Improved entropy and VIKOR mixed unmanned aerial vehicle model selection method and system

Abstract

The invention relates to the technical field of target decision, and discloses an improved entropy and VIKOR hybrid unmanned aerial vehicle model selection method and system, which converts a qualitative index into a weight index and builds a standardized forward matrix together with a quantitative index, overcomes the defect that qualitative factors are difficult to objectively measure, and improves the accuracy of model selection. The model selection decision can comprehensively cover key factors influencing the unmanned aerial vehicle; by improving the normalization mode of the negative index, the redundancy step of forward normalization and then standardization is omitted, and the data processing amount and complexity are reduced while the data comparability is ensured; according to the method, index weights are distributed objectively by adopting an entropy method, a group utility value, an individual regret value and a compromise index are calculated in combination with a VIKOR method, and double verification of acceptable advantages and acceptable stability is performed on a sorting result, so that objectivity and scientificity of an unmanned aerial vehicle model selection result are ensured, and the unmanned aerial vehicle model selection efficiency is improved. The unmanned aerial vehicle scheme capable of achieving the optimal balance between the optimal overall performance and no obvious short plate can be reliably identified.

Description

Technical Field

[0001] This invention relates to the field of target decision-making, and in particular to an improved method and system for selecting unmanned aerial vehicles (UAVs) that combines entropy and VIKOR. Background Technology

[0002] In recent years, with the rapid development and widespread application of drone technology, a wide variety of drone products with different functions have emerged in the market.

[0003] Currently, the selection and procurement of drones mainly relies on fragmented information obtained from on-site inspections and manufacturer presentations, and decisions are made based on personal experience and subjective judgment. There is a lack of systematic planning of drone market performance parameters. In the selection of drone parameters, the focus is often only on meeting immediate data production indicators and specific terrain conditions, without comprehensively considering the application scenario expansion throughout the equipment's life cycle, performance redundancy, and long-term cost-effectiveness. As a result, the selected drone models often only match some of the current explicit needs, and after actual deployment, their deficiencies in mission adaptability, performance redundancy, or maintenance economy are gradually exposed. This makes the drones ultimately purchased not the most suitable in essence, and there is always a mismatch between drones and real business needs. Summary of the Invention

[0004] The purpose of this invention is to address the technical problem that the subjective selection of existing drones leads to drones not being the most suitable, and to provide an improved drone selection method and system that combines entropy value and VIKOR.

[0005] To achieve the above-mentioned objectives, the embodiments of the present invention provide the following technical solutions:

[0006] An improved method for selecting drones that combines entropy and VIKOR, comprising the following sub-steps:

[0007] Select the drone takeoff type based on business needs, select multiple options based on the drone takeoff type, and determine the parameter configuration of each drone based on the options;

[0008] The parameters of the UAV are quantified to form a set of qualitative indicators and a set of quantitative indicators.

[0009] Construct a standardized positive matrix from the set of quantitative indicators and the data within the set of quantitative indicators;

[0010] The entropy weights are determined for the standardized positive matrix using the entropy method, resulting in a weight vector.

[0011] The weighted normalized matrix is ​​obtained by using the weight vector and the standardized positive matrix. The weighted normalized matrix is ​​then used to obtain the group utility value and individual regret value for each option using VIKOR. The combination of the group utility value and individual regret value yields the compromise index.

[0012] Sort the compromise indices of each solution from smallest to largest. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, then select the solution corresponding to the first compromise index. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, then select the compromise solution as the final solution. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, then select the candidate solution as the final solution.

[0013] To address the issue of subjective drone selection leading to drones not being the most suitable, this application transforms qualitative indicators (such as operational capabilities and after-sales service levels) into calculable data. It also employs improved normalization to directly construct a standardized positive matrix for negative indicators (such as market price and deployment time), simplifying the process while maintaining data accuracy. The entropy method is used to objectively determine indicator weights, and the VIKOR method is used to calculate the group utility value, individual regret value, and compromise index of each option. Finally, through a dual verification mechanism of acceptable advantages and acceptable stability, the optimal model with balanced overall performance, no significant shortcomings, and best suited for long-term business development is identified from multiple options. This solves the problem of subjective drone selection leading to drones not being the most suitable. Furthermore, by establishing a complete quantitative mapping and decision-making rule from requirements to parameters, this application unifies fragmented performance parameters and long-term business needs within an objective calculation framework, enabling the final selection result to systematically balance various performance indicators and be tailored to the best drone model.

[0014] Compared with existing technologies, the beneficial effects of this application are as follows: By transforming qualitative indicators into standardized weighted indicators and forming a standardized positive matrix together with quantitative indicators, the deficiency of qualitative factors being difficult to objectively measure is overcome, enabling the selection decision to comprehensively cover key factors affecting the entire life cycle of UAV use; by improving the normalization processing method of negative indicators, the redundant steps of first positiveizing and then standardizing are eliminated, directly forming a standardized matrix, which reduces the amount and complexity of data processing while ensuring data comparability and improving computational efficiency; by using the entropy method to objectively allocate indicator weights and combining the VIKOR method to calculate the group utility value, individual regret value, and compromise index, the ranking results are double-verified for acceptable advantages and acceptable stability, ensuring the objectivity and scientific nature of UAV selection results, and enhancing the robustness and interpretability of the decision, which can reliably identify the most suitable UAV solution that achieves the best balance between optimal overall performance and no significant shortcomings.

[0015] Furthermore, an improved method for selecting drones that combines entropy and VIKOR, wherein determining the parameter configuration of each drone through the scheme includes the following sub-steps:

[0016] The take-off type of the drone is selected according to different business needs. The take-off type of the drone includes vertical take-off drones, fixed-wing drones, and compound-wing drones. Multiple drones are determined as a plan based on the selected take-off type.

[0017] The scheme determines the parameter configuration for each UAV, which includes a set of flight requirement indicators, a set of UAV parameter indicators, a set of environmental requirement indicators, and a set of other factor indicators.

[0018] In the aforementioned solutions, existing technologies typically rely on isolated, fragmented lists of performance parameters or empirically typical configurations to determine specific parameter configurations for UAV selection. This lack of a systematic mapping and structured decomposition from business objectives to technical specifications leads to a one-sided and haphazard parameter configuration process. It fails to ensure that the selected parameter set can fully and balancedly support mission objectives and adapt to environmental constraints and other non-technical factors throughout the entire lifecycle, thus creating the potential for selection bias. This application addresses the technical problem of the lack of systematic and complete parameter configuration by introducing structured parameter configuration sub-steps: selecting the takeoff type based on business needs and determining the flight requirement index set, UAV parameter index set, environmental requirement index set, and other factor index set for each UAV. Specifically, firstly, based on the inherent characteristics of different business needs such as image mapping and patrol monitoring, a primary selection is made among UAV takeoff types such as vertical takeoff, fixed-wing, and compound-wing, ensuring that the platform's basic characteristics match the core mission mode and avoiding fundamental mismatch. Then, by systematically constructing and filling flight requirement indicator sets, UAV parameter indicator sets, environmental requirement indicator sets, and other factor indicator sets, abstract business objectives are decomposed layer by layer and transformed into specific, measurable technical parameters covering all dimensions of technical performance, external constraints, and operational support. This lays a complete and consistent data foundation for subsequent quantitative analysis and decision-making. This application constructs a systematic parameter configuration link from top-level mission to bottom-level specifications by defining the type based on business needs, the solution based on the type, and the structured parameter set based on the solution. This not only overcomes the shortcomings of fragmented and one-sided parameter selection in traditional methods but also ensures that the parameter configuration of each alternative solution is a complete and structured description of its comprehensive technical capabilities and applicable boundaries.

[0019] Furthermore, an improved method for selecting unmanned aerial vehicles (UAVs) that combines entropy and VIKOR is proposed, wherein the flight requirement index set includes minimum personnel configuration, supporting equipment, and operating conditions.

[0020] The set of parameters for the UAV includes deployment time, cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, size, and vertical takeoff height.

[0021] The environmental requirements index set includes maximum takeoff altitude, minimum ambient temperature, maximum ambient temperature, and wind resistance level;

[0022] The other factor indicators include market price, control capability, insurance payout indicators, after-sales service level, and application scenarios.

[0023] In the aforementioned solutions, existing technologies, when constructing UAV selection and evaluation systems, typically list general performance parameters or focus only on the limited technical indicators of the flight platform itself. They lack a systematic and structured attribution and classification of various factors affecting the equipment's lifecycle applicability and economy. This results in a one-sided parameter set construction and missing dimensions. Consequently, the evaluation system cannot accurately depict the complexity of mission requirements, the severity of environmental constraints, and the operational support dependencies that UAVs must address in actual business deployments. This leads to a disconnect between the selection criteria and real-world application scenarios, making it difficult to select truly suitable models. This application addresses this by constructing a structured parameter system consisting of four dimensions: a flight requirement indicator set, a UAV parameter indicator set, an environmental requirement indicator set, and other factor indicator sets. Specifically, the flight requirement indicator set (such as minimum personnel configuration and operating conditions) directly addresses the organizational and resource constraints of business execution; the UAV parameter indicator set (such as cruising speed, endurance, and maximum takeoff weight) comprehensively describes the core technical capabilities of the UAV; the environmental requirement indicator set (such as wind resistance level and operating temperature) clarifies the physical deployment boundaries of the equipment; and the other factor indicator set (such as market price, after-sales service, and application scenarios) covers operational and market factors affecting long-term holding costs and efficiency. This solves the technical problem of inapplicable UAV selection due to vague parameters when constructing a UAV selection evaluation system. This application systematically deconstructs and categorizes UAV selection parameters into the above four interrelated yet distinct indicator sets, achieving a leap from single performance to structured evaluation dimensions. This ensures that the evaluation of each alternative fully covers all key decision factors from technical feasibility to economic rationality, and from current operations to long-term maintenance, thus providing a solid, comprehensive, and highly scenario-based data foundation for subsequent quantitative modeling and scientific decision-making, fundamentally avoiding selection bias caused by missing evaluation dimensions.

[0024] Furthermore, an improved method for selecting drones that combines entropy and VIKOR, wherein the formation of qualitative and quantitative indicator sets includes the following sub-steps:

[0025] The indicators that cannot be measured by specific numerical values ​​in the indicator set are formed into a qualitative indicator set, and the qualitative indicators in the qualitative indicator set are converted into weighted indicators.

[0026] The quantitative indicators in the indicator set are integrated to form a quantitative indicator set, and the quantitative indicator set is normalized.

[0027] The quantitative indicator set includes a positive indicator set and a negative indicator set;

[0028] The set of qualitative indicators includes supporting equipment, operating conditions, control capabilities, after-sales service level, and application scenarios;

[0029] The positive indicator set includes cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, vertical takeoff height, maximum takeoff altitude, maximum ambient temperature, wind resistance rating, and insurance payout indicators.

[0030] The set of negative indicators includes minimum staffing, deployment time, minimum ambient temperature, market price, and size.

[0031] In the aforementioned solutions, existing technologies typically employ a single, general quantitative approach for all indicators when constructing the data foundation for selection. This leads to either the neglect of qualitative indicators that cannot be directly measured with specific numerical values, resulting in a lack of evaluation dimensions, or a heavy reliance on subjective scoring based on the personal experience of experts or users. Consequently, the quantitative results vary from person to person and from time to time, lacking objective standards and consistency, and significantly undermining the scientific rigor and repeatability of the selection model. This application addresses the core technical problem of strong subjectivity and unclear logic in selection data processing by explicitly establishing qualitative and quantitative indicator sets, and further distinguishing the quantitative indicators into positive and negative indicator sets. Specifically, clear and operable grading and value assignment rules were established for the qualitative indicator set (such as operating conditions and application scenarios), objectively transforming them into comparable weighted indicator values. This completely replaced subjective experience-based scoring. By pre-dividing the quantitative indicator set into a positive indicator set where larger values ​​are better (such as endurance and wind resistance) and a negative indicator set where smaller values ​​are better (such as deployment time and market price), the process paved the way for efficient and targeted normalization processing, ensuring the clarity and directness of the data processing logic. This application, through the implementation of rule-based assignment of qualitative indicators and refined classification of quantitative indicators, constructs objective, consistent, and standardized high-quality input from the data source, eliminating the arbitrariness and instability caused by relying on subjective experience-based scoring. Moreover, through clear indicator classification, a concise and efficient data processing flow is established for all subsequent calculation steps, thereby improving the overall reliability, interpretability, and objectivity and fairness of UAV selection and decision-making results.

[0032] Furthermore, an improved method for selecting UAVs that combines entropy and VIKOR, wherein the normalization of the quantitative indicator set includes the following steps:

[0033] The quantitative indicator set is divided into a positive indicator set and a negative indicator set. Let the positive vector indicator corresponding to the maximum value of the positive indicator be 1. The positive indicator set is then normalized into a positive vector indicator using the following formula:

[0034] ;

[0035] in, For the i-th positive vectorization index, For the i-th positive indicator, This represents the maximum value of the positive indicator.

[0036] Let the negative vectorization index corresponding to the minimum value of the negative index be 1. The formula for normalizing the set of negative indices into negative vectorization indices is:

[0037] ;

[0038] in, For the i-th negative vectorization index, For the i-th negative indicator, This represents the minimum value of the negative indicator.

[0039] In the aforementioned solutions, existing technologies, when normalizing quantitative indicator sets, cannot directly compare negative indicators with positive indicators under the same logic for normalization. This forces the introduction of an additional positive preprocessing step (such as taking the reciprocal or using the difference method). This two-step approach of first positive and then normalizing not only increases the complexity of data processing and computational steps, but also, during the mathematical transformation of the first positive transformation (especially nonlinear transformations), may alter or distort the inherent proportional relationships between the original data, introducing unnecessary precision loss. Consequently, the standardized data used for decision-making cannot fully and objectively reflect the actual situation of the original indicators. This application addresses the technical problems of redundancy and vulnerability to data precision in existing technologies when processing negative indicators by pre-dividing the quantitative indicator set into positive and negative indicator sets and applying a proprietary one-step normalization formula to each. Specifically, for the positive indicator set, entropy is used for efficient normalization; for the negative indicator set, an improved entropy is innovatively applied, thus simultaneously achieving the dual purpose of positive and normalization in a single step. This application eliminates the essential intermediate conversion step in existing technologies by implementing positive and negative indicator classification and innovative processing of negative indicators. This not only significantly simplifies the overall data processing flow and reduces the amount of computation, but also fundamentally avoids data distortion that may be caused by additional mathematical transformations. It ensures that the objective proportional relationship between the original values ​​of all quantitative indicators is strictly maintained during the process of converting them into standardized data, thereby providing a more accurate and comparable data foundation for subsequent objective weight calculation and scheme ranking.

[0040] Furthermore, an improved method for selecting UAVs that combines entropy and VIKOR is proposed, wherein a standardized positive matrix is ​​constructed using the following formula:

[0041] ;

[0042] in, To standardize the positive matrix, The evaluation index for the j-th column of the i-th row of the scheme;

[0043] The evaluation indicators include negative vectorized indicators, positive vectorized indicators, and weighted indicators.

[0044] In the aforementioned schemes, existing technologies often treat different types of indicators in a fragmented manner. Quantitative indicators may undergo decentralized normalization, while qualitative indicators are scored independently. This results in isolated evaluation data for each scheme across all dimensions, lacking a central dataset with consistent mathematical meaning. Consequently, the entire decision-making process lacks a single, objective, and verifiable data source, making systematic error tracing and consistency verification difficult. This application addresses the technical problems of inconsistent data foundations and fragmented computational sources in the decision-making process by explicitly defining and implementing the construction of a standardized positive matrix. Specifically, this application treats all indicators processed in the preliminary steps—including negative vectorized indicators transformed from negative indicators, positive vectorized indicators transformed from positive indicators, and weighted indicators transformed from qualitative indicators—as homogeneous evaluation indicators, strictly integrating them into a unified matrix expression according to the organization of schemes (rows) and evaluation indicators (columns). Each element of this matrix has been pre-converted into a standardized form where larger values ​​are considered better, thus ensuring that all evaluation dimensions are in a completely fair and comparable state before entering the core algorithm. This application, by explicitly defining the standardized positive matrix as the core data hub, establishes for the first time a single, authoritative, standardized data platform that aggregates all evaluation information in UAV selection decision-making. This changes the situation of data silos and fragmented calculation processes in traditional methods. It not only provides unique, pure, and homogeneous high-quality data input for subsequent entropy weight calculation and VIKOR multi-attribute ranking, thus ensuring the rigor of the entire mathematical model operation and the credibility of the results, but also makes the entire decision-making process highly traceable and repeatable.

[0045] Furthermore, an improved method for selecting drones by combining entropy and VIKOR, wherein the group utility value and individual regret value of each option are obtained using VIKOR, and the compromise index is obtained by combining the group utility value and individual regret value, includes the following sub-steps:

[0046] Multiply each column of the evaluation index in the standardized positive matrix by the entropy value weight corresponding to the weight vector to obtain the weighted normalized matrix;

[0047] Find the maximum and minimum values ​​of each column of evaluation indicators after normalization to form the ideal solution and the negative ideal solution;

[0048] Calculate the group utility value of each evaluation index of the UAV in each scheme compared to the ideal solution;

[0049] Calculate the individual regret value of each evaluation index for the drone in the worst-case scenario for each scheme;

[0050] The group utility value and individual regret value of each option are standardized, and the compromise index is calculated by combining the standardized values.

[0051] Through the above scheme, this application constructs a complete, closed-loop computational chain from data standardization and objective weighting to multi-attribute optimization decision-making. First, the standardized positive matrix is ​​analyzed using the entropy method. Based on the dispersion (information entropy) of each indicator data, the difference coefficient is objectively calculated and normalized to obtain the weight vector, thereby eliminating the subjective bias of manually setting weights and ensuring the fairness of the evaluation system. The weight vector is applied to the standardized matrix to obtain the weighted normalized matrix. Based on this, the VIKOR multi-criteria compromise ranking method is introduced: by defining the ideal solution and the negative ideal solution, the group utility value reflecting the overall approximation of the optimal level of the scheme and the individual regret value measuring the maximum local weakness of the scheme are calculated respectively. These two key values ​​are standardized and combined into a compromise index. This application successfully transforms the complex UAV selection problem into a comprehensive ranking problem of all alternatives. Through this improved entropy-VIKOR hybrid model, decision-makers can not only select the optimal solution based on the objective ranking of the compromise index, but also clearly reveal the balance between the overall advantages and developmental disadvantages of each solution through dual measures of group utility and individual regret. This provides a deeper insight beyond a single ranking for the final decision, making the selection result not only optimal, but also a robust solution that achieves the best balance between overall performance and risk aversion. This fundamentally meets the core requirements of long-term applicability and economy in UAV procurement.

[0052] Furthermore, an improved method for selecting UAVs that combines entropy and VIKOR is provided, wherein the criteria for determining acceptable advantages are as follows:

[0053] ;

[0054] in, The minimum value of all compromise indices. The second smallest value of all compromise indices, where m is the total number of rows in the scheme;

[0055] The criteria for determining acceptable stability are:

[0056] or ;

[0057] in, This represents the group utility value of the solution corresponding to the first compromise index. This represents the individual regret value for the solution corresponding to the first compromise index. To achieve the optimal group utility value, This represents the optimal individual regret value.

[0058] The existing technology, after ranking multi-attribute decision-making, typically selects based solely on the simple order of compromise indices, directly using the top-ranked solution as the final output. This fails to determine the statistical significance of the top-ranked solution's advantage. When the compromise indices of the first and second-ranked solutions are extremely close, firstly, the optimal position is very fragile; even minor data fluctuations or weight changes can lead to a reversal of the ranking, resulting in poor robustness of the selection outcome. Secondly, it lacks verification of the structural stability of the top-ranked solution itself, failing to identify whether the solution has serious shortcomings in certain key dimensions and wins solely based on its overall score. This could lead to the selected optimal solution being unusable in actual business due to a single unmet metric, resulting in high decision-making risk and ultimately, incompatibility in UAV selection. This application addresses the technical problems of fragile results and uncontrollable risks in traditional methods by introducing and rigorously defining two mathematical judgment conditions: acceptable advantage and acceptable stability. Specifically, the acceptable advantage condition is quantified, requiring the optimal solution's compromise index to maintain a clear minimum advantage threshold related to the total number of solutions compared to the second-best solution, thereby ensuring the statistical significance and stability of its leading position. Acceptable stability conditions are enforced through logical judgment, requiring that the top-ranked solution possesses at least one of the optimal values ​​among all solutions: either its group utility value or its individual regret value. This ensures that the solution either leads in overall benefit or has no significant major weakness, representing a balanced development or a solution with prominent advantages. This application elevates the final decision-making process from a simple ranking-selection to a rigorous ranking-verification-decision process by adding a quantitative threshold verification of acceptable advantage and a structural optimality verification of acceptable stability. This provides decision-makers with a clear and reasonable path for subsequent analysis, enhancing the method's practicality and flexibility.

[0059] Furthermore, an improved method for selecting UAVs that combines entropy and VIKOR is provided, wherein if the solution corresponding to the first compromise index does not meet the acceptable advantage, then the subsequent solutions corresponding to the first compromise index are integrated into a set of compromise solutions, as shown in the formula:

[0060] ;

[0061] in, For the k-th compromise solution, Let be the compromise index corresponding to the k-th compromise solution, and m be the total number of rows in the solution;

[0062] If the solution corresponding to the first compromise index satisfies acceptable advantage but not acceptable stability, then the candidate solution set is initialized. ,set up The process begins with a loop to determine the final set of candidate solutions for integration, based on the following criteria:

[0063] like and ,but Add to the candidate solution set. ;

[0064] in, The solution corresponding to the first compromise index. Let p be the compromise index corresponding to the p-th candidate solution. Let p be the p-th possible candidate solution. This is a set of candidate solutions.

[0065] The existing technology, in multi-attribute decision-making, often lacks clear and systematic rules to determine the subsequent choice when the top-ranked solution cannot be the unique optimal solution due to insignificant advantages or shortcomings. Common practices include simply recommending the top N options, but the value of N is arbitrary; or relying entirely on the decision-maker's subjective reassessment, leading to interruptions in the decision-making process and inconsistent standards. This lack of rules causes the final decision to revert to ambiguity and empiricism at critical moments, weakening the value of the entire quantitative selection model and failing to guarantee the mathematical rationality of the generated candidate set (e.g., whether the differences between solutions are significant). This application addresses the technical problem of unclear decision paths and unfounded candidate solution generation when the unique optimal solution is not found by strictly defining the mathematical construction rules for the compromise solution set and candidate solution set. Specifically, when the first solution does not meet the acceptable advantage requirement, this application defines a compromise solution set, which automatically categorizes all solutions whose compromise index difference with the first solution is less than a stability threshold into the compromise solution set, acknowledging that their differences are not significant. When the first solution meets the advantage condition but not the acceptable stability requirement, an algorithm that initializes the set and iteratively checks the difference between the next solution and the last solution in the current set, starting with the first solution. If the difference is less than the threshold, it is included, until the difference is greater than or equal to the threshold, thus forming a candidate solution set that may include the first solution and its closely related subsequent solutions. This application provides a clear and automated execution logic for the final branch of the decision-making process. All possible cases after the compromise index sorting are guided to output results with clear mathematical definitions, realizing full quantification and closed-loop management of the decision path.

[0066] An improved UAV selection system combining entropy and VIKOR is provided, comprising a UAV parameter configuration module, a UAV parameter quantification module, a UAV parameter unification module, a weight vector calculation module, a compromise index calculation module, and a UAV scheme selection module.

[0067] The drone parameter configuration module selects the drone takeoff type according to business needs, selects multiple schemes according to the drone takeoff type, and determines the parameter configuration of each drone through the scheme;

[0068] The UAV parameter quantification module quantifies the UAV parameter configuration, forming a qualitative index set and a quantitative index set; the UAV parameter unification module constructs a standardized positive matrix from the data in the quantitative index set and the quantitative index set.

[0069] The weight vector calculation module determines the entropy weights of the standardized positive matrix using the entropy method, thus obtaining the weight vector.

[0070] The compromise index calculation module obtains a weighted normalized matrix through a weight vector and a standardized positive matrix. The weighted normalized matrix uses VIKOR to obtain the group utility value and individual regret value of each solution. The combination of the group utility value and individual regret value obtains the compromise index.

[0071] The UAV solution selection module sorts the compromise index of each solution from smallest to largest. When the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, the solution corresponding to the first compromise index is selected. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, a compromise solution is selected as the final solution. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, a candidate solution is selected as the final solution. Attached Figure Description

[0072] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0073] Figure 1 This is a flowchart of an improved method for selecting drones that combines entropy and VIKOR.

[0074] Figure 2 A flowchart for quantifying the parameter configuration of a drone.

[0075] Figure 3 A flowchart for obtaining the compromise index using VIKOR.

[0076] Figure 4 This is a structural diagram of an improved UAV selection system that combines entropy and VIKOR. Detailed Implementation

[0077] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0078] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance, or suggesting any such actual relationship or order between these entities or operations. Additionally, the terms "connected," "linked," etc., can refer to a direct connection between elements or an indirect connection via other elements.

[0079] This invention is achieved through the following technical solutions, such as... Figure 1 As shown, an improved method for selecting drones that combines entropy and VIKOR includes the following steps:

[0080] S1: Select the drone takeoff type according to business needs, select a solution according to the drone takeoff type, and determine the parameter configuration of each drone based on the solution;

[0081] S11: Select the drone takeoff type according to different business needs. The drone takeoff type includes vertical takeoff drones, fixed-wing drones, and compound-wing drones. Determine multiple drones as a plan based on the selected drone takeoff type.

[0082] S12: Determine the parameter configuration for each UAV through the scheme. The parameter configuration includes a flight requirement index set, a UAV parameter index set, an environmental requirement index set, and a other factor index set.

[0083] The flight requirement index set includes minimum personnel configuration, supporting equipment, and operating conditions;

[0084] The set of parameters for the UAV includes deployment time, cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, size, and vertical takeoff height.

[0085] The environmental requirements index set includes maximum takeoff altitude, minimum ambient temperature, maximum ambient temperature, and wind resistance level;

[0086] The other factor indicators include market price, control capability, insurance payout indicators, after-sales service level, and application scenarios.

[0087] S2: Quantify the parameter configuration of the UAV to form a set of qualitative indicators and a set of quantitative indicators;

[0088] like Figure 2 As shown, S21: Form a qualitative indicator set from the indicators in the indicator set that cannot be measured by specific values, and convert the qualitative indicators in the qualitative indicator set into weighted indicators.

[0089] The set of qualitative indicators includes supporting equipment, operating conditions, control capabilities, after-sales service level, and application scenarios;

[0090] The supporting equipment will use whether RTK needs to be set up and the operating handle as weighted indicators:

[0091] No RTK or controller needs to be set up; the weighting index is set to 1.0.

[0092] No RTK setup is required, but a controller is needed; the weighting metric is set to 0.8.

[0093] RTK needs to be set up but no controller is needed; set the weighting index to 0.6.

[0094] RTK needs to be set up and a controller is required. The weighting index is set to 0.4.

[0095] The operational conditions are set with airspace approval and professional pilots as weighted indicators:

[0096] No airspace approval or professional pilots are required; the weighted indicator is set to 0.

[0097] No airspace approval is required, but professional pilots are needed; the weighting index is set at 0.5.

[0098] Airspace approval and professional pilots are required; the weighted index is set to 1.

[0099] The operational capability is set with assembly difficulty and the degree of human intervention during flight as weighted indicators:

[0100] It is easy to assemble and requires very little human intervention during flight; the weighted index is set at 0.9.

[0101] The assembly is complex and requires human intervention throughout the flight process, so the weighted index is set at 0.7;

[0102] Assembly is generally possible, but human intervention may be required occasionally during flight. Therefore, the weighted index is set at 0.8.

[0103] The after-sales service level will be set with timeliness and professional reliability as weighted indicators:

[0104] After-sales service is timely, professional, and reliable; the weighted indicator is set at 0.9.

[0105] After-sales service may not be timely, but professional and reliable; a weighted indicator of 0.7 is set for this service.

[0106] After-sales service is timely but not professional or reliable; a weighted indicator of 0.5 is set for this.

[0107] The after-sales service is untimely, unprofessional, and unreliable; a weighted indicator of 0.3 is set for this.

[0108] The application scenario uses the type of data acquired as a weighting metric:

[0109] It can acquire image data and laser point cloud data, and the weighting index is set to 1.0;

[0110] Acquire image data or laser point cloud data, and set the weighting index to 0.5.

[0111] Existing technologies for handling qualitative indicators in UAV selection typically rely on subjective scoring or simple grading based on expert experience. This approach suffers from core problems such as vague evaluation criteria, high subjectivity and arbitrariness, and poor consistency in scoring results across different experts or scenarios. This makes it difficult to fairly and repeatedly integrate qualitative indicators with precise quantitative indicators within the same mathematical model, thus weakening the objectivity and reliability of the overall selection decision. This application establishes a standardized set of rules for converting qualitative to quantitative data by clearly defining clear, specific, and observable judgment scenarios and their corresponding weighted values ​​for each qualitative indicator. This transforms subjective and vague descriptive information into objective, discrete numerical data. It fundamentally solves the problem of quantifying qualitative indicators in decision-making models. The built-in standardized conversion rules significantly reduce subjective bias in human evaluation, ensure the consistency of data sources, and allow all candidate solutions to be compared under a unified and transparent benchmark, significantly enhancing the scientific rigor, repeatability, and practical guidance value of the final selection results.

[0112] S22: Integrate the quantitative indicators in the indicator set into a quantitative indicator set, and normalize the quantitative indicator set;

[0113] S221: Divide the quantitative indicator set into a positive indicator set and a negative indicator set. Let the positive vector indicator corresponding to the maximum value of the positive indicator be 1. Normalize the positive indicator set into a positive vector indicator using the following formula:

[0114] ;

[0115] in, For the i-th positive vectorization index, For the i-th positive indicator, This represents the maximum value of the positive indicator.

[0116] The positive indicator set includes cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, vertical takeoff height, maximum takeoff altitude, maximum ambient temperature, wind resistance rating, and insurance payout indicators.

[0117] In this embodiment, there are three drones. The first drone has a flight time of 4 hours, the remaining two drones have flight times of 3 hours and 2 hours respectively, and 4 hours is the maximum flight time. Therefore, the positive vectorization index of the flight time of the first drone is 1, and the positive vectorization index of the flight time of the second drone is... The positive vectorization index of the third drone's endurance is .

[0118] S222: Let the negative vectorization index corresponding to the minimum value of the negative index be 1. Normalize the set of negative indices into negative vectorization indices using the following formula:

[0119] ;

[0120] in, For the i-th negative vectorization index, For the i-th negative indicator, This is the minimum value of the negative indicator;

[0121] The set of negative indicators includes minimum staffing, deployment time, minimum ambient temperature, market price, and size.

[0122] In existing technologies, processing negative indicators typically requires a positive transformation (such as taking the reciprocal or using the difference method). This extra step not only increases the complexity and time consumption of data processing, but more importantly, it may distort the proportional relationships of the original data during the mathematical transformation, leading to a loss of data accuracy and consequently affecting the objectivity of subsequent weight calculations and decision ranking. This application introduces a formula to directly normalize the set of negative indicators, transforming it into a value range within... The negative vectorization index, which is defined as an interval and whose smaller value is better, unifies the index processing framework at the algorithm logic level. It eliminates the independent positiveization stage that may introduce errors, simplifies and cohedes the data processing flow, significantly improves computational efficiency, and more importantly, strictly maintains the relative proportion between the original data, avoiding the accuracy loss caused by intermediate transformations. This allows subsequent entropy weight calculation and multi-attribute decision analysis to be based on a more direct and accurate data foundation.

[0123] In this embodiment, there are three drones. The market price of the first drone is 200,000 yuan, and the market prices of the remaining two drones are 300,000 yuan and 500,000 yuan respectively. 200,000 yuan is the minimum market price. Therefore, the negative quantitative index of the market price of the first drone is 1, and the negative quantitative index of the market price of the second drone is... (To two decimal places), the negative vectorized index of the price of the third drone is: .

[0124] S3: Construct a standardized positive matrix from the quantitative indicator set and the data within it, using the following formula:

[0125] ;

[0126] in, To standardize the positive matrix, The evaluation index for the j-th column of the i-th row of the scheme;

[0127] The evaluation indicators include negative vectorized indicators, positive vectorized indicators, and weighted indicators.

[0128] In this embodiment, to explain the standardized positive matrix, we assume that only the maximum takeoff weight, minimum personnel configuration, and control capability are considered. Assume there are three drones to choose from: Drone A has a maximum takeoff weight of 25kg, a minimum personnel configuration of 2, and a control capability of 0.8; Drone B has a maximum takeoff weight of 15kg, a minimum personnel configuration of 1, and a control capability of 0.8; and Drone C has a maximum takeoff weight of 30kg, a minimum personnel configuration of 3, and a control capability of 0.7. Then the standardized positive matrix is:

[0129] ;

[0130] To make it clearer, it is explained in the form of a standardized positive matrix table, as shown in Table 1.

[0131] Table 1: Example table of standardized positive matrices;

[0132] ;

[0133] S4: Determine the entropy weights of the standardized positive matrix using the entropy method to obtain the weight vector.

[0134] S41: The entropy value of each evaluation index in the standardized positive matrix is ​​determined using the entropy method. The formula is as follows:

[0135] ;

[0136] in, Let j be the entropy value of the evaluation index in column j. Let m be the probability matrix, and m be the total number of rows in the scheme;

[0137] S42: Calculate the difference coefficient and entropy weight of each evaluation indicator in the standardized positive matrix using entropy values, and integrate the entropy weights to obtain the weight vector. The formula is:

[0138] ;

[0139] in, Let be the difference coefficient of the evaluation index in column j, and n be the total number of columns of evaluation indicators. Let the entropy value weight of the evaluation index in column j be denoted as . It is the weight vector, which is the set of entropy weights of all evaluation indicators.

[0140] S5: Obtain the weighted normalized matrix through the weight vector and the standardized positive matrix. Use VIKOR to obtain the group utility value and individual regret value of each solution from the weighted normalized matrix. Combine the group utility value and individual regret value to obtain the compromise index.

[0141] like Figure 3 As shown in S51: Multiply each column of the evaluation index in the standardized positive matrix by the entropy value corresponding to the weight vector to obtain the weighted normalized matrix. The formula is:

[0142] ;

[0143] in, The evaluation index is the j-th column of the scheme in the i-th row of the weighted normalization matrix;

[0144] S52: Find the maximum and minimum values ​​for each column of evaluation indicators after normalization, forming the ideal solution and the negative ideal solution, using the following formula:

[0145] ;

[0146] in, For the ideal solution, It is a negative ideal solution;

[0147] S53: Calculate the group utility value of the UAV for each evaluation index and the ideal solution in each scheme, using the following formula:

[0148] ;

[0149] in, Let be the group utility value (i.e., the weighted distance sum) of the i-th scheme.

[0150] It is important to note that the smaller the group utility value, the closer it is to the ideal state.

[0151] S54: Calculate the individual regret value of each evaluation index for the UAV in the worst-case scenario for each scheme, using the following formula:

[0152] ;

[0153] in, Let be the individual regret value for the i-th solution;

[0154] It is important to note that the lower the individual regret value, the more the requirements are met.

[0155] S55: Standardize the group utility value and individual regret value for each option, and then calculate the compromise index by combining the standardized values. The formula is as follows:

[0156] ;

[0157] in, Let be the compromise index of the i-th option. For decision coefficients, To achieve the optimal group utility value, For all schemes The minimum value in, The worst-case group utility value. For all schemes The maximum value in, The optimal individual regret value. For all schemes The minimum value in, The worst individual's regret value. For all schemes The maximum value in.

[0158] In the example of strength, .

[0159] S6: Sort the compromise index of each solution from smallest to largest. When the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, select the solution corresponding to the first compromise index. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, select the compromise solution as the final solution. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, select the candidate solution as the final solution.

[0160] The criteria for judging an acceptable advantage are:

[0161] ;

[0162] in, The minimum value of all compromise indices. The second smallest value of all compromise indices, where m is the total number of rows in the scheme (i.e., the number of schemes).

[0163] The criteria for acceptable stability are: or ;

[0164] in, This represents the group utility value of the solution corresponding to the first compromise index. This represents the individual regret value for the solution corresponding to the first compromise index.

[0165] If the solution corresponding to the first compromise index does not satisfy the acceptable advantage, then the set of compromise solutions integrated by subsequent solutions corresponding to the first compromise index is as follows:

[0166] ;

[0167] in, For the k-th compromise solution, Let be the compromise index corresponding to the k-th compromise solution;

[0168] In this example, it is assumed that there are 4 options, and the compromise index of each option is sorted from smallest to largest. , , , , It does not meet the acceptable advantage requirement. , , , That is, the compromise solution set is .

[0169] It is important to note that the compromise solution indicates that no solution has a significant advantage, the top solutions have similar overall performance, and the choice depends on other secondary criteria or preferences, requiring manual selection from the set of compromise solutions.

[0170] If the solution corresponding to the first compromise index satisfies acceptable advantage but not acceptable stability, then initialize the candidate solution set. ,set up The process begins with a loop to determine the final set of candidate solutions for integration, based on the following criteria:

[0171] like and ,but Add to the candidate solution set. ;

[0172] in, The solution corresponding to the first compromise index. Let p be the compromise index corresponding to the p-th candidate solution. Let p be the p-th possible candidate solution. This is a set of candidate solutions.

[0173] In this example, it is assumed that there are 4 options, and the compromise index of each option is sorted from smallest to largest. , , , , If acceptable advantage is satisfied, then acceptable stability is not satisfied. , , The loop condition is not met, i.e., the candidate solution set is not satisfied. .

[0174] like Figure 4 As shown, an improved UAV selection system combining entropy and VIKOR is proposed, comprising a UAV parameter configuration module, a UAV parameter quantification module, a UAV parameter unification module, a weight vector calculation module, a compromise index calculation module, and a UAV scheme selection module.

[0175] The drone parameter configuration module selects the drone takeoff type according to business needs, selects multiple schemes according to the drone takeoff type, and determines the parameter configuration of each drone through the scheme;

[0176] The UAV parameter quantification module quantifies the UAV parameter configuration, forming a qualitative index set and a quantitative index set; the UAV parameter unification module constructs a standardized positive matrix from the data in the quantitative index set and the quantitative index set.

[0177] The weight vector calculation module determines the entropy weights of the standardized positive matrix using the entropy method, thus obtaining the weight vector.

[0178] The compromise index calculation module obtains a weighted normalized matrix through a weight vector and a standardized positive matrix. The weighted normalized matrix uses VIKOR to obtain the group utility value and individual regret value of each solution. The combination of the group utility value and individual regret value obtains the compromise index.

[0179] The UAV solution selection module sorts the compromise index of each solution from smallest to largest. When the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, the solution corresponding to the first compromise index is selected. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, a compromise solution is selected as the final solution. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, a candidate solution is selected as the final solution.

[0180] It should be noted that the solution corresponding to the first compromise index has obvious advantages, but it has a certain weakness. The candidate solution set includes the first one with obvious advantages, as well as subsequent solutions that are only slightly different from the previous solution and may pose a threat. The selection needs to be made manually from the candidate solution set.

[0181] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An improved method for selecting unmanned aerial vehicles (UAVs) that combines entropy and VIKOR, characterized in that, Includes the following sub-steps: Select the drone takeoff type based on business needs, select multiple options based on the drone takeoff type, and determine the parameter configuration of each drone based on the options; The parameters of the UAV are quantified to form a set of qualitative indicators and a set of quantitative indicators. Construct a standardized positive matrix from the set of quantitative indicators and the data within the set of quantitative indicators; The entropy weights are determined for the standardized positive matrix using the entropy method, resulting in a weight vector. The weighted normalized matrix is ​​obtained by using the weight vector and the standardized positive matrix. The weighted normalized matrix is ​​then used to obtain the group utility value and individual regret value for each option using VIKOR. The combination of the group utility value and individual regret value yields the compromise index. Sort the compromise indices of each solution from smallest to largest. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, then select the solution corresponding to the first compromise index. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, then select the compromise solution as the final solution. If the solution corresponding to the first compromise index satisfies acceptable advantage but cannot satisfy acceptable stability, then select the candidate solution as the final solution.

2. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, The process of determining the parameter configuration for each drone includes the following sub-steps: The drone takeoff type is selected according to different business needs. The drone takeoff type includes vertical takeoff drones, fixed-wing drones, and compound-wing drones. Multiple drones are determined as a plan based on the selected drone takeoff type. The scheme determines the parameter configuration for each UAV, which includes a set of flight requirement indicators, a set of UAV parameter indicators, a set of environmental requirement indicators, and a set of other factor indicators.

3. The improved UAV selection method combining entropy and VIKOR as described in claim 2, characterized in that, The flight requirement index set includes minimum personnel configuration, supporting equipment, and operating conditions; The set of parameters for the UAV includes deployment time, cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, size, and vertical takeoff height. The environmental requirements index set includes maximum takeoff altitude, minimum ambient temperature, maximum ambient temperature, and wind resistance level; The other factor indicators include market price, control capability, insurance payout indicators, after-sales service level, and application scenarios.

4. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, The formation of the qualitative and quantitative indicator sets includes the following sub-steps: The indicators that cannot be measured by specific numerical values ​​in the indicator set are formed into a qualitative indicator set, and the qualitative indicators in the qualitative indicator set are converted into weighted indicators. The quantitative indicators in the indicator set are integrated to form a quantitative indicator set, and the quantitative indicator set is normalized. The set of qualitative indicators includes supporting equipment, operating conditions, control capabilities, after-sales service level, and application scenarios; The quantitative indicator set includes a positive indicator set and a negative indicator set; The positive indicator set includes cruise speed, endurance, maximum takeoff weight, service ceiling, controlled distance, vertical takeoff height, maximum takeoff altitude, maximum ambient temperature, wind resistance rating, and insurance payout indicators. The set of negative indicators includes minimum staffing, deployment time, minimum ambient temperature, market price, and size.

5. The improved UAV selection method combining entropy and VIKOR as described in claim 4, characterized in that, The normalization of the quantitative index set includes the following steps: The quantitative indicator set is divided into a positive indicator set and a negative indicator set. Let the positive vector indicator corresponding to the maximum value of the positive indicator be 1. The positive indicator set is then normalized into a positive vector indicator using the following formula: ； in, For the i-th positive vectorization index, For the i-th positive indicator, This represents the maximum value of the positive indicator. Let the negative vectorization index corresponding to the minimum value of the negative index be 1. The formula for normalizing the set of negative indices into negative vectorization indices is: ； in, For the i-th negative vectorization index, For the i-th negative indicator, This represents the minimum value of the negative indicator.

6. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, The formula for constructing the standardized positive matrix is ​​as follows: ； in, To standardize the positive matrix, The evaluation index for the j-th column of the i-th row of the scheme; The evaluation indicators include negative vectorized indicators, positive vectorized indicators, and weighted indicators.

7. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, The process of obtaining the group utility value and individual regret value for each option using VIKOR, and then combining the group utility value and individual regret value to obtain the compromise index, includes the following sub-steps: Multiply each column of the evaluation index in the standardized positive matrix by the entropy value weight corresponding to the weight vector to obtain the weighted normalized matrix; Find the maximum and minimum values ​​of each column of evaluation indicators after normalization to form the ideal solution and the negative ideal solution; Calculate the group utility value of each evaluation index of the UAV in each scheme compared to the ideal solution; Calculate the individual regret value of each evaluation index for the drone in the worst-case scenario for each scheme; The group utility value and individual regret value of each option are standardized separately, and the compromise index is calculated by combining the standardized values.

8. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, The criteria for determining the acceptable advantage are: ； in, The minimum value of all compromise indices. The second smallest value of all compromise indices, where m is the total number of rows in the scheme; The criteria for determining acceptable stability are as follows: or ; in, This represents the group utility value of the solution corresponding to the first compromise index. This represents the individual regret value for the solution corresponding to the first compromise index. To achieve the optimal group utility value, This represents the optimal individual regret value.

9. The improved UAV selection method combining entropy and VIKOR as described in claim 1, characterized in that, If the solution corresponding to the first compromise index does not satisfy the acceptable advantage, then the set of compromise solutions integrated by subsequent solutions corresponding to the first compromise index is as follows: ； in, For the k-th compromise solution, Let be the compromise index corresponding to the k-th compromise solution, and m be the total number of rows in the solution; If the solution corresponding to the first compromise index satisfies acceptable advantage but not acceptable stability, then the candidate solution set is initialized. ,set up The process begins with a loop to determine the final set of candidate solutions for integration, based on the following criteria: like and ,but Add to the candidate solution set. ; in, The solution corresponding to the first compromise index. Let p be the compromise index corresponding to the p-th candidate solution. Let p be the p-th possible candidate solution. This is a set of candidate solutions.

10. An improved UAV selection system combining entropy and VIKOR, characterized in that, It includes a drone parameter configuration module, a drone parameter quantification module, a drone parameter unification module, a weight vector calculation module, a compromise index calculation module, and a drone solution selection module. The drone parameter configuration module selects the drone takeoff type according to business needs, selects multiple schemes according to the drone takeoff type, and determines the parameter configuration of each drone through the scheme; The UAV parameter quantification module quantifies the UAV parameter configuration, forming a qualitative index set and a quantitative index set; the UAV parameter unification module constructs a standardized positive matrix from the data in the quantitative index set and the quantitative index set. The weight vector calculation module determines the entropy weights of the standardized positive matrix using the entropy method, thus obtaining the weight vector. The compromise index calculation module obtains a weighted normalized matrix through a weight vector and a standardized positive matrix. The weighted normalized matrix uses VIKOR to obtain the group utility value and individual regret value of each solution. The combination of the group utility value and individual regret value yields the compromise index. The UAV solution selection module sorts the compromise index of each solution from smallest to largest. When the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, the solution corresponding to the first compromise index is selected. If the solution corresponding to the first compromise index cannot satisfy acceptable advantage, a compromise solution is selected as the final solution. If the solution corresponding to the first compromise index satisfies both acceptable advantage and acceptable stability, a candidate solution is selected as the final solution.

Images