by indicator coupling can be reduced.
[0020] (2) Dynamic adaptability and real-time optimization: Through reinforcement learning, the intuitive fuzzy weights are automatically adjusted according to the needs of business changes, quickly responding to changes in business needs during the system design phase, optimizing evaluation results, and ensuring that the evaluation results are consistent with the current goals. Attached Figure Description
[0021] Figure 1 This is a flowchart of a system architecture evaluation method based on an improved intuitionistic fuzzy hierarchical analysis method according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the technique of dynamically adjusting the weights of an intuitionistic fuzzy matrix based on reinforcement learning, according to an embodiment of the present invention. Detailed Implementation
[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments provided by this invention without inventive effort are within the scope of protection of this invention.
[0023] Obviously, the accompanying drawings described below are merely some examples or embodiments of the present invention. Those skilled in the art can apply the present invention to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this invention, modifications to design, manufacturing, or production based on the technical content disclosed in this invention are merely conventional technical means and should not be construed as insufficient disclosure of the present invention.
[0024] In this invention, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this invention may be combined with other embodiments without conflict.
[0025] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "a," "an," "an," "the," and similar words used in this invention do not indicate quantity limitation and may indicate singular or plural. The terms "comprising," "including," "having," and any variations thereof used in this invention are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms "connected," "linked," "coupled," and similar words used in this invention are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "A plurality" used in this invention refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships may exist; for example, "A and / or B" can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects have an "or" relationship. The terms "first," "second," and "third" used in this invention are merely to distinguish similar objects and do not represent a specific ordering of the objects.
[0026] To address the challenges of conflicting and balancing multi-dimensional indicators and the difficulty of dynamic adaptive evaluation in system architecture scheme evaluation, this invention proposes a system architecture scheme evaluation method based on an improved intuitionistic fuzzy hierarchical analysis (AHP). This method, building upon AHP, introduces intuitionistic fuzzy numbers and relative importance scales to construct an intuitionistic fuzzy relative importance judgment matrix. It then determines the initial weights of this matrix and, combined with reinforcement learning, designs a state space, action space, and reward function for training a reinforcement learning model. The weights of the intuitionistic fuzzy judgment matrix are dynamically adjusted, and finally, a weighted average is calculated to compare the comprehensive scores of each system architecture scheme, thereby improving the scientific rigor and rationality of selecting complex information system architecture schemes.
[0027] (1) Introduce intuitionistic fuzzy numbers and relative importance scales to construct intuitionistic fuzzy relative importance judgments and deal with the conflict of multidimensional indicators and subjective weight bias in the evaluation of system architecture schemes.
[0028] First, clarify the core objective of the system architecture evaluation (i.e., primary indicator C). i (e.g., performance, cost, security, reliability, etc.) The core objectives are decomposed into a hierarchical structure (e.g., decomposed into secondary indicators C). ijDomain experts were invited to make intuitive fuzzy relative importance judgments on the decomposed indicators. By introducing intuitive fuzzy numbers (membership degree μ, non-membership degree ν, hesitation degree π), experts were allowed to express their judgments using triples (μ, ν, π). Membership degree μ represents the expert's degree of acceptance of the indicator's relative importance, non-membership degree ν represents the expert's degree of rejection of the indicator's relative importance, and hesitation degree π... π =1- μ - ν To represent the degree of uncertainty among experts, an intuitionistic fuzzy relative importance judgment matrix is constructed based on expert ratings to accurately quantify experts' fuzzy judgments on indicators; then, r is used... ij =μ ij / (ν ij +μ ij This is converted into a fuzzy priority relation matrix, and normalized weights are calculated to obtain the intuitive fuzzy weight values of each level of indicators; based on the comprehensive weight W... ij =w i ×w ij The comprehensive fuzzy weights of the lowest-level indicators are calculated; combined with the specific assignment and normalization of the lowest-level indicators, the indicator values of the lowest-level indicators are obtained; finally, the lowest-level indicator values and their comprehensive fuzzy weights are weighted and summed (by multiplying the indicator scores of each scheme by their weights), and the comprehensive score of each scheme is calculated. The schemes are then sorted from highest to lowest comprehensive score, and the optimal scheme is selected. Figure 1 As shown.
[0029] (2) By combining reinforcement learning, the weights of the intuition fuzzy judgment matrix are dynamically adjusted to solve the problem of dynamic adaptability in the evaluation of complex system architecture schemes caused by changes in business requirements during the system design phase.
[0030] By combining reinforcement learning training models, intuitionistic fuzzy weights are automatically adjusted to quickly respond to changes in requirements. First, a state space design is performed, converting the intuitionistic fuzzy relative importance judgment matrix into real numbers using a scoring function (scoring function S=(μ-ν) / (μ+ν+π)). The real numbers of the intuitionistic fuzzy relative importance judgment matrix, weight allocation, and performance changes serve as inputs to the reinforcement learning model (taking performance improvement as an example). Second, a discrete action space design is performed, defining adjustments to the weights, such as increasing performance weights or decreasing cost weights, and defining action constraints to ensure the weights satisfy non-negativity and normalization conditions. Finally, a reward function is designed, such as a performance improvement reward R. 性能 =Performance Score new -Performance Score old Add penalties for decreased scores on other metrics, such as R. 惩罚 =λ ∑ i不等于性能 (Score old -Score new The final reward is R=R性能 -R 惩罚 Next, the reinforcement learning model is trained. The adjusted weights are updated in the intuitionistic fuzzy weight matrix, and the reward is calculated based on the weight changes and updated in the reinforcement learning model. Training stops when the reward function stabilizes or reaches the preset number of iterations. Figure 2 As shown.
[0031] Therefore, the system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method provided by this invention can significantly improve the scientificity and rationality of complex system architecture selection. Compared with existing evaluation methods, it can achieve the following beneficial effects: (1) Scientific weighing of multi-dimensional indicators of the system: By constructing an intuitive fuzzy relative importance judgment matrix, it is easier to compare multiple schemes. Furthermore, by using the membership degree, non-membership degree, and hesitation degree of the intuitive fuzzy number, the rationality of the evaluation results can be verified from multiple dimensions, avoiding the dominance of a single indicator in decision-making. At the same time, the evaluation indicators of complex system architecture schemes are decomposed into hierarchical structures, such as first-level indicators, second-level indicators, third-level indicators, etc. Through hierarchical comparison and weight allocation, the evaluation bias caused by indicator coupling can be reduced.
[0032] (2) Dynamic adaptability and real-time optimization: Through reinforcement learning, the intuitive fuzzy weights are automatically adjusted according to the needs of business changes, quickly responding to changes in business needs during the system design phase, optimizing evaluation results, and ensuring that the evaluation results are consistent with the current goals.
[0033] Specifically, the technical process of the complex system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method in this invention is as follows: Figure 1 As shown, the specific implementation steps for this example are as follows: (1) Establish a system architecture scheme indicator evaluation system.
[0034] Evaluating a system architecture solution requires considering many indicators, and these indicators are hierarchically divided to obtain an indicator system for evaluating the system architecture.
[0035] (2) Construct a matrix to judge the relative importance of indicators.
[0036] First, based on the relative importance of indicators at the same membership level, a relative importance scale table based on intuitionistic fuzziness is constructed. The scale table is shown in Table 1. The scale can reflect the relative importance of two indicators. The more important the former is than the latter, the larger the corresponding membership value; the more important the latter is than the former, the larger the corresponding non-membership value.
[0037] Table 1. Relative Importance Scaling of Indicators Based on Intuitive Fuzzy Numbers
[0038] Based on the relative importance scale in the table above, an intuitionistic fuzzy judgment matrix is constructed. Considering the consistency issue of the judgment matrix, the range of hesitation values is [0, 0.1]. The relative importance of indicators at different membership levels is shown in Tables 2 to 6.
[0039] Table 2. Relative Importance of Indicators Belonging to System Architecture
[0040] Table 3. Relative Importance of Indicators Belonging to Ability C1
[0041] Table 4. Relative Importance of Indicators Belonging to Capability C2
[0042] Table 5. Relative Importance of Indicators Belonging to Ability C3
[0043] Table 6. Relative Importance of Indicators Belonging to Ability C4
[0044] (3) Determine the weights of the intuition fuzzy relative importance judgment matrix.
[0045] Based on the relative importance table of indicators in Table 2, we first use r ij =μ ij / (ν ij +μ ij Convert it into a fuzzy priority relation matrix, and calculate the fuzzy priority relation matrix in Table 2. Then, normalized weights are calculated: Each row and are respectively The sum is The normalized weights are: .
[0046] Similarly, the weight values of the indicators at each level in Table 2-6 are shown in Table 7.
[0047] Table 7 Weighting Results of Indicators at Different Levels
[0048] (4) Apply the operation rules.
[0049] Based on the comprehensive fuzzy weight W ij =w i ×w ij The comprehensive fuzzy weights of the 15 lowest-level indicators are shown in Table 8.
[0050] Table 8. Comprehensive Fuzzy Weight Table of Bottom-Level Indicators
[0051] It should be noted that when calculating the comprehensive fuzzy weight, starting from the top layer, the comprehensive fuzzy weight of the previous layer indicator is multiplied by the weight of the next layer indicator to obtain the comprehensive fuzzy weight of the next layer indicator. This process is repeated layer by layer until the comprehensive fuzzy weight of the bottom layer indicator is obtained. In particular, the intuitive fuzzy weight of the top layer indicator is the comprehensive fuzzy weight of the top layer indicator.
[0052] (5) Generate a score.
[0053] First, specific capability values are calculated for the underlying metrics (secondary metrics in this embodiment) of different architecture schemes to be evaluated. Then, the metric values are normalized to obtain the example shown in Table 9.
[0054] Table 9. Table of underlying indicator values for each system architecture
[0055] By weighting and summing the underlying indicator values given in Table 9 and the comprehensive weights of the underlying indicators listed in Table 8, the final performance values of the four different system architectures can be obtained: The performance value of system architecture A is: E1 = 0.58728; The performance value of system architecture B is: E2 = 0.66312; The system architecture performance value is: E3 = 0.82187; The system architecture D performance value is: E4 = 0.87809; It can be seen that the proposed system architectures C and D both show significant performance improvements, with system architecture D exhibiting the best performance.
[0056] (6) Dynamic adjustment of weights in intuitive fuzzy judgment matrix.
[0057] First, the state space is designed by converting the intuitionistic fuzzy relative importance judgment matrix into real numbers using a scoring function (score function S=(μ-ν) / (μ+ν+π)). The real numbers of the intuitionistic fuzzy relative importance judgment matrix, the weight assignment, and the performance changes are used as inputs to the reinforcement learning model. Second, the discrete action space is designed by defining adjustments to the weights, such as increasing performance weights or decreasing cost weights, and defining action constraints to ensure that the weights satisfy non-negativity and normalization conditions. Finally, the reward function is designed, such as the performance improvement reward R. 性能 =Performance Score new -Performance Score old Add penalties for decreased scores on other metrics, such as R. 惩罚 =λ×∑ i不等于性能 (Score old -Scorenew The final reward is R=R 性能 -R 惩罚 Next, the reinforcement learning model is trained, the adjusted weights are updated into the intuitionistic fuzzy weight matrix, the reward is calculated based on the weight changes and updated into the reinforcement learning model, and training is stopped when the reward function is stable or reaches the preset number of iterations. Based on the updated matrix weights, step (4) is repeated to apply the operation rules and (5) to generate scores, so as to obtain the scores of each system architecture scheme after weight adjustment.
[0058] Example of adjustment process: enter: Initial weights: w = [0.3, 0.5, 0.2]; Intuitive fuzzy matrix (3×3): ; Triggering condition: User feedback that "performance requirements have increased".
[0059] step: 1. Detect changes in demand: Identify signals of "increased performance requirements".
[0060] 2. Perform the action: The reinforcement learning strategy selects the action of "increasing performance weights"; Adjust weights: w 性能 ←0.3+0.1=0.4, and the other weights are normalized to [0.4, 0.429, 0.171].
[0061] 3. Calculate the reward: Performance score: Recalculated using the intuitionistic fuzzy weighted average method, score from S 性能,old =0.65 increased to S 性能,new =0.72; Cost score: decreased from 0.55 to 0.48; Reward: R = (0.72 - 0.65) - 0.5 * (0.55 - 0.48) = 0.07 - 0.035 = 0.035.
[0062] 4. Update strategy: Reinforcement learning algorithms (such as Q-learning or DQN) adjust the strategy by updating weights based on rewards.
[0063] 5. Convergence judgment: If the rewards are stable for several consecutive rounds, stop adjusting.
[0064] In summary, this invention proposes a system architecture evaluation method based on an improved intuitionistic fuzzy hierarchical analysis (AHP). It combines the advantages of intuitionistic fuzzy set theory and AHP. Intuitionistic fuzzy set theory allows experts to express uncertainty and hesitation in their evaluations. Compared to the 1-9 scale of traditional AHP, intuitionistic fuzzy numbers and relative importance scales more accurately reflect the subjective cognitive differences of experts, avoiding the simplification of fuzziness by linear scales. Simultaneously, the weighted fusion of fuzzy numbers reduces the excessive influence of a single expert opinion on the results, minimizing subjective bias. Furthermore, the introduction of a reinforcement learning model to dynamically adjust the relative importance judgment weights of the intuitionistic fuzzy matrix enables rapid response to changes in business requirements, providing quantifiable decision-making basis and facilitating horizontal comparison of different architecture schemes, thereby improving the scientific rigor and rationality of complex information system architecture selection.
[0065] Furthermore, the present invention also provides a computer device / apparatus / system, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method.
[0066] The present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.
[0067] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, systems, and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or block diagrams.
[0068] These computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more blocks in a block diagram.
[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, thereby providing steps for implementing the functions specified in one or more flowcharts and / or one or more blocks in a block diagram.
[0070] It should be noted that, depending on the implementation needs, the various steps / components described in this invention can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.
[0071] Those skilled in the art will readily understand that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A system architecture evaluation method based on an improved intuitionistic fuzzy hierarchical analysis method, characterized in that, The method includes: Determine the top-level indicators for evaluating system architecture solutions, decompose the top-level indicators into a hierarchical structure, and construct a multi-level evaluation indicator system for system architecture solutions. Based on the relative importance of indicators at the same membership level, an index relative importance scaling table based on intuitionistic fuzzy numbers is constructed. Based on the index relative importance scaling table based on intuitionistic fuzzy numbers, the relative importance between indices at the same membership level is determined, so as to construct the intuitionistic fuzzy relative importance judgment matrix for each membership level. The intuitionistic fuzzy relative importance judgment matrix is converted into a fuzzy priority relation matrix, and the intuitionistic fuzzy weights of each level index are determined based on the fuzzy priority relation matrix of each membership level. Reinforcement learning is used to dynamically adjust intuitionistic fuzzy weights based on the needs of changing business operations. Using the intuitive fuzzy weight of the top-level indicator as the comprehensive fuzzy weight, the comprehensive fuzzy weight of the upper-level indicator is multiplied by the intuitive fuzzy weight of the lower-level indicator layer by layer to obtain the comprehensive fuzzy weight of the lower-level indicator, until the comprehensive fuzzy weight of the bottom-level indicator is obtained. Obtain the index values of each underlying indicator for the system architecture scheme to be evaluated, and sum the index values with the corresponding comprehensive fuzzy weights to obtain the comprehensive score of the system architecture scheme to be evaluated.
2. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 1, characterized in that, The method also includes: Obtain the comprehensive score of all system architecture schemes to be evaluated, and select the optimal system architecture scheme based on the comprehensive score.
3. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 1, characterized in that, The relative importance scale table of indicators based on intuitionistic fuzzy numbers includes membership degree μ and non-membership degree ν, which are used to represent the relative importance between two indicators: the more important the former indicator is relative to the latter indicator, the larger the value of membership degree μ; the more important the latter indicator is relative to the former indicator, the larger the value of non-membership degree ν.
4. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 1, characterized in that, Each element in the intuitionistic fuzzy relative importance judgment matrix is represented by a triple (μ, ν, π), where μ is the membership degree, representing the degree of recognition of the importance of the former indicator relative to the latter indicator; ν represents the degree of non-membership, indicating the degree of negation; π represents the degree of hesitation, which satisfies π = 1 - μ - ν, indicating the degree of uncertainty.
5. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 4, characterized in that, The membership degree μ and the non-membership degree ν range from 0.5 to 0.9, and the hesitation degree π ranges from 0 to 0.
1.
6. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to any one of claims 3 to 5, characterized in that, The intuitionistic fuzzy relative importance judgment matrix is transformed into a fuzzy priority relation matrix. Based on the fuzzy priority relation matrix of each membership level, the intuitionistic fuzzy weights of each level's indicators are determined, including: The intuitive fuzzy relative importance judgment matrix is converted into a fuzzy priority relation matrix using the formula r=μ / (μ+ν), where r is the fuzzy priority relation value; The fuzzy priority relation matrix is normalized to obtain the intuitive fuzzy weights of each level of indicators.
7. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 4 or 5, characterized in that, Using reinforcement learning to dynamically adjust intuitionistic fuzzy weights based on changing business needs, including: The relative importance judgment matrix of intuitionistic fuzziness is converted into a real matrix using a scoring function. The real matrix, the current intuitionistic fuzziness weight allocation, and the demand information of business changes are used as the state input of the reinforcement learning model. Define a discrete action space, where actions are used to increase or decrease the intuitive fuzzy weights of the indicators, and satisfy non-negativity and normalization constraints. Design a reward function that includes a positive reward for improved target metric scores and a penalty for decreased scores on non-target metric scores. The reinforcement learning model is trained, and the action is adjusted by selecting intuitionistic fuzzy weights according to the current policy. The intuitionistic fuzzy weights are updated, the adjusted comprehensive score and reward value are calculated, and the reinforcement learning model is updated. Training stops when the reward value stabilizes or reaches the preset number of iterations, resulting in an adjusted intuitive fuzzy weight that adapts to changing business needs.
8. The system architecture evaluation method based on the improved intuitionistic fuzzy hierarchical analysis method according to claim 7, characterized in that, The scoring function is: S = (μ-ν) / (μ+ν+π); The reward function is: R = R_target - R_penalty, where R_target is the positive reward, i.e., the improvement in the target metric score; and R_penalty is the penalty, R_penalty = λ × Σ. 非目标指标 (Old score of indicator - new score of indicator), where λ is the penalty coefficient.
9. A computer device / equipment / system, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.
10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 8.