Introducing the comparative weighting evaluation based on the best criteria method: a multi-fuzzy framework for aviation fuel efficiency Authors: Serkan Eti, Fatih Karaman, Serhat Yüksel, Hasan Dinçer, İpek Yaylalı Published: 2025 ABSTRACT AbstractFuel consumption in the aviation industry is one of the most critical cost items in terms of both economic and environmental sustainability. Although numerous studies have examined the determinants of aircraft fuel efficiency, there remains no consensus in the literature regarding which factors are most influential. This lack of clarity creates a major research gap that limits the ability of airlines to allocate resources effectively, make strategic plans, and implement operational improvements. To address this gap, the present study introduces the Comparative Weighting Evaluation based on the Best Criteria (COWEB) method — a novel multi-fuzzy decision-making framework designed to identify and prioritize the most critical factors influencing aircraft fuel consumption. The distinct originality of the COWEB model lies in its comparative structure, which evaluates all criteria relative to the most significant one, thereby reducing cognitive load and enhancing consistency. Moreover, the integration of Triangular, Pythagorean, and Fermatean fuzzy environments enables comprehensive modeling of expert uncertainty across multiple dimensions. According to the results, “engine efficiency and technology” and “pilot behavior and flight management” are the two most influential criteria, with respective weights of 0.254 and 0.227. While the COWEB technique yields stable results under all fuzzy environments, the comparative SIWEC analysis reveals minor inconsistencies, confirming the superior stability and reliability of the proposed model. By minimizing the number of required comparisons, the COWEB method improves expert consistency, reduces evaluation time, and provides a clearer differentiation among closely related factors.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-32988-7.Keywords: Aircraft fuel consumption, COWEB technique, Fuzzy logic, Engine efficiency, Pilot behaviour, Sustainability, Decision-making modelsSubject terms: Engineering, Mathematics and computing Introduction Fuel consumption in aircraft is one of the most important issues in aviation. The amount of fuel consumed is a significant cost component in aircraft. Consequently, this has a significant negative impact on the financial performance of companies1. If the necessary measures are not taken, this situation leads to a number of significant financial problems2. As a result, airlines lose their economic competitiveness. Furthermore, excessive fuel consumption leads to the release of more carbon dioxide gas into the atmosphere. This is considered one of the most significant causes of global warming. Therefore, reducing aircraft fuel consumption is crucial to reduce environmental pollution. Many different factors affect aircraft fuel consumption. The air resistance an aircraft experiences while moving through the air is related to aerodynamic efficiency3. Less resistance means the engine consumes less power, which directly leads to fuel savings. Furthermore, the use of new aircraft types can also have an impact on fuel consumption. New aircraft models are generally lighter, quieter, and more efficient. This makes a significant difference in both environmental and economic terms. In addition, the heavier the aircraft, the more thrust it requires to generate lift. This causes the engine to use more fuel. Technological advancements are another important factor in this process4. Newer-generation engines consume significantly less fuel than older models. Finally, fuel consumption in aircraft is also affected by flight altitude. The higher the aircraft flies, the lower the air density. This allows the engine to maintain the same speed while consuming less power. Numerous factors influence aircraft fuel costs. However, it is not appropriate to assign equal importance to all of them. For businesses to effectively utilize their limited resources, it’s essential to identify the most critical factors. This is crucial for strategic planning and operational efficiency. Identifying the most important criteria allows businesses to develop concrete strategies that can be implemented in the short term. Numerous studies exist in the literature examining the factors affecting aircraft fuel consumption5. However, most of these studies address factors at a general level and fail to reveal their relative importance6. The limited number of studies identifying the most critical factors is considered a significant “missing gap” in the aviation field7. This deficiency creates uncertainty both theoretically and practically8. Airline companies lack clear guidance on which factors to prioritize. This leads to consequences such as increased costs, resource waste, and a departure from sustainability goals. Furthermore, the lack of systematic identification of fuel efficiency-focused strategic priorities in the academic literature creates a significant theoretical gap9. Therefore, the need for a new and comprehensive priority analysis is increasingly necessary. Such an analysis will both support the strategic decision-making processes of airline companies and make a significant scientific contribution by filling the existing gap in the literature. This study aims to define essential indicators of aircraft fuel consumption. The study is motivated by the need for airlines to reduce fuel costs, achieve sustainability goals, and optimize strategic resource utilization. As a result of the literature review, eight key criteria affecting fuel consumption are identified. A unique decision-making model is developed to identify the most critical of these criteria. The proposed model, called comparative weighting evaluation based on the best criteria (COWEB), is based on a multi-criteria decision-making approach. The unique feature of the model is its comparative evaluation principle and its integration with various fuzzy sets. To achieve this, the COWEB technique is applied in conjunction with triangular, Pythagorean, and Fermatean fuzzy set structures. This integrated structure enables more realistic modeling of uncertainties and more effective evaluation of expert opinions in the decision-making process. In addition, to test the validity and reliability of the proposed method, a comparative analysis is conducted with the SIWEC approach. The study not only develops a theoretical model but also demonstrates the practical applicability of the proposed approach. In this direction, the research seeks to answer the following fundamental questions: (1) Which are the most critical factors affecting aircraft fuel consumption? (2) To what extent is the COWEB model an effective method in determining the importance levels of these factors? (3) Does the integration of Triangular, Pythagorean, and Fermatean fuzzy sets increase the accuracy and consistency of the decision-making process? (4) Does the proposed COWEB approach provide more reliable and flexible results than the existing SIWEC method? This study contributes to the aviation sustainability and decision-making literature by developing a new comparative weighting model designed to identify and prioritize the key factors influencing aircraft fuel consumption. The research introduces the COWEB method, which represents a methodological advancement by integrating multiple fuzzy environments (Triangular, Pythagorean, and Fermatean) to capture expert uncertainty more comprehensively. The novelty of the developed model is evident in both the structure of the weighting technique and the integration of the fuzzy sets used. In this respect, the study distinguishes itself from existing methods in the literature, both methodologically and practically. (1) This study develops a new weighting technique called COWEB. Various criticisms of existing weighting methods were considered in the development of this new technique10. In pairwise comparison-based methods such as AHP and DEMATEL, the large number of questions reduces the consistency of experts and compromises the reliability of the data11. In objective weighting methods such as LOPCOW and Entropy, alternatives must be determined in advance to calculate criterion weights. This makes it difficult to use these techniques when analysis is aimed at criteria only12. In methods like CIMAS, experts cannot fully reflect the relative importance of important criteria because they lack a reference value. Considering these shortcomings, the COWEB technique was developed. In the COWEB approach, the most important criterion is first determined, and other factors are compared based on this criterion. Thus, the model provides a more concrete and reliable basis for comparison. In this respect, COWEB offers several key advantages over existing methods: it has higher reliability, enables a clearer separation of criteria with similar importance levels, and significantly reduces data collection time. While AHP collects evaluations for each criterion pair, COWEB only constructs pairs with the most important criterion13. This requires collecting more evaluations in expert systems. In other words, this means the evaluation form is longer14. In BWM, however, the best and worst criteria must be determined. This requires evaluators to rank each criterion twice. This complicates the creation of the form and the collection of evaluations. (2) To enable a more comprehensive and flexible analysis, the COWEB technique is implemented in conjunction with Triangular, Pythagorean, and Fermatean fuzzy set structures15. This integration allows the model to manage uncertainties more effectively. The Triangular fuzzy set al.lows for a simple and intuitive expression of uncertainties in expert opinions. The Pythagorean fuzzy set al.lows for a more flexible definition of membership and dissenting membership degrees, thus improving the accuracy of the evaluation process. The Fermatean fuzzy set is a more generalized version of this structure, allowing experts to reflect their uncertainty perceptions within a broader framework16. Using these three structures together contributes to holistic modeling of the decision-making process. Furthermore, a comparative analysis is conducted with the SIWEC approach, widely used in the literature, to assess the validity and reliability of the proposed method. This comparison demonstrates that the COWEB technique offers higher reliability and success not only theoretically but also in practice. Thus, the proposed model represents a superior alternative to existing methods in the literature in terms of both accuracy and applicability. The organization of the manuscript is explained below. Section “Literature review” underlines the missing gap in the literature. Section “Methodology” indicates the details of the proposed model. Section “Comparative weighting evaluation based the best criteria (COWEB)” denotes the main results of this model. Section “COWEB with triangular fuzzy sets” makes a comprehensive discussion based on the analysis results. Section “COWEB with pythagorean fuzzy sets” focuses on the main conclusions and limitations. Literature review Aerodynamic efficiency affects aircraft fuel consumption. In the literature, Guo et al.17 addressed the relationship between aerodynamic drag and fuel efficiency, emphasizing the importance of aerodynamic shape modification in optimizing aircraft performance. The study demonstrates that reducing aerodynamic drag can improve fuel efficiency, and therefore, aerodynamic shape modification is an important parameter in aircraft design. Luo et al.18 addressed the multidisciplinary optimization of aircraft aerodynamics using distributed propulsion systems (DP). The study emphasizes the importance of coordinating different functions, such as the aerodynamic structure, propulsion, energy, and control systems. Jang et al.19 compared the fuel efficiency of the Airbus A380 and A350, finding that the A350 is approximately twice as efficient with the same load, and that the economic sustainability of the A380 is directly related to the balance between revenue and cost. Kretov and Tiniakov20 highlight the potential of high-span wings to increase aerodynamic efficiency, particularly in passenger aircraft designed with composite materials and featuring foldable flight wings. The study demonstrates that this design can improve fuel efficiency by reducing drag, and therefore, increasing aerodynamic efficiency has a significant impact on reducing fuel consumption. Another factor affecting fuel consumption is aircraft weight. Zhang et al.21 developed a comprehensive mathematical framework that provides high-precision and high-resolution fuel calculation methods by establishing a link between flight dynamics and fuel consumption. The related study also addresses the effect of aircraft weight on fuel consumption. The study indicates that increasing aircraft weight directly increases fuel consumption, especially during energy-intensive phases such as takeoff, climb, and landing. Zhao et al.22 analyzed the impact of aircraft weight on fuel consumption using a data-driven fuel consumption prediction model. The study shows that carrying additional fuel directly increases fuel consumption by increasing total weight. This demonstrates that optimizing aircraft weight is a critical strategy for fuel efficiency, especially for long-haul flights. Engine efficiency and technology also affect aircraft fuel consumption. Chen et al.23 state that adaptive cycle engines are one of the most innovative technologies with the potential to increase fuel efficiency in aviation. They state that these engines can achieve a dynamic balance between both high performance and low fuel consumption. Furthermore, it is stated that ACE architectures can reduce fuel consumption by 10–20% compared to traditional turbofan engines. Sadeq et al.24 comprehensively examined next-generation combustion technologies developed to increase the fuel efficiency of modern aircraft engines. The study emphasizes that homogeneous charge compression ignition (HCCI) and reactivity-controlled compression ignition (RCCI) systems can significantly reduce fuel consumption by increasing engine thermal efficiency. Liu and Yang25, while studying the effects of Fischer–Tropsch fuel mixtures on engine performance, conclude that the compatibility between fuel composition and engine-discharge/design control laws, especially those directly related to engine technology, is a critical element affecting fuel consumption. The adoption of new aircraft types also plays a significant role in this process. Babuder et al.26, examining the impact of sustainable aviation technologies on the current operational ecosystem, indicated that new aircraft types provide significant improvements in fuel consumption. Oliveira et al.27, focusing on fleet modernization in the Brazilian airline industry over a two-decade period, concluded that the adoption of new-generation aircraft reduces operating costs, improves service quality, and reduces emissions. These findings highlight the potential for increasing fuel efficiency of new aircraft types and the strategic importance of fleet modernization. Pilot behavior and flight management also impact fuel consumption. Wunderli et al.28 examined the effectiveness of pilot assistance systems in reducing fuel consumption during energy-efficient landings. The study shows that pilots’ flight management decisions, combined with the energy optimization strategies suggested by the system, lead to fuel savings. Furthermore, it is explained that pilots’ effective use of these support systems helps prevent unnecessary fuel consumption during the descent and approach phases. These findings reveal that pilot behaviors and decision-making mechanisms during flight have a direct impact on fuel efficiency. Xiong et al.29 developed a model to estimate fuel consumption during the climb, cruise, and landing phases of flight. Using data in this model, they aimed to more accurately analyze the effects of pilots’ flight management decisions on fuel consumption. The findings reveal that the management strategies pilots implement during flight have a significant impact on fuel efficiency. Yoo and Jeon30 investigated the behaviors pilots exhibit during maneuvers and the effects of these behaviors on fuel consumption. The study examined pilots’ attitudes toward environmental protection and fuel conservation. The analyses revealed that the management strategies pilots employ during flight have a significant impact on fuel efficiency. These findings demonstrate that pilot behavior and decision-making during flight have a direct impact on fuel efficiency. Cruise altitude and speed selection are also factors to consider in this process. Ekici et al.31 comprehensively examined the effects of different cruising altitudes on fuel consumption and emissions in different commercial aircraft types. The results show that appropriate cruise altitude selection is critical for minimizing fuel consumption and reducing emissions. These findings emphasize that optimizing flight level in airline operations is an important tool for energy efficiency and environmental sustainability. Madjid et al.32 examined the effects of different cruising speeds on fuel consumption in a Cessna 172 aircraft. The results show that actual fuel consumption exceeds Pilot Operating Handbook (POH) estimates. The findings highlight the importance of selecting optimal cruise speeds to increase fuel efficiency and reduce operational costs. Route planning, taxiing, and ground maneuvers also play a role in this process. Chen et al.33 examined the effects of aircraft taxiway optimization on fuel consumption. The study revealed that taxi duration, number of turns, and flight timing directly impact fuel consumption. It also emphasized that conflicts and obstacles encountered during taxiing increase fuel consumption, while optimizing the number of engines and taxiways plays a critical role in improving fuel efficiency. Edemeka34 aimed to develop a fuel consumption and emissions inventory for local aircraft fleets using AIMS (Airline Information Management System) data. The study demonstrated that AIMS data provides a critical resource for fuel consumption calculations by reflecting real-time flight activity. Furthermore, accurate fusion of flight data increases the reliability of fuel efficiency analyses by minimizing the effects of out-of-sequence flights and delays. Curtis et al.35 examined the effects of planning energy-efficient flight paths on busy routes on fuel consumption. Findings indicate that optimizing flight routes and energy-efficient planning is an effective strategy to reduce fuel consumption in airline operations. Environmental and meteorological conditions can also affect aircraft fuel consumption. Némethová and Vagner36, in their comprehensive study of wind effects on aircraft fuel consumption and costs, revealed that headwind conditions increase fuel consumption, while tailwind conditions contribute to fuel savings. Furthermore, wind direction and intensity are highlighted as critical parameters that directly affect the relationship between flight duration and fuel consumption. These findings highlight the importance of considering meteorological conditions in flight planning for fuel efficiency and operational costs. Nguyen et al.9 used deep reinforcement learning methods to examine takeoff routes that minimize both fuel consumption and noise impacts. The results show that energy-efficient takeoff strategies that consider wind conditions are effective in reducing fuel consumption and increasing operational efficiency. Galant et al.37 analyzed the effects of wind on fuel consumption and harmful exhaust gas emissions on specific flight routes. The study revealed that headwind conditions increase fuel consumption, while tailwind conditions increase fuel efficiency. Furthermore, wind speed and direction were evaluated as direct determinants of flight performance and emissions. These results demonstrate that considering meteorological parameters in flight planning is critical for both fuel savings and environmental sustainability. Several previous studies have applied multi-criteria decision-making techniques to aviation fuel management and efficiency assessment. For instance, AHP and TOPSIS have been widely used to evaluate fuel-saving strategies and aircraft selection alternatives by considering economic, environmental, and operational factors38,39. Some studies have also adopted DEMATEL and SWARA to determine the causal relationships among criteria affecting fuel consumption and to derive expert-based weights for technical parameters40,41. However, these methods typically require large numbers of pairwise comparisons or rely on pre-defined alternatives, which may reduce the consistency of expert evaluation42. Moreover, most of these approaches handle uncertainty in a limited way, usually through crisp or single-type fuzzy sets. The COWEB model developed in this study addresses these shortcomings by reducing the cognitive burden on experts, integrating multiple fuzzy environments to manage uncertainty comprehensively, and providing more consistent and reliable weighting results in the context of aviation fuel efficiency analysis. Literature review reveals that the factors affecting aircraft fuel consumption are multidimensional, interconnected, and interactive. Research demonstrates that improving aerodynamic efficiency and reducing aircraft weight directly reduce fuel consumption. Furthermore, advances in engine technologies have led to higher thermal efficiency and lower fuel consumption. New-generation aircraft models and fleet modernization are reported to combine aerodynamic and engine efficiency to save fuel while also reducing maintenance costs and environmental impacts. From an operational perspective, it is understood that factors such as pilot behavior, flight management decisions, cruise altitude, and speed selection directly affect fuel efficiency. Furthermore, optimizing route planning, taxi times, and ground maneuvers improve operational efficiency, while meteorological conditions—particularly wind direction and strength—have a decisive impact on fuel consumption and emissions43,44. Current literature evaluates fuel consumption not only as the amount of energy expended during flight, but also as an integrated result of aircraft design, engine technology, operational management, environmental conditions, and human factors. Therefore, it is believed that a holistic approach should be taken to improve fuel efficiency, encompassing technology, engineering, the environment, and human factors. It has been observed that the literature contains limited studies on the factors affecting aircraft fuel consumption, which focus on single concepts. Furthermore, it is understood that a clear operational sequence for reducing fuel consumption in the civil aviation sector is not yet established. It is known that most airlines utilize the appropriate options listed, based on their fleet and operational structures. The study aims to conduct research to increase fuel efficiency and rank the factors affecting fuel consumption. Methodology This section describes the details of the newly developed method, its advantages, calculations and process with fuzzy numbers. The main motivations for the development of this method were criticisms of other multi-criteria decision-making models. This developed method eliminates the need for alternative methods such as CRITIC, LOPCOW, or MEREC in calculating criterion weights. Thus, it offers a unique methodology for studies requiring only criterion weighting. Additionally, the fact that other criteria are evaluated based only on the most important criterion provides an advantage over methods like AHP and DEMATEL. Because methods like AHP and DEMATEL require data collection for each criterion pair, they prevent experts with time constraints or heavy workloads from dedicating time. In these cases, experts either avoid making assessments or conduct superficial assessments to speed up the process. This proposed method overcomes this problem by reducing the number of questions. In this study, the integration of Triangular, Pythagorean, and Fermatean fuzzy environments was intentionally designed to capture different dimensions of uncertainty in expert evaluations. The Triangular fuzzy set provides a simple and intuitive representation of uncertainty, making it particularly effective for modeling linguistic evaluations in cases where experts express approximate or interval-based judgments. The Pythagorean fuzzy set expands this capability by allowing a more flexible balance between membership and non-membership degrees, which helps capture the hesitation and partial confidence often observed in decision-making under ambiguity. The Fermatean fuzzy set, as a generalized structure of the previous two, enables experts to represent their uncertainty perceptions within a broader and more tolerant mathematical space. In the context of aviation fuel efficiency, where expert judgments involve technical, operational, and environmental uncertainties simultaneously, using all three fuzzy environments ensures that the model can accommodate varying levels of vagueness, hesitation, and confidence. This integration thus enhances the robustness and realism of the decision-making process. When comparing the proposed method with methods like SIWEC and CIMAS, the presence of a concrete concept underlying the evaluation process demonstrates the superiority of the method. Because methods like SIWEC and CIMAS evaluate criteria individually, they cannot fully reveal the distinction when the difference in importance between criteria is small45. Since two criteria are analyzed by expressing them as important, the difference in the results cannot be revealed. However, when the proposed methodology uses the same criterion, the distinction is more clearly demonstrated by evaluating the two criteria according to the underlying criterion. Finally, considering SWARA, another advantage of the method is that it does not require calculations for each expert46. Comparing it with other methods demonstrates the proposed method’s reliability and advantage in complex problems. Furthermore, the inclusion or exclusion of expert weights provides flexibility, allowing for parametric flexibility. Comparative weighting evaluation based the best criteria (COWEB) Each expert determines the most important criterion among the criteria, and the most important criterion is scored as “1.” Each criterion is then evaluated according to the scale in Table 1, based on the most important criterion. Table 1Scale.TermScaleEqual Importance (EI)1Relatively Low Importance (RLI)0.5Low Importance (LI)0.3Very Low Importance (VLI)0.25Extreme Low Importance (ELI)0.2 Scale. As a result of each expert’s evaluation, the evaluation matrix in Eq. (1) is constructed using scale equivalents of linguistic terms. Wherein e and n are the numbers of experts and criteria, respectively. In other words, if the jth criterion is the most important criterion for the ith expert, the value of xij is 1. Otherwise, xij is the scale value of the evaluation of the jth criterion with respect to the most important criterion for the ith expert. Next, the elements of evaluation matrix are normalized with the help of Eq. (2). Equation (2) normalizes the evaluations of each expert. In other words, the relative importance of each expert regarding the criteria is obtained. Then, normalized elements are summed using Eq. (3). Thanks to Eq. (3), the sum of the relative importance of each criterion among experts is obtained. If desired, expert weightings can be used during this aggregation process. This results in a sum of expert-weighted relative importance scores. In this case, the new form of Eq. (3) becomes Eq. (4). Wherein ew refers to expert weighting. Finally, weights of criteria are defined by Eq. (5). A high w value indicates that the importance of that criterion is high. To clarify the methodological distinction of the proposed COWEB model, Table 2 presents a comparative analysis with two recent and advanced MCDM models: the Z-Number Extension of the Parsimonious Best Worst Method (Z-PBWM) and the Parsimonious Spherical Fuzzy Analytic Hierarchy Process (PSF-AHP). Table 2Comparative analysis of COWEB and recent extended MCDM Models.FeatureCOWEB (Proposed)Z-Number Extension of Parsimonious Best Worst Method (Z-PBWM)Parsimonious Spherical Fuzzy Analytic Hierarchy Process (PSF-AHP)Main StructureComparative weighting based on the best criterionPairwise best–worst comparisons with Z-number uncertainty representationHierarchical pairwise comparisons using spherical fuzzy setsCore InnovationSimplifies comparison process by using one reference criterion; integrates multiple fuzzy environmentsIncorporates Z-numbers to capture reliability of expert judgmentsUtilizes spherical fuzzy numbers to improve expressiveness in uncertainty modelingCognitive Load on ExpertsVery low (requires minimal comparisons)Moderate (still requires structured pairwise comparisons)High (full pairwise matrix required)Type of Uncertainty RepresentationMulti-fuzzy (Triangular, Pythagorean, Fermatean)Z-numbers (fuzzy + reliability components)Spherical fuzzy numbersComputation ComplexityLow; simple normalization-based calculationModerate; additional reliability parameter computationsHigh; multiple membership functions processedOutput ConsistencyHigh; stable results across fuzzy environmentsModerate; depends on reliability weightsVariable; sensitive to fuzzification levelApplicability to Aviation ProblemsHigh; suitable for operational and strategic factor analysisMedium; complex interpretation for practical systemsMedium; requires more data and expert inputMain AdvantageReduces comparisons, improves consistency, and integrates multi-fuzzy logicEnhances reliability modeling in expert evaluationsCaptures more nuanced uncertainty through spherical membershipMain LimitationDoes not rank alternatives directlyIncreases computational demand with added reliability dimensionRequires extensive pairwise judgments Comparative analysis of COWEB and recent extended MCDM Models. As shown in Table 2, the COWEB model significantly reduces expert workload while maintaining high consistency across multiple fuzzy environments. Unlike the compared methods, which either increase computational complexity or require extensive pairwise evaluations, COWEB offers a balanced approach between methodological simplicity and uncertainty modeling depth. To strengthen the comparative framework, the evaluation characteristics of the three models were further examined. First, the Z-PBWM47 and PSF-AHP48 studies were added to the reference list to properly acknowledge their methodological origins. Compared with these methods, COWEB requires significantly fewer pairwise inputs, since only comparisons with respect to the most important criterion are needed. For example, when evaluating eight criteria, PSF-AHP requires 28 pairwise comparisons and Z-PBWM requires at least 14 structured best–worst comparisons, whereas COWEB requires only 7 comparisons. This difference substantially reduces the time of evaluation and lowers the cognitive burden on experts. Moreover, the absence of a full pairwise comparison matrix in COWEB eliminates inconsistencies that may arise when experts struggle with large comparison sets. These characteristics collectively justify why COWEB provides a more efficient and reliable weighting structure, especially in real-world decision-making environments where expert time is limited. COWEB with triangular fuzzy sets Each expert determines the most important criterion among the criteria, and the most important criterion is scored as “1.” Each criterion is then evaluated according to the scale in Table 349, based on the most important criterion. Table 3Scale for triangular fuzzy Numbers.TermabcEqual Importance (EI)0.7511Relatively Low Importance (RLI)0.50.751Low Importance (LI)0.250.50.75Very Low Importance (VLI)00.250.5Extreme Low Importance (ELI)000.25 Scale for triangular fuzzy Numbers. As a result of each expert’s evaluation, the evaluation matrix in Eq. (6) is constructed using scale equivalents of linguistic terms. Wherein . In other words, if the jth criterion is the most important criterion for the ith expert, the a, b and c values of are 1. Otherwise, is the scale value of the evaluation of the jth criterion with respect to the most important criterion for the ith expert. Next, the elements of evaluation matrix are normalized with the help of Eq. (7). If desired, expert weightings can be used during this aggregation process. This results in a sum of expert-weighted relative importance scores. In this case, the new form of Eq. (8) becomes Eq. (9). Wherein ew refers to expert weighting. Finally, weights of criteria are defined by Eqs. (10) and (11). A high w value indicates that the importance of that criterion is high. COWEB with pythagorean fuzzy sets Each expert determines the most important criterion among the criteria, and the most important criterion is scored as “(1,0)” Each criterion is then evaluated according to the scale in Table 450, based on the most important criterion. Table 4Scale for pythagorean fuzzy Numbers.TermMembership (µ)Non-Membership (v)Equal Importance (EI)0.960.27Relatively Low Importance (RLI)0.830.46Low Importance (LI)0.650.54Very Low Importance (VLI)0.460.73Extreme Low Importance (ELI)0.380.84 Scale for pythagorean fuzzy Numbers. As a result of each expert’s evaluation, the evaluation matrix in Eq. (12) is constructed using scale equivalents of linguistic terms. Wherein . In other words, if the jth criterion is the most important criterion for the ith expert, the and values of are 1 and 0, respectively. Otherwise, is the scale value of the evaluation of the jth criterion with respect to the most important criterion for the ith expert. Next, the elements of evaluation matrix are defuzzified with the help of Eq. (13). Then, the elements of defuzzified matrix are normalized with the help of Eq. (14). Then, normalized elements are summed using Eq. (15). Finally, weights of criteria are defined by Eq. (16). A high w value indicates that the importance of that criterion is high. COWEB with fermatean fuzzy sets Each expert determines the most important criterion among the criteria, and the most important criterion is scored as “(1,0)” Each criterion is then evaluated according to the scale in Table 551, based on the most important criterion. Table 5Scale for fermatean fuzzy Numbers.TermMembership (β)Non-Membership (α)Equal Importance (EI)0.90.1Relatively Low Importance (RLI)0.70.2Low Importance (LI)0.40.5Very Low Importance (VLI)0.10.75Extreme Low Importance (ELI)01 Scale for fermatean fuzzy Numbers. As a result of each expert’s evaluation, the evaluation matrix in Eq. (17) is constructed using scale equivalents of linguistic terms. Wherein . In other words, if the jth criterion is the most important criterion for the ith expert, the and values of are 1 and 0, respectively. Otherwise, is the scale value of the evaluation of the jth criterion with respect to the most important criterion for the ith expert. Next, the elements of evaluation matrix are defuzzified with the help of Eq. (18). Then, the elements of defuzzified matrix are normalized with the help of Eq. (19). Then, normalized elements are summed using Eq. (20). Finally, weights of criteria are defined by Eq. (21). A high w value indicates that the importance of that criterion is high. Numerical example with AHP, BWM, and COWEB A case study is conducted to compare the speed and efficiency of the performance of methods with different criteria. Problems are created first with three criteria, then with four criteria, and finally with seven criteria. To solve these problems, feedback is collected from decision-makers specific to the three methods, and analyses are conducted. The detailed timeline of the three methods for the same problems is summarized in Table 6. Table 6Time performance of Methods.Number of CriteriaMethodsEvaluation Collection TimeAnalysis TimeTotal Time3COWEB10 min15 min25 minBWM15 min20 min35 minAHP40 min10 min50 min4COWEB15 min17 min32 minBWM30 min22 min42 minAHP45 min12 min57 min7COWEB35 min20 min55 minBWM40 min25 min65 minAHP55 min15 min70 min Time performance of Methods. When Table 6 is examined, although the AHP method analysis seems fast, the total time is very long due to the long data collection process. Numerical sample datasets are shared in the attached Tables A5 - A7. The numerical analysis results obtained with these criteria are summarized in Table 7. Table 7Outputs of criteria weights of numerical Examples.Number of CriteriaCriteriaCOWEBBWMAHP3CR10.5270.7520.571CR20.3040.1810.286CR30.1680.0670.1434CR10.4130.5970.484CR20.2980.2320.319CR30.1670.1160.124CR40.1220.0550.0747CR10.2860.3850.374CR20.2340.2340.230CR30.1360.1170.170CR40.1030.0940.114CR50.0890.0780.054CR60.0770.0590.035CR70.0760.0330.024 Outputs of criteria weights of numerical Examples. As can be seen from Table 7, the results of the three methods give the same order of priority. However, as shown in Table 6, COWEB gives this priority order faster than other methods. Additionally, the CR values ​​of AHP are 0.000, 0.021, and 0.049, respectively. Moreover, three alternatives are ranked by SAW using these criterion weights. The results are presented in Table 8. Table 8Outputs of SAW for numerical Examples.Number of CriteriaAlternativesCOWEBBWMAHP3ALTRN10.9660.9800.968ALTRN20.6170.6400.621ALTRN30.6700.6280.6644ALTRN10.9620.9780.974ALTRN20.5970.6250.615ALTRN30.6390.6270.5997ALTRN10.8670.9040.918ALTRN20.5060.5440.555ALTRN30.7290.6960.703 Outputs of SAW for numerical Examples. As can be seen from Table 8, when the number of criteria is 3, the alternative rankings of COWEB and BWM are the same. In AHP, the order of the second and third alternatives is reversed. Similarly, in the example with 4 criteria, the alternative rankings of COWEB and AHP are the same. This time, the rankings of BMW differ. In the final example, the alternative rankings of the three methods are the same. Accordingly, COWEB demonstrated resilience to the rankings of the alternatives when the number of criteria is varied. In the other methods, the ordering of the three alternatives changes with the number of criteria. Additionally, a sensitivity analysis is performed for the example with three criteria. In this context, the criteria weights are increased or decreased by 10% in each scenario. This allows us to analyze the impact of minimal changes in criteria weights on the analysis results. The sensitivity analysis for numerical example is exhibited in Table 9. Table 9Sensitivity analysis for numerical Example.COWEBBWMAHP10%−10%10%−10%10%−10%SN1SN2SN3SN1SN2SN3SN1SN2SN3SN1SN2SN3SN1SN2SN3SN1SN2SN3ALTRN1111111211211112112ALTRN2333333122122323323ALTRN3222222333333231231 Sensitivity analysis for numerical Example. As can be seen from Table 9, while the ranking of the alternatives remains constant in COWEB, some rankings differ in BWM and AHP. This demonstrates the robustness of the model. Analysis This section shows that the results of method and comparative with SIWEC by various fuzzy sets. Within the scope of the numerical example of the manuscript, the criteria affecting the fuel consumption of aircraft are evaluated. These criteria are determined as aerodynamic efficiency (AEREFF), weight (WEIGHT), engine efficiency and technology (ENEFTC), new aircraft types and modernization (NAIRMD), pilot behaviour and flight management (PBFMNG), cruise altitude and speed selection (CRASPS), route planning, taxiing and ground maneuvers (RPTGMN), and environmental and meteorological conditions (ENVMCN). New technologies and data analysis techniques regarding fuel consumption that have emerged in recent years allow for more accurate examination and mitigation of the effects of these factors29. Factors affecting aircraft fuel consumption encompass not only technical but also operational and environmental components. Elements such as wing geometry, fuselage form, surface roughness, and laminar flow/turbulence management reduce drag, resulting in lower fuel consumption. According to a report by the International Council on Clean Transportation52, new-generation aircraft manufactured after 2020 use an average of 6% less fuel than older models. Each additional kilogram of payload directly increases fuel consumption. This has a negative impact due to both the payload weight (passenger, cargo) and the “fuel carrying cost” resulting from carrying more fuel. Xiong et al.29 state that modeling the weight-fuel relationship for different flight phases is crucial. The aircraft’s weight distribution (CG) affects flight stability. Furthermore, carrying unnecessary fuel paradoxically increases fuel costs. Excess fuel means more weight and more fuel consumption. Engine specific fuel consumption (SFC) is a critical variable that determines low fuel consumption. New-generation turbofan engines offer better thermodynamic efficiency and lower fuel consumption. New-generation turbofan engines and hybrid electric propulsion systems provide fuel savings of up to 10% by increasing thermodynamic efficiency53. The structural and technological improvements offered by new aircraft types result in lower fuel consumption compared to older aircraft. According to the ICCT report “Fuel Burn of New Commercial Jet Aircraft: 1960 to 2024,” new aircraft can be, on average, 6% more efficient than previous generations compared to production standards. Bernardo and Fageda’s54 study, while examining the impact of new aircraft on CO₂ emissions, explains the need to balance lower fuel consumption with sectoral expansion. A comparison of the A380 and the A350 revealed that the A350 is approximately twice as efficient under the same load condition19. Pilot flight management choices can significantly affect fuel consumption. In particular, “eco-flight” methods, including the best climb profile, cruise at constant speed, and continuous descent during descent, provide fuel savings. Furthermore, practices such as using a single engine during the taxi phase provide significant fuel savings on short-haul flights. Therefore, providing fuel efficiency awareness training to pilots by airlines plays a significant role in improving operational efficiency. Air density decreases at higher altitudes, which reduces drag and saves fuel. However, this choice must be balanced against performance limits, fuel planning, and flight safety limits. Some studies show that aircraft deviation from optimal speed can have significant effects on fuel consumption. Approaches such as optimal climb rates, avoiding erratic maneuvers and power changes, and continuous descent during descent reduce fuel consumption. Xiong et al.29 examine the impact of flight profile choices on fuel cost by examining the climb, dilution, and descent phases separately with fuel consumption models. The required distance may not be the shortest route; Wind direction, airspace restrictions, meteorological conditions, and air traffic control (ATC) restrictions play a role in shaping the route. Wang et al.2, while developing a comprehensive fuel forecasting framework, show that route-based forecast errors can be reduced to levels as low as 0.1%. Factors such as taxi time at the airport, engine power usage, braking, and waiting times also affect fuel consumption. The single-engine taxi strategy can provide significant fuel savings, especially for large aircraft. Fuel use depends not only on the technical specifications of the aircraft but also on how flight operations are planned. Factors such as route selection, flight altitude, speed profile, and taxi times can significantly affect total fuel consumption. Xiong et al.29 showed that machine learning methods increase the accuracy of flight profile-based fuel estimation and thus eliminate planning errors. Headwind reduces the effective speed of flight, causing more fuel consumption, while tailwind provides savings. Zhao et al.22 modeled the effects of meteorological variables (e.g., wind, outside air temperature) on fuel efficiency during the landing approach phase. In addition, high temperature and low air pressure reduce engine efficiency by reducing air density. In this case, the aircraft may have to use more fuel to produce more thrust. They included meteorological variables in the fuel estimation model along with factors such as pressure altitude, speed, and distance. The primary criterion for selecting a decision maker is industry experience. Furthermore, experience in a decision-making position is a significant factor in selecting a decision maker. Evaluations are collected from ten decision makers with at least 15 years of experience in flight management and planning. Half of the experts are pilots, and the other half are airline managers as operations managers and engineers. Pilots have been pilots for at least 15 years, while managers have been managers or engineers in airline companies for at least 17 years. The linguistic evaluations are shared in Table 10. Table 10Linguistic Evaluations.AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM1VLIVLI1VLIEILIELILIDM2VLIVLI1ELIRLILILILIDM31ELIRLILIRLIVLIVLIVLIDM4ELILIEIELI1LILIVLIDM5LIVLI1LILIELIVLIELIDM6ELIVLIRLI1RLILIVLILIDM7ELIVLIEILI1ELIELIELIDM8LIELI1LIEIELILIELIDM9VLILI1LIRLILIVLIELIDM10LILIRLIVLI1LILILI Linguistic Evaluations. Weighting of criteria by real numbers The linguistic evaluations in Table 10 are converted to real numbers in Table 1 to obtain the evaluation matrix defined in Eq. (1). The evaluation matrix is illustrated in Table 11. Table 11Evaluation matrix (Real Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.250.251.000.251.000.330.200.33DM20.250.251.000.200.500.330.330.33DM31.000.200.500.330.500.250.250.25DM40.200.331.000.201.000.330.330.25DM50.330.251.000.330.330.200.250.20DM60.200.250.501.000.500.330.250.33DM70.200.251.000.331.000.200.200.20DM80.330.201.000.331.000.200.330.20DM90.250.331.000.330.500.330.250.20DM100.330.330.500.251.000.330.330.33 Evaluation matrix (Real Number). Afterwards, the elements in Table 11 are normalized with the help of Eq. (2). In other words, the element is divided by the row total. For example, first element is The normalized matrix is displayed in Table 12. Table 12Normalized matrix (Real Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.070.070.280.070.280.090.060.09DM20.080.080.310.060.160.100.100.10DM30.300.060.150.100.150.080.080.08DM40.050.090.270.050.270.090.090.07DM50.110.090.340.110.110.070.090.07DM60.060.070.150.300.150.100.070.10DM70.060.070.300.100.300.060.060.06DM80.090.060.280.090.280.060.090.06DM90.080.100.310.100.160.100.080.06DM100.100.100.150.070.290.100.100.10 Normalized matrix (Real Number). Next, the elements ​​of each criterion are summed. In this manuscript, since the weights of the decision makers are assumed to be equal, Eq. (3) is used. In other words, first z value equals to 0.07 + 0.08 + 0.30 + 0.05 + 0.11 + 0.06 + 0.06 + 0.09 + 0.08 + 0.1 = 1.01 Finally, the weights of criteria are defined by Eq. (5). The weight of AEREFF is calculated as The results are shared in Table 13. Table 13Sums and weights of criteria (Real Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNz1.010.792.541.072.140.850.810.78Weight0.1010.0790.2540.1070.2140.0850.0810.078 Sums and weights of criteria (Real Number). The most essential criteria affecting the fuel consumption of aircraft are engine efficiency and technology and pilot behaviour and flight management with 0.254 and 0.214, respectively. For the comparative analysis, the SIWEC method is performed. Based on Table 11, the SIWEC steps are applied. In other words, the values ​​are divided by the largest value in the matrix. They are then multiplied by the standard deviation and summed. The sum is normalized to obtain the criterion weights from SIWEC. The comparative results between COWEB and SIWEC with real numbers are presented in Fig. 1. Comparative results with real numbers. When Fig. 1 is examined, the order of the criteria for the two methods is the same. Additionally, the correlation between the weight values ​​is 0.999. This demonstrates the reliability of the results and demonstrates the accuracy of the proposed method. Weighting of criteria by triangular fuzzy numbers The linguistic evaluations in Table 10 are transformed into Triangular fuzzy numbers in Table 3 to obtain the evaluation matrix formed in Eq. (6). The evaluation matrix is shown in Table A1 in appendix. Next, the elements in Table 13 are normalized with the help of Eq. (7). First element is calculated as . The normalized matrix is expressed in Table A2 in appendix. Next, the elements ​​of each criterion are summed using Eq. (8). Finally, the weights of criteria are defined by Eqs. (10) and (11). The results are shared in Table 14. Table 14Sums and weights of criteria (Triangular fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNz0.3240.8872.4370.1330.7482.2971.5422.6044.4130.4261.1302.7851.2912.3594.2700.2700.8362.4630.1830.8002.3720.1770.6362.141Fuzzy Weight0.0140.0890.5610.0060.0750.5290.0670.2601.0150.0180.1130.6410.0560.2360.9830.0120.0840.5670.0080.0800.5460.0080.0640.493Weight0.2210.2030.4470.2570.4250.2210.2110.188 Sums and weights of criteria (Triangular fuzzy Number). The most important criteria affecting the fuel consumption of aircraft are engine efficiency and technology and pilot behaviour and flight management with 0.447 and 0.425, respectively. For the comparative analysis, the SIWEC method is performed. Based on Table 9, the SIWEC steps are applied. In other words, the values ​​are divided by the largest c value in the matrix. They are then multiplied by the standard deviation and summed. The sum is normalized to obtain the criterion weights from SIWEC. The comparative results between COWEB and SIWEC with triangular fuzzy numbers are presented in Fig. 2. Comparative results with triangular fuzzy numbers. When Fig. 2 is examined, the order of the criteria for the two methods is the similar. Only the fourth and fifth priorities are reversed in the two methods. This demonstrates the reliability of the results and demonstrates the accuracy of the proposed method. Weighting of criteria by pythagorean fuzzy numbers The linguistic evaluations in Table 10 are transformed into Pythagorean fuzzy numbers in Table 3 to create the evaluation matrix formed in Eq. (12). The evaluation matrix is given in Table A3 in appendix. Afterwards, the elements in Table A3 are defuzzified with the help of Eq. (13). First element is established as . The other elements are estimated with similar calculation. The defuzzified matrix is summarized in Table 15. Table 15Defuzzified matrix (Pythagorean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.680.682.000.681.851.130.441.13DM20.680.682.000.441.481.131.131.13DM32.000.441.481.131.480.680.680.68DM40.441.131.850.442.001.131.130.68DM51.130.682.001.131.130.440.680.44DM60.440.681.482.001.481.130.681.13DM70.440.681.851.132.000.440.440.44DM81.130.442.001.131.850.441.130.44DM90.681.132.001.131.481.130.680.44DM101.131.131.480.682.001.131.131.13 Defuzzified matrix (Pythagorean fuzzy Number). Next, the elements in Table 15 are normalized using Eq. (14). The normalized matrix is shared in Table 16. Table 16Normalized matrix (Pythagorean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.080.080.230.080.220.130.050.13DM20.080.080.230.050.170.130.130.13DM30.230.050.170.130.170.080.080.08DM40.050.130.210.050.230.130.130.08DM50.150.090.260.150.150.060.090.06DM60.050.080.160.220.160.130.080.13DM70.060.090.250.150.270.060.060.06DM80.130.050.230.130.220.050.130.05DM90.080.130.230.130.170.130.080.05DM100.120.120.150.070.200.120.120.12 Normalized matrix (Pythagorean fuzzy Number). After that, the total normalized value ​​of each criterion is computed via Eq. (15). Finally, the weights of criteria are defined by Eq. (16). The results are shared in Table 17. Table 17Sums and weights of criteria (Pythagorean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNz1.020.892.141.171.961.010.940.88Weight0.1020.0890.2140.1170.1960.1010.0940.088 Sums and weights of criteria (Pythagorean fuzzy Number). The most important criteria affecting the fuel consumption of aircraft are engine efficiency and technology and pilot behaviour and flight management with 0.214 and 0.196, respectively. Based on Table A3, the weights are established from SIWEC. In other words, the values ​​are defuzzified and divided by the maximum value in the matrix. Normalized values are multiplied by the standard deviation and then summed. The total value is normalized to obtain the criterion weights from SIWEC. The comparative results between COWEB and SIWEC with Pythagorean fuzzy numbers are visualized in Fig. 3. Comparative results with pythagorean fuzzy numbers. When Fig. 3 is examined, the order of the criteria for the two methods is the same. Moreover, the correlation between the weight values ​​is 1. This demonstrates the reliability of the results and demonstrates the accuracy of the proposed method. Weighting of criteria by fermatean fuzzy numbers The linguistic evaluations in Table 10 are transformed into Fermatean fuzzy numbers in Table 4 to create the evaluation matrix constructed in Eq. (16). The evaluation matrix is exhibited in Table A4 in appendix. Afterwards, the elements in Table A4 are defuzzified with the help of Eq. (15). First element is obtained as . The other elements are computed with similar process. The defuzzified matrix is displayed in Table 18. Table 18Defuzzified matrix (Fermatean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.580.582.000.581.730.940.000.94DM20.580.582.000.001.340.940.940.94DM32.000.001.340.941.340.580.580.58DM40.000.941.730.002.000.940.940.58DM50.940.582.000.940.940.000.580.00DM60.000.581.342.001.340.940.580.94DM70.000.581.730.942.000.000.000.00DM80.940.002.000.941.730.000.940.00DM90.580.942.000.941.340.940.580.00DM100.940.941.340.582.000.940.940.94 Defuzzified matrix (Fermatean fuzzy Number). Then, the normalized elements are calculated using Eq. (15). The normalized matrix is illustrated in Table 19. Table 19Normalized matrix (Fermatean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNDM10.080.080.270.080.240.130.000.13DM20.080.080.270.000.180.130.130.13DM30.270.000.180.130.180.080.080.08DM40.000.130.240.000.280.130.130.08DM50.160.100.330.160.160.000.100.00DM60.000.080.170.260.170.120.080.12DM70.000.110.330.180.380.000.000.00DM80.140.000.310.140.260.000.140.00DM90.080.130.270.130.180.130.080.00DM100.110.110.160.070.230.110.110.11 Normalized matrix (Fermatean fuzzy Number). Next, the total value of normalized criterion is calculated using Eq. (15). Finally, the weights of criteria are defined by Eq. (21). The results are shared in Table 20. Table 20Sums and weights of criteria (Fermatean fuzzy Number).AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNz0.920.812.541.142.270.830.840.65Weight0.0920.0810.2540.1140.2270.0830.0840.065 Sums and weights of criteria (Fermatean fuzzy Number). The most significant criteria affecting the fuel consumption of aircraft are engine efficiency and technology and pilot behaviour and flight management with 0.254 and 0.227, respectively. Using the values in Table A4 in appendix, the weights are defined from SIWEC. In other words, the values ​​are defuzzified. The defuzzified values are divided by the maximum defuzzified value in the matrix. Normalized values are multiplied by the standard deviation and then summed. The total value is normalized regarding SIWEC. The comparative results between COWEB and SIWEC with Fermatean fuzzy numbers are shown in Fig. 4. Comparative results with fermatean fuzzy numbers. When Fig. 4 is examined, the order of the criteria for the two methods is the same. In addition, the correlation value between two model’s results ​​is 1. This demonstrates the reliability of the results and demonstrates the accuracy of the proposed method. Comparative priorities In summary, the priorities of the criteria are obtained by ranking the weight values ​​obtained from the method. The criteria priorities of COWEB and SIWEC analysis according to different number systems are given in Table 21. Table 21Criteria priorities.AEREFFWEIGHTENEFTCNAIRMDPBFMNGCRASPSRPTGMNENVMCNOur Method47132568Our Method with Triangular Fuzzy Sets47132568Our Method with Pythagorean Fuzzy Sets47132568Our Method with Fermatean Fuzzy Sets47132658SIWEC47132568SIWEC with Triangular Fuzzy Sets57132468SIWEC with Pythagorean Fuzzy Sets47132568SIWEC with Fermatean Fuzzy Sets47132658 Criteria priorities. Examining the priorities in Table 20, the method’s results and SIWEC’s results for real numbers are identical. This increases the reliability of the method. The consistency of the proposed method’s results with other number systems demonstrates the robustness of the method. Furthermore, the variation in the SIWEC method’s priority results across number systems demonstrates the superiority of the proposed method. A minor shift was observed between the “Cruise Altitude and Speed Selection” (CRASPS) and “Route Planning, Taxiing, and Ground Maneuvers” (RPTGMN) criteria under the Fermatean fuzzy set in both COWEB and SIWEC analyses. This variation can be explained by the broader representational flexibility of Fermatean fuzzy numbers, which allow experts to express uncertainty over a wider confidence domain than Pythagorean or Triangular fuzzy sets. As a result, small differences in expert hesitation or partial agreement may slightly affect the normalization process and alter adjacent rankings when criteria have very close importance levels. However, the relative weight differences between CRASPS and RPTGMN remain minimal (below 0.01), confirming that the model maintains high stability across all fuzzy environments. This consistency indicates that the COWEB framework is robust, and that minor positional changes among similarly weighted criteria do not undermine the overall reliability or interpretive validity of the results. Moreover, Best-Worst model is used for comparative results. The criterion with the highest value is selected as the best criterion by averaging the evaluations of each decision maker as the data. Similarly, the criterion with the lowest value is selected as the worst criterion. This is how the model’s inputs are obtained. The comparative results of the three methods are given in Table 22. Table 22Comparative results with BWM.Best-Worst ModelSIWECCOWEBAEREFF0.1000.0960.101WEIGHT0.0570.0780.079ENEFTC0.3320.2570.254NAIRMD0.1330.1030.107PBFMNG0.1990.2250.214CRASPS0.0800.0830.085RPTGMN0.0660.0800.081ENVMCN0.0330.0770.078 Comparative results with BWM. Pearson and Spearman correlation coefficients are examined for similarity between the proposed method and the other two methods. The Spearman correlation coefficient is 1 with both models. The Pearson correlation coefficient between the BWM and the proposed model is 0.960, while the Pearson correlation coefficient between the SIWEC and the proposed model is 0.999. Sensitivity analysis The results are presented in the previous sections, with the decision makers taken equally. For the sensitivity analysis, different weight values ​​are obtained and analyzed, provided that the total weight of the decision makers is 1. In other words, Eq. (4) is used instead of Eq. (3). Ten scenarios were constructed by systematically assigning a substantially higher weight to a different decision maker in each scenario (for example, a weight of 0.50 for one expert and 0.055 for each of the remaining experts), while ensuring that the total weight always summed to 1. The results are shared in Table 23. Table 23Sensitivity Analysis.Sc1Sc2Sc3Sc4Sc5Sc6Sc7Sc8Sc9Sc10AEREFF4444444444WEIGHT7777777777ENEFTC1111111111NAIRMD3333333333PBFMNG2222222222CRASPS5555555555RPTGMN6666666666ENVMCN8888888888 Sensitivity Analysis. The criteria priorities remained completely unchanged across all ten scenarios, demonstrating that the COWEB model’s results are highly robust to variations in the assigned weights of the decision makers. This stability indicates that even when one expert is given a disproportionately large influence in the weighting structure, the resulting ranking of the criteria does not shift in any way. In other words, the model consistently produces the same priority order regardless of how the relative importance of the decision makers is altered. Such invariance confirms that the COWEB framework is not sensitive to fluctuations in expert weights and that its outcomes are driven primarily by the underlying structure of the method rather than by the dominance of any single evaluator. Discussion The criteria most affecting aircraft fuel consumption are “engine efficiency and technology.” This result demonstrates that engine performance is a decisive factor in terms of fuel economy and environmental sustainability in aviation. Engine efficiency depends on multidimensional factors such as the efficiency of the combustion process and the level of technological sophistication. Therefore, even small advances in engine technologies can lead to significant reductions in fuel costs in the long run. The importance of this finding is frequently emphasized in the literature55. It is stated that new-generation engine designs, hybrid propulsion systems, and lightweight composite materials significantly reduce fuel consumption. In this context, the results offer important implications from both academic and practical perspectives. First and foremost, at the policy level, aviation authorities and governments should prioritize research support programs that encourage the development of engine technologies56. Furthermore, tax advantages, subsidies, and carbon credits that encourage the use of environmentally friendly and highly efficient engines can be expanded. From a strategic perspective, it is critical that airlines prioritize engine efficiency as a priority criterion in their fleet renewal policies. Optimizing periodic engine maintenance processes in existing fleets, digitizing engine performance monitoring systems, and supporting them with AI-powered predictive maintenance applications will also increase fuel efficiency57. Furthermore, collaborations between manufacturers and universities can accelerate the development of innovative solutions in engine technologies. Consequently, the “engine efficiency and technology” finding should be viewed not only as a technical assessment but also as a strategic priority that can form the cornerstone of energy efficiency, cost reduction, and environmental sustainability policies. Another important finding from this study is that pilot behavior and flight management are among the factors that most significantly impact aircraft fuel consumption. This result demonstrates that not only technical infrastructure or equipment but also human factors play a decisive role in fuel efficiency. Pre-flight planning and speed management directly determine pilots’ impact on fuel consumption. This finding is frequently emphasized in the literature58. It is particularly noted that even small changes in pilot decision-making processes can result in fuel savings of 3 to 10%. This demonstrates that human-centered strategies, as well as technical modernization, should form the basis of sustainable aviation policies. At the policy level, it is crucial for airline authorities and regulatory bodies to mandate fuel efficiency-focused pilot training programs. Simulation-based training can increase fuel consumption awareness and foster more efficient flying habits59. Furthermore, analyzing pilots’ flight data with performance-based monitoring systems will contribute to the continuous improvement of individual fuel management performance. Strategically, airlines can provide pilots with real-time optimization recommendations by leveraging data analytics and AI-supported decision support systems in their flight management processes. This allows variables such as route planning, speed control, and altitude selection to be managed more effectively. Furthermore, implementing reward and motivation programs that encourage positive pilot behaviors aimed at fuel conservation will support a culture of sustainability60. Ultimately, pilot behavior and flight management demonstrate that human performance, along with technical innovations, should be central to fuel efficiency policies; in this regard, they are considered a critical element at both operational and strategic levels. Beyond the leading criteria, the analysis also provides meaningful insights regarding the relative positioning of other factors. Although aircraft weight was expected to play a major role in fuel consumption, its lower ranking can be attributed to recent advances in material science and aircraft design. The widespread adoption of lightweight composite structures and modern load optimization systems has already mitigated much of the weight-related inefficiency in current fleets. Similarly, factors such as cruise altitude and speed selection and route planning appear in the middle range of priorities because airlines have already implemented standardized procedures and digital optimization tools to manage these aspects effectively. The relatively lower priority of environmental and meteorological conditions reflects the limited controllability of these external factors by airlines, even though they remain operationally significant. Taken together, the overall priority structure suggests that airlines should direct their strategic investments primarily toward enhancing engine technology, pilot training, and intelligent flight management systems, while maintaining incremental improvements in aircraft design, route optimization, and operational sustainability programs. This distribution of priorities highlights a shift from purely physical efficiency measures toward human–technological integration as the main driver of fuel efficiency and sustainability in modern aviation. Conclusion This study aims to address the gap in the literature regarding the identification of the most critical factors affecting aircraft fuel consumption and their prioritization. To achieve this purpose, a comprehensive literature review is conducted, and eight key criteria affecting fuel consumption are identified. A new decision-making model, called COWEB, is developed to determine the relative importance of the identified criteria. This model offers a unique framework where the most important criterion is compared with other factors, providing a more reliable, flexible, and shorter evaluation process compared to existing weighting methods. Furthermore, the model can be integrated with Triangular, Pythagorean, and Fermatean fuzzy set structures, enabling more realistic analysis of uncertainties. Analysis results indicate that the most significant factors affecting aircraft fuel consumption are “engine efficiency and technology” and “pilot behavior and flight management.” While the COWEB technique produces consistent results across different fuzzy sets, differences are observed in comparative analyses conducted with the SIWEC method. This finding demonstrates that the COWEB model is a stable and reliable analysis tool. The study contributes to the literature from both methodological and practical perspectives. On the one hand, it offers a theoretical innovation by proposing a new weighting approach, and on the other hand, it provides a scientific basis for determining strategic priorities for fuel efficiency in the aviation sector. The research focused solely on the aviation sector, and the obtained results are limited to the dynamics of the sector. Therefore, testing the applicability of the developed COWEB model across different sectors will increase its generalizability. Future studies could expand the model’s practical application areas by conducting similar analyses in different fields such as automotive, energy, or maritime. However, one of the main limitations of the proposed model is that it focuses solely on calculating criteria weights. The current model does not rank alternatives, and therefore, the strategic applicability of the identified criteria is not directly tested. To overcome this limitation in future research, the developed model could be integrated with alternative-based multi-criteria decision-making methods. For example, identified strategies could be defined as alternatives and ranked. Such an approach would enable not only the importance of criteria but also the determination of the most appropriate strategic option for decision-makers. Furthermore, future studies could increase the model’s methodological flexibility by testing its applicability in uncertain environments beyond fuzzy sets. Such extensions will contribute to the COWEB technique becoming a more powerful and comprehensive decision support tool at both theoretical and practical levels. Although the proposed model effectively prioritizes the criteria influencing aircraft fuel consumption, it is limited to the weighting phase and does not incorporate explicit alternatives such as aircraft types or operational strategies. This restriction stems from the study’s aim to identify fundamental factors rather than rank alternatives. Nevertheless, this limitation also indicates a potential extension for future research, where the COWEB method could be integrated with alternative-based MCDM techniques (e.g., TOPSIS, MARCOS, or VIKOR) to evaluate specific aircraft models or operational scenarios. Such an extension would allow the framework to provide not only criterion-level insights but also actionable strategic decisions for airlines. One of the theoretical and practical limitations of the COWEB method is that the best criterion can only be selected by the decision maker. If the expert cannot determine the best criterion, the method forces them to choose one. Theoretically, if more than one criterion is paramount, the model becomes limited. Since the COWEB model does not rely on a full pairwise comparison matrix, a traditional mathematical consistency ratio such as the Consistency Ratio used in AHP or the consistency index applied in BWM cannot be computed. This is because COWEB requires only comparative assessments with respect to a single reference criterion rather than evaluating all criteria against each other. Therefore, consistency in the classical sense is not applicable to the structure of this model. Instead, the internal reliability of COWEB is demonstrated through its robustness across different fuzzy environments and its strong correlation with widely used weighting methods such as SIWEC and the Best–Worst Method. The stability of rankings and the high similarity of weight distributions across multiple evaluation settings confirm that the model provides coherent and dependable results even without a traditional consistency ratio. Although the COWEB model provides important advantages in terms of simplicity and reduced cognitive effort, it is not superior to AHP or BWM in every aspect. The most fundamental limitation is that COWEB does not generate a full pairwise comparison matrix, and therefore it cannot compute a traditional consistency index such as the Consistency Ratio in AHP or the consistency measure used in BWM. For example, when evaluating eight criteria, AHP requires 28 pairwise comparisons and allows the analyst to explicitly measure the logical coherence of these comparisons through the Consistency Ratio. Similarly, BWM requires a structured set of comparisons between the best and worst criteria, enabling a mathematical consistency check based on deviation minimization. COWEB, on the other hand, requires only seven comparisons when eight criteria are used because all evaluations are anchored to the most important criterion rather than being compared with each other. This design greatly reduces expert workload but also removes the possibility of measuring internal inconsistency in the classical sense. Therefore, COWEB is more suitable for decision-making contexts where expert time is highly limited, while AHP and BWM remain more appropriate when consistency measurement and full comparison frameworks are essential. This trade-off clarifies that COWEB should be seen as a complementary approach rather than a universally superior replacement for AHP or BWM. While not for every problem, rapid decision-making is important for some strategic decisions. In some decision-making problems, speed and efficiency are more important than consistency. In real life, while consistency is necessary for some problems, rapid decision-making is crucial for others. The newly developed COWEV may not be the optimal choice for complex problems requiring consistency. Techniques such as AHP and BWM can be considered for solving such complex issues. However, some problems require rapid resolution. In this case, although AHP and BWM can perform detailed analysis, they require significant time. This can be an obstacle to achieving the goal of rapid decision-making. Therefore, a new approach capable of faster analysis is needed in addition to techniques such as AHP and BWM. The COWEB technique makes a significant contribution to these issues with its rapid analysis. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Author contributions S.E., F.K., S.Y., H.D. and İ.Y. wrote the main manuscript text and S.E., F.K. and S.Y. prepared figures and tables. All authors reviewed the manuscript. Data availability All data analyzed during this study is included in this article. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All data analyzed during this study is included in this article.