Green transformation decision support system based on civil aviation carbon emission characteristics

Abstract

The invention discloses a green transformation decision support system based on civil aviation carbon emission characteristics, and the system comprises a multi-source data fusion module which is used for integrating Chinese civil aviation core element data, geographic data and carbon emission data; the core element dynamic optimization configuration model module is used for realizing the minimization of the net present value of the total cost of the system in the planning period by embedding a civil aviation core element dynamic optimization configuration model under the perspective of carbon cost; the model solving module is used for solving the core element dynamic optimization configuration model; the space-time visualization module is used for realizing visual display of civil aviation element space distribution and carbon emission space-time characteristics; and the scene simulation and iteration module automatically updates the model through parameter adjustment and feeds back an optimization result. Through the scheme, the carbon emission reduction target and the operation benefit can be effectively balanced, the core resource configuration accuracy, the decision response efficiency and the anti-risk capability are greatly improved, and an omnibearing technical support is provided for green transformation and efficient operation of the civil aviation industry.

Description

Technical Field

[0001] This invention generally relates to the field of carbon emission technology. More specifically, this invention relates to a green transition decision support system based on the carbon emission characteristics of civil aviation. Background Technology

[0002] The decision-making process for the green transformation of the civil aviation industry needs to take into account operational efficiency, emission reduction benefits and dynamic adaptability. However, the problems with existing technical solutions directly restrict the scientific nature and feasibility of the decision-making.

[0003] In existing decision-making models, carbon costs are not integrated into a unified optimization framework along with investment and operating costs. Furthermore, carbon prices are often static and fixed, failing to couple carbon market fluctuations (such as spot price fluctuations of 40-80 yuan / ton) with regionally differentiated emission reduction policies (such as strict quota management in key regions). This leads to an imbalance between economic and emission reduction objectives, resulting in either failure to meet emission reduction targets or excessive increases in operating costs. Simultaneously, existing models primarily focus on local optimization of single elements (fleet, routes, airports), neglecting the dynamic relationships between them (such as the constraint coupling between fleet renewal and route carbon emissions, and airport adaptability). They particularly lack long-term planning adaptability, failing to support dynamic scenarios such as fleet iteration and technological upgrades (such as SAF application) across cycles up to 2060, resulting in significant deviations between local and global optima. Moreover, core civil aviation data, geospatial data, and carbon emission data exhibit multi-source heterogeneity, and existing tools lack efficient spatiotemporal alignment mechanisms, leading to inconsistencies in data spatiotemporal dimensions (coordinate deviations, inconsistent timestamps). In particular, carbon emission data can only calculate the total amount for an industry / enterprise, lacking refined data at the levels of flight routes, aircraft types, and time periods. This makes it impossible to accurately depict the spatiotemporal heterogeneity of emissions, affecting the quality of model input. Furthermore, with existing tools supporting only a single set of scenario parameters, it is difficult to simulate complex scenarios such as carbon tax adjustments, demand fluctuations (±30%), and efficiency improvements of new aircraft models. Moreover, parameter updates rely on manual operation, and when key parameters such as carbon market prices and regional carbon quotas change, the model cannot be automatically triggered to iterate and solve the problem. The decision-making response lags behind policy and market dynamics, and the weak visualization and interaction capabilities also limit the practicality of decision-making.

[0004] In view of this, there is an urgent need to provide a green transformation decision support system based on the characteristics of civil aviation carbon emissions, so as to construct a targeted mechanism that integrates carbon cost quantification, multi-factor collaboration, refined data support and dynamic iteration. Summary of the Invention

[0005] In order to at least address one or more of the technical problems mentioned above, this invention proposes a green transition decision support system based on the characteristics of carbon emissions in civil aviation in several aspects.

[0006] In a first aspect, the present invention provides a green transformation decision support system based on the characteristics of civil aviation carbon emissions, characterized by comprising: a multi-source data fusion module, used to integrate core element data, geographic data, and carbon emission data of Chinese civil aviation, achieving data alignment through spatial interpolation and temporal smoothing, and outputting the data in the form of a distance matrix; a core element dynamic optimization configuration model module, which minimizes the net present value of the total system cost within the planning period by embedding a dynamic optimization configuration model of civil aviation core elements from the perspective of carbon cost; the dynamic optimization configuration model of civil aviation core elements adopts a method of coupling graph theory network modeling and mixed integer dynamic programming, wherein the total cost includes investment cost, operating cost, and dynamic carbon cost; a model solving module, which solves the dynamic optimization configuration model of core elements using mathematical modeling system tools, and outputs fleet configuration, route flight frequency, and carbon emission results for each planning period; a spatiotemporal visualization module, which visualizes the spatial distribution of civil aviation elements and the spatiotemporal characteristics of carbon emissions using geographic information system tools; and a scenario simulation and iteration module, which automatically updates the model and provides feedback on optimization results by supporting parameter adjustments for different carbon policies and demand fluctuation scenarios.

[0007] In a second aspect, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the functions of the green transition decision support system based on the carbon emission characteristics of civil aviation as described in any one of the first aspects.

[0008] The green transition decision support system based on civil aviation carbon emission characteristics, as described above, enables the global unified quantification and dynamic adaptation of carbon costs by embedding a dynamic optimization configuration model of civil aviation core elements from a carbon cost perspective into the core element dynamic optimization configuration model module. Furthermore, in some embodiments, the dynamic carbon cost and multi-dimensional constraints in the core element dynamic optimization configuration model allow for the dynamic coordination of multiple elements and the coupling of long-term constraints. By improving the accuracy of multi-source data fusion and the granularity of carbon emission characterization, and by implementing scenario simulation and dynamic iterative feedback mechanisms, the accuracy of green transition decisions can be improved. Attached Figure Description

[0009] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0010] Figure 1 An exemplary structural block diagram of a green transition decision support system based on the carbon emission characteristics of civil aviation, according to an embodiment of the present invention, is shown. Figure 2 An exemplary structural block diagram of an electronic device according to an embodiment of the present invention is shown. Detailed Implementation

[0011] 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, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Figure 1 An exemplary structural block diagram of a green transition decision support system 100 based on the carbon emission characteristics of civil aviation, according to an embodiment of the present invention, is shown.

[0013] like Figure 1 As shown, the green transformation decision support system 100 includes: a multi-source data fusion module 101, which integrates core element data of China's civil aviation, geographic data and carbon emission data, achieves data alignment through spatial interpolation and temporal smoothing, and outputs the data in the form of a distance matrix.

[0014] Specifically, the multi-source data fusion module is the core data input support of the system. Its core objective is to solve the problems of spatiotemporal inconsistency, insufficient accuracy, and poor correlation of multi-source heterogeneous data in civil aviation. It provides highly consistent and accurate foundational data for subsequent models through a complete technical solution of "data acquisition - preprocessing - spatiotemporal alignment - quality verification - standardized output". In one implementation scenario, the specific implementation method is as follows: First, data collection and source adaptation must integrate at least the following three core data types to cover all dimensions of civil aviation elements, geospatial data, and carbon emissions. All data sources are authoritative and compliant. Specific details are as follows: Core Civil Aviation Data: Industry-wide data from 2020 to 2024 was obtained from the Statistical Information Center of the Civil Aviation Administration of China, including geographic information (latitude and longitude coordinates, runway length, terminal capacity, number and type of aircraft stands) for 248 transport airports in 31 provinces (municipalities and autonomous regions), parameters for 15 types of aircraft from 18 major airlines (such as A320neo, B737-800, B787-9, etc., with parameters including number of seats, fuel efficiency curves, engine models, maximum takeoff weight, and monthly utilization limit), and operational data for 1,687 scheduled routes nationwide (average daily flight frequency, actual flight distance, takeoff and landing airports, load factor, and crew configuration standards in 2024). Real-time flight dynamic data (takeoff / landing time, delay duration, and real-time fuel consumption) is synchronized from the airlines' Flight Operations Control (FOC) systems, with an update frequency of once per minute.

[0015] Geospatial data: Using the 1:100,000 scale electronic map from the National Geographic Information Public Service Platform, topographic elevation data and administrative boundary data of the national airspace at a 1km×1km grid were obtained. Vegetation cover and surface temperature data of the flight area can be extracted from data of Fengyun-3F (FY-3F) and Gaofen series satellites to correct the environmental impact factors in carbon emission calculations.

[0016] Carbon emission data: The basic data comes from the Civil Aviation Administration's "Annual Report on Energy Conservation and Emission Reduction in the Civil Aviation Industry". The refined data is calculated by the "bottom-up" method - combining flight trajectory data (flight altitude, speed, attitude) from the ADS-B Automatic Dependent Surveillance-Broadcast System and measured data from the Aircraft Engine Emission Monitoring System (EEMS), and calculating the carbon emissions per flight by route-aircraft type-time period, with data granularity reaching "route-aircraft type-15 minutes".

[0017] Furthermore, to address issues such as noise, bias, and missing data in multi-source data, categorized preprocessing techniques can be employed: Spatial data correction: To address the issue of original coordinate deviations (±80-120m) in some small and medium-sized airports, spatial interpolation optimization was performed using the inverse distance weighted method (IDW). Using high-precision GPS coordinates (deviation ≤10m) of 30 hub airports, including Beijing Daxing and Shanghai Pudong, as reference control points, a search radius of 5km and a weighting coefficient p=2 were set to correct the coordinates of 120 small and medium-sized airports. The final coordinate deviation was controlled within ±30m, meeting the requirement of "precise waypoint positioning" in route planning. Kriging interpolation was used to fill in missing values ​​in the terrain elevation data, ensuring that the airspace grid data integrity was ≥99%.

[0018] Time-series data smoothing: Flight operation data (such as delay duration and fuel consumption) are susceptible to outliers due to weather and air traffic control. A combination of "3σ criterion detection + weighted moving average smoothing" can be used. Specifically, the 3σ criterion (mean ± 3 standard deviations) is first used to screen for extreme outliers (such as a single delay exceeding 4 hours or fuel consumption deviating from the mean by 50%). Then, a weighted moving average method with a 24-hour window is used—with a weight of 0.7 for the nearest 12 hours and 0.3 for the furthest 12 hours within the window. This preserves the time-series trend while eliminating sudden interference, achieving an outlier removal rate of over 95%. For carbon emission time-series data, linear interpolation is used to fill short-term gaps (missing time ≤ 1 hour), while long-term gaps (> 1 hour) are replaced with the mean of data from the same route, aircraft type, and period, improving data integrity to 98.5%.

[0019] Attribute data standardization: This involves standardizing fleet parameters provided by different airlines (e.g., fuel efficiency units such as kg / km and L / seat). (100km) will be uniformly converted to "kg / seat" according to ICAO standards. km. The actual flight distance data is used uniformly (instead of the straight-line distance). The actual mileage of the flight segment is calculated based on the ADS-B trajectory data, and the correction error is less than a predetermined threshold, for example, the threshold is set to 2%.

[0020] Secondly, a three-dimensional alignment technique combining "timestamp unification + spatial grid matching + attribute association mapping" is employed to achieve spatiotemporal alignment and standardized output, ensuring consistency across multi-source data. Specifically, this may include: Time dimension alignment: All time-series data are uniformly converted to 1-minute timestamps in UTC+8 time zone. For example, the flight departure time is accurate to "2024-08-15 07:32:00". Carbon emission data is mapped to specific timestamps according to flight phase (takeoff, climb, cruise, descent), realizing the precise correlation of "time-event-data".

[0021] Spatial Dimension Alignment: Using the EqualEarth projected coordinate system, the national airspace is divided into 1km×1km grid units, with each grid assigned a unique ID. Airport locations, flight routes, and carbon emission points are all mapped to corresponding grids, achieving spatial association through grid IDs. For example, for an A320neo operating the Beijing-Shanghai route (actual distance 1318km), its carbon emission data during the cruise phase is mapped to over 1200 grids located between 36°-31° North latitude and 116°-121° East longitude.

[0022] Attribute-dimensional association: The cosine similarity algorithm (threshold ≥ 0.95) is used to establish an association mapping of "aircraft type-route-carbon emission-airport". For example, the carbon emission data of "B787-9" aircraft type, "Beijing-Los Angeles" route, and "2024-08-15 14:00-16:00" cruise phase are associated with the wide-body aircraft stand resource data of Beijing Capital Airport to ensure unique matching of multi-source data for the same operational event.

[0023] Furthermore, a three-level verification mechanism can be established to ensure data quality. Specifically: Level 1 verification (completeness) requires single-type data coverage ≥98%; Level 2 verification (consistency) requires spatiotemporal alignment accuracy ≥99%; and Level 3 verification (accuracy) requires the deviation of key parameters (such as aircraft fuel efficiency and route distance) from measured data to be ≤3%. Data failing verification is automatically returned to the preprocessing stage for re-optimization. The final output format is an Excel distance matrix and a CSV data table. The matrix dimensions are "airport × airport × aircraft type × cycle," including core parameters such as route distance, carbon emission coefficient, and aircraft stand suitability. The data table is stored in a "timestamp-grid ID-data type-value" structure, supporting direct reading by tools such as GAMS and Python.

[0024] Furthermore, in the core element dynamic optimization configuration model module 102, a civil aviation core element dynamic optimization configuration model from a carbon cost perspective is embedded to minimize the net present value of the total system cost during the planning period. The civil aviation core element dynamic optimization configuration model adopts a method that couples graph theory network modeling with mixed integer dynamic programming, wherein the total cost includes investment cost, operating cost, and dynamic carbon cost.

[0025] The core element dynamic optimization configuration model module is the decision-making core of the green transformation decision support system of this invention. In one implementation scenario, it adopts a coupling technology of "graph theory network modeling + mixed integer dynamic programming" to achieve long-term dynamic optimization of core civil aviation elements (fleet, airport, routes). Its core innovation lies in the deep coupling of dynamic carbon costs with multi-dimensional constraints to ensure optimal synergy between "economy and emission reduction". The specific implementation method is as follows: First, in a graph theory network modeling embodiment, the civil aviation transportation system is abstracted as a weighted undirected graph G=(V,E,W), and the element association mapping is realized through the definition of nodes, edges, and weights. Specifically, this includes: Node (V) Definition: This includes two types of nodes: First, airport nodes (248 in total), whose attributes include airport capacity (maximum number of takeoffs and landings per hour, e.g., 60 flights / hour at Beijing Capital Airport, 55 flights / hour at Shanghai Hongqiao Airport), aircraft stand type (number of narrow-body aircraft stands, number of wide-body aircraft stands), and regional emission reduction level (key area / cooperative area / general area); Second, aircraft type nodes (15 types in total), whose attributes include number of seats (e.g., 186 seats for the A320neo, 335 seats for the B787-9), fuel consumption per unit hour (2.3 tons / hour for the A320neo, 4.1 tons / hour for the B787-9), and carbon emission factor per unit hour (…). The A320neo has a CO2 / hour capacity of 5.8 tons, and the B787-9 has a CO2 / hour capacity of 10.5 tons. (Monthly utilization limit) Narrow-body aircraft: 75 hours / month; Wide-body aircraft: 85 hours / month.

[0026] Edge (E) definition: An edge represents a flight route between two airport nodes (1687 in total), and its attributes include the actual distance of the route (e.g., Beijing-Shanghai 1318km, Guangzhou-Chengdu 1690km), predicted passenger flow, etc. Based on passenger flow data from the past five years, the ARIMA model predicts that, for example, the average daily passenger flow on the Beijing-Shanghai route will be 12,000 in Q1 2025, and the expected load factor will be... Trunk routes 0.85, feeder routes 0.75), regional policy coefficient ( (1.3 for routes in the Beijing-Tianjin-Hebei region, 1.2 for routes in the Yangtze River Delta region, and 1.0 for other routes).

[0027] Weight (W) definition: The edge weight is the comprehensive cost of "operating cost + carbon cost", where operating cost includes fuel cost, crew cost and airport charges, and carbon cost is dynamic carbon price × total carbon emissions. The weight value is dynamically adjusted according to the planning cycle and scenario parameters. For example, the weight of the Beijing-Shanghai route (operated by A320neo) in Q1 2025 is RMB128,000 / flight, and the weight is adjusted to RMB145,000 / flight in Q1 2030 (after the carbon price increase).

[0028] Secondly, the objective function and parameters are defined. The objective function, for example, focuses on minimizing the net present value of the total system cost over the period of 2030-2060 (31 years, divided into 124 planning periods by quarter). Specific parameters and calculation examples are as follows:

[0029] In the above formula, z represents the net present value of the total system cost, which is the net present value of all costs in the planning period (2030-2060) discounted to 2060, and is the optimization objective (minimization) of the model. t represents the planning period number. For example, if the planning period is divided into quarters, Q1 of 2030 is t=1, Q4 of 2060 is t=124, and there are a total of 124 periods.

[0030] Discount rate ( ): The discount rate for period t. It is dynamically set according to the planning stages: 3.85% for 2030-2035 ("16th Five-Year Plan"), 3.65% for 2036-2040 ("17th Five-Year Plan"), 3.45% for 2041-2045 ("18th Five-Year Plan"), 3.25% for 2046-2050 ("19th Five-Year Plan"), and 3.05% for 2051-2060 ("20th Five-Year Plan"), in line with the long-term interest rate trend.

[0031] Total annual cost ( ): Total cost for year t. This includes investment costs, operating costs, and dynamic carbon costs, and is the core cost item for a single period. A specific calculation example is taken for Q1 of 2025 (t=20): a. Investment cost: The purchase cost of 10 newly introduced A320neo aircraft (US$110 million / aircraft × exchange rate of 6.9 = RMB759 million / aircraft) + the residual value recovery of 5 retired old B737-700 aircraft (US$30 million / aircraft × exchange rate of 6.9 = RMB207 million / aircraft), the net investment cost = 10 × 759 - 5 × 207 = RMB6.555 billion.

[0032] b. Operating costs: Fuel costs (A320neo operating the Beijing-Shanghai route, fuel consumption per flight: 5.2 tons × 8 yuan / L = 41,600 yuan / flight × 14 flights per week × 13 weeks = 7,572,800 yuan) + Crew costs (2 pilots + 3 flight attendants, hourly wages of 1,000 yuan and 400 yuan respectively, flight duration 2.5 hours × 14 flights × 13 weeks = 20,020,000 yuan) + Airport fees (peak-hour narrow-body aircraft parking fees: 12,000 yuan / flight × 14 flights × 13 weeks = 2,184,000 yuan). Annual operating cost per route = 7,572,800 + 2,002 + 2,184,000 = 29,776,800 yuan. The total operating cost for 1,687 routes nationwide is approximately 50.2 billion yuan.

[0033] c. Dynamic carbon cost: A320neo single-flight carbon emissions are 28 tons of CO2 × dynamic carbon price (benchmark price 70 yuan / ton × carbon market volatility coefficient 1.1 × regional policy coefficient 1.2 = 92.4 yuan / ton) × 14 flights × 13 weeks = 28 × 92.4 × 182 = 465,800 yuan / route, and the total carbon cost for all routes nationwide is about 7.85 billion yuan.

[0034] SALV is the residual value of assets, such as the remaining value of the fleet at the end of 2060, which is uniformly calculated at 15% of the purchase price of the aircraft.

[0035] In summary, Q1ANNC(t) in 2025 = 65.55 + 502 + 78.5 = 646.05 billion yuan.

[0036] SALV refers to the residual value of assets. For example, the remaining value of the fleet at the end of 2060, calculated uniformly at 15% of the purchase price of the aircraft, is approximately RMB 23 billion.

[0037] Furthermore, the core element dynamic optimization configuration model adopts multi-dimensional constraint implementation rules, specifically including four types of core constraints. These constraints are deeply linked to mathematical expressions and actual operational rules to ensure the feasibility of the optimization results. Specifically: Passenger flow balance constraint: For each airport node i, the difference between inflow and outflow passenger flow in period t shall not exceed 10% of the airport's transfer capacity, as expressed mathematically:

[0038] in, For the departing passenger flow of airport i in period t, For the arrival passenger flow of airport i in period t, Let represent the transfer capacity of airport i in period t (e.g., Beijing Daxing Airport's transfer capacity is 20,000 passengers / day). For example: Beijing Capital International Airport in Q1 2025. =8 million visits, =7.8 million passenger trips, transit capacity =1.5 million passenger trips, with a constraint requiring the difference between inflow and outflow to be ≤150,000 passenger trips. The model ensures this constraint is met by adjusting the frequency of flight routes. Predict passenger flow for route j to i in period t. The total number of passengers traveling from airport j to airport i in period t can be predicted using the ARIMA model.

[0039] Capacity supply and demand constraints: The capacity of a single route must cover 1.05 times the predicted passenger flow (with room for fluctuations), the mathematical expression is: ,in, A(t) is the set of aircraft models, which is the set of all aircraft models that participate in operation during the t-th period, and can include 15 mainstream aircraft models. The frequency of flights operated by aircraft type a on route i to j in cycle t. The number of seats for aircraft type A is the standard layout passenger seat number for aircraft type A, which can be obtained from the Civil Aviation Administration's aircraft type airworthiness database. The expected load factor for aircraft type a operating on routes i to j in period t is a reasonable target value that can be set based on historical load factor data, with higher values ​​for trunk routes than for feeder routes. 1.05 is the capacity reserve coefficient, a safety factor set to cope with fluctuations in passenger demand (such as last-minute bookings) to ensure that capacity covers extreme demand.

[0040] The following is a specific example illustrating the Guangzhou-Chengdu route in Q1 2025. =500,000 passengers / quarter, A320neo ( =186 seats), A330-300 ( (301 seats) operated the flight. The values ​​are 0.85 and 0.88 respectively, calculated by the model. =84 classes / quarter =26 shifts / quarter, total capacity = 84×186×0.85+26×301×0.88≈525,000 passenger trips, which meets the constraint of 1.05×50=525,000 passenger trips.

[0041] Fleet / crew constraints: The monthly flight time of each aircraft type shall not exceed the utilization limit, and the available fleet time shall not exceed the crew's flight capacity. The mathematical expression is as follows: (Quarterly constraint, 3 represents the number of months in the quarter), and simultaneously satisfy... .

[0042] In the above formula, The flight time for a single flight on route i to j operated by aircraft type a (e.g., A320neo operating the Beijing-Shanghai route h=2.5 hours, B787-9 operating the Beijing-Los Angeles route h=12 hours). Let be the fleet size of aircraft type a in the t-th period. 3 represents the upper limit of monthly utilization rate for model a, and 3 is the periodic conversion factor.

[0043] The total number of crew configurations for aircraft type a in period t (based on mainstream airline standards, 4 crews per narrow-body aircraft and 6 crews per wide-body aircraft, such as Air China's A320neo fleet). =50 aircraft, =50×4=200 sets); - The maximum monthly flight time for a single crew is strictly in accordance with the requirements of CCAR-121: pilots shall not fly more than 100 hours per month and 900 hours per year (referring to the actual operating standards of a certain airline, taking an average of 70-80 hours per month as the model baseline); flight attendants shall not fly more than 110 hours per month and 1200 hours per year, and shall be required to have at least 36 hours of continuous rest within 7 consecutive calendar days. This refers to the maximum monthly flight time for a single flight crew, for example: Air China's B787-9 fleet in Q1 2025. =30 aircraft =85 hours / month =30 × 6 = 180 sets =80 hours / month, maximum quarterly crew availability = 180 × 80 × 3 = 43200 hours. Maximum quarterly fleet availability = 30 × 85 × 3 = 7650 hours. The model takes the minimum of the two as the upper limit of the constraint to ensure a balance between crew flight compliance and fleet utilization.

[0044] Supplementary dynamic constraints: Fleet updates must meet the following requirements Furthermore, the annual retirement rate is ≤5% (referencing the fleet renewal standards of China Eastern Airlines and China Southern Airlines). The flight time in the first quarter after the introduction of new aircraft models will be calculated at the upper limit of 80%, with time reserved for crew training and aircraft model integration.

[0045] Carbon quota constraint: The total carbon emissions of a region in period t shall not exceed the policy quota, expressed mathematically as follows: ,in: The carbon emission factor per unit hour for model a (based on EEMS measured data, such as 5.8 tons of CO2 / hour for A320neo and 6.2 tons of CO2 / hour for B737-800). The emission reduction coefficients for routes i to j are set as follows: 1.3 for the Beijing-Tianjin-Hebei route, 1.2 for the Yangtze River Delta route, and 1.0 for other regions, based on the "Provincial Greenhouse Gas Emission Quota Allocation Scheme". The carbon allowance is calculated for a region's quarterly quota. For example, the civil aviation allowance for East China in Q1 2025 is 50 million tons × (1-5%)^(2025-2020) = 38.68 million tons. For instance, in Q1 2025, an airline in East China operated 120 routes, including Shanghai-Beijing, with a total flight time of 86,000 hours. Its total carbon emissions were 86,000 × 5.8 × 1.2 = 593,760 tons, which did not exceed the airline's allocated regional allowance of 1.2 million tons, thus meeting the requirements.

[0046] Furthermore, in the model solving module 103, a mathematical modeling system ( GAMS The tool solves the dynamic optimization configuration model of the core elements, outputting fleet configuration, flight frequency, and carbon emission results for each planning cycle. Specifically, a coupled architecture of "mixed-integer dynamic programming + adaptive genetic algorithm" can be used to efficiently solve high-dimensional decision variables (15 aircraft types × 1687 routes × 124 cycles) and multi-constraint coupled problems. The details are explained below: 1. Optimize the algorithm parameters to better suit the company's needs, including: Mixed-integer dynamic programming parameters: The problem is decomposed into layers according to the planning period. For the short-term period (1-5 years), a fine time step (quarterly) is used, with the state variables defined as "fleet size + airport capacity + remaining carbon allowances". For the long-term period (6-30 years), a coarse time step (annual) is used, and the state variables are simplified to "fleet structure + regional emission reduction progress", reducing computational complexity. A Bellman equation stage payoff function is introduced. ,in The penalty coefficient for exceeding carbon emission standards is set at twice the carbon price, such as 92.4 yuan / ton × 2 = 184.8 yuan / ton.

[0047] Adaptive genetic algorithm parameters: The population size is dynamically set to 800-1200 based on the decision variable dimensions. The crossover probability (Pc) and mutation probability (Pm) are dynamically adjusted based on the fitness value: the top 20% of excellent individuals (low-cost, high-emission-reduction solutions) have Pc=0.6 and Pm=0.03; the middle 60% of ordinary individuals have Pc=0.9 and Pm=0.07; and the bottom 20% of inferior individuals have Pc=0.95 and Pm=0.1. The crossover operation adopts "route cluster grouping crossover" (divided into 7 major regions such as North China and East China to avoid the logical break of cross-regional aircraft type-route matching), and the mutation operation incorporates the actual operating rules of airlines (such as narrow-body aircraft prioritizing mutation of short-route frequencies and wide-body aircraft prioritizing mutation of long-route aircraft types).

[0048] 2. Optimize the solution process and efficiency, including: 1) Initialization: Based on the actual operating data (aircraft type-route matching, flight frequency) of three airlines, namely Air China and China Eastern Airlines in 2024, 30% of the initial population was generated, and the remaining 70% was randomly generated and screened through hard constraints (removing solutions with excessive airport capacity and crew timeout). The initial feasibility rate was ≥85%.

[0049] 2) Iterative solution: Dynamic programming decomposes the problem periodically, and outputs a subset of feasible solutions that satisfy the time constraints in each period. Genetic algorithm performs discrete optimization on the feasible solutions. Through the collaborative mechanism of "local optimum within the period - global integration across periods", the local optimum trap is avoided.

[0050] 3) Convergence judgment: Set dual convergence conditions: the total cost fluctuation is ≤0.1% and the constraint satisfaction rate is ≥99.8% for 50 consecutive iterations, and the solution time is controlled within 1.5 hours (70% efficiency improvement compared to the single dynamic programming algorithm).

[0051] 4) Solution Verification: Compare the optimization results with China Southern Airlines' actual operating data in 2024. The frequency deviation of a single route is ≤10% and the carbon emission error is ≤3% to ensure the practicality of the solution.

[0052] Furthermore, in the spatiotemporal visualization module 104, the spatial distribution of civil aviation elements and the spatiotemporal characteristics of carbon emissions are visualized using ArcGIS tools. Specifically, in one embodiment, the spatiotemporal visualization module can integrate multiple tools such as ArcGIS, Python, and SigmaPlot to achieve an intuitive presentation of data and optimization results. The specific implementation details are as follows: 1. Spatial Visualization Implementation. First, based on the 1km×1km grid layer of ArcGIS, airport nodes, route networks and carbon emission heat maps are overlaid. A hierarchical color scheme is used to display carbon emission intensity (low intensity ≤3 tons CO2 / grid, medium intensity 3-8 tons, high intensity >8 tons). For example, in Q1 2025, the high emission grids in the Beijing-Tianjin-Hebei airspace are concentrated within a 10km radius of Beijing Capital International Airport and Beijing Daxing International Airport, which intuitively reflects the key areas for regional emission reduction.

[0053] Secondly, the system dynamically displays the spatial distribution of the fleet configuration, using different colors to mark aircraft types (red for narrow-body aircraft, blue for wide-body aircraft, and green for regional aircraft). Clicking on an airport node allows users to view detailed data such as the frequency of takeoffs and landings of the aircraft type, the occupancy rate of parking spaces, and carbon emissions for that airport. It also supports multi-period comparisons (such as changes in fleet structure between 2025 and 2030).

[0054] 2. Data Visualization and Interactive Functions. First, by integrating SigmaPlot with Python, multi-dimensional analytical charts are generated: a cost-emissions reduction benefit scatter plot (horizontal axis represents total system cost, vertical axis represents carbon emission reduction, and the optimal solution is marked under different carbon policy scenarios), an aircraft type carbon emission comparison curve (showing the carbon emission trends of 15 types of aircraft per flight), and a bar chart of periodic quota usage progress, to support decision-makers in quickly identifying core optimization directions. Second, an interactive interface has been developed, supporting data filtering by region (e.g., the Yangtze River Delta), route type (trunk / feeder), and aircraft type. Users can customize the time range (e.g., 2025-2030) to view dynamic changes. Charts can be exported to PDF and Excel formats, adapting to the needs of airline strategic planning report preparation.

[0055] Furthermore, in the scenario simulation and iteration module 105, the model is automatically updated and optimization results are fed back by supporting parameter adjustments for different carbon policies and demand fluctuation scenarios. In one embodiment, the scenario simulation and iteration module supports multi-dimensional scenario parameter adjustments across "policy-market-technology" to achieve dynamic model adaptation, as detailed below: 1. Scene parameters and company adaptation settings include the following scenarios.

[0056] Policy scenarios: carbon tax rate (50-200 yuan / ton), regional carbon quota adjustment (±10%), emission reduction rate target (5%-15%), referencing the dynamics of national carbon market policies and the actual carbon quota acquisition by airlines (e.g., Air China's carbon quota utilization rate in 2024 was 85%).

[0057] Market scenarios: fluctuations in passenger demand (±30%, peak seasons during Spring Festival / summer travel season), and fluctuations in jet fuel prices (6-12 yuan / L), based on passenger flow forecast data from OTA platforms and trends in international crude oil futures prices.

[0058] Technical scenarios: Improved fuel efficiency of new aircraft models (15%-35%), SAF compatibility ratio (0%-50%), referencing the technical parameters of new aircraft models such as the Airbus A350neo and Boeing 777X, as well as the actual SAF refueling ratio of China Eastern Airlines (reaching 3% in 2024).

[0059] 2. Regarding the dynamic iteration and feedback mechanism, a monthly parameter update mechanism is first established. By automatically synchronizing carbon price data from the Shanghai Environment and Energy Exchange, carbon quota adjustment announcements from the Civil Aviation Administration of China, and airline fleet renewal plans, the model is triggered to re-solve, and the iteration results are pushed to the visualization module in real time. Secondly, it supports custom scenario combination simulations, such as the scenario of "carbon tax of 150 yuan / ton + SAF adaptation of 30% + demand growth of 20%". The model outputs the fleet renewal plan (e.g., adding 15 A320neo aircraft in 2026), route frequency adjustment plan (increasing trunk route frequency by 10%), and carbon emission reduction (22% reduction compared to the baseline scenario) under this scenario, and generates a feasibility analysis report, including cost calculation, policy compliance assessment, and operational risk warnings (e.g., crew training needs, airport adaptation and transformation), providing full-process support for airlines' green transformation decisions.

[0060] Through the embodiments provided by this invention, the collaborative innovation among the modules enables multi-dimensional optimization, including: the multi-source data fusion module employs categorized preprocessing, three-dimensional spatiotemporal alignment, and three-level quality verification technologies to effectively address the issues of spatiotemporal inconsistency and insufficient accuracy of multi-source heterogeneous data in civil aviation, providing high-consistency and high-accuracy basic data support for subsequent models. The core element dynamic optimization configuration model, through graph theory network modeling and mixed-integer dynamic programming coupling technology, deeply integrates dynamic carbon costs with multi-dimensional constraints, achieving long-term dynamic collaborative optimization of fleets, airports, and routes, achieving synergistic optimality between economic costs and emission reduction benefits. The model solving module adopts a dual-algorithm coupling architecture to improve the solution efficiency of high-dimensional decision variables, avoid local optimum traps, and ensure the practicality of optimization results. The spatiotemporal visualization module integrates multiple tools to achieve intuitive presentation of data and results, supports multi-dimensional interactive analysis, and helps decision-makers quickly grasp the core optimization direction. The scenario simulation and iteration module supports multi-dimensional scenario parameter adjustment and dynamic iterative feedback, ensuring that the system can adapt to policy, market, and technological changes in real time, providing scientific, efficient, and long-term adaptable decision support for the green transformation of civil aviation.

[0061] Furthermore, in one embodiment, the alignment process of the multi-source data fusion module includes: matching the multi-source data in a spatial grid of 1km×1km size and timestamps of 1-minute time level to achieve consistency in the spatial-temporal dimensions, so that the data coverage is greater than a preset threshold.

[0062] First, a unified spatiotemporal benchmark is established for multi-source data. For example, a dedicated spatiotemporal coordinate system for civil aviation is established, spatially based on the National Geodetic Coordinate System 2000 (CGCS2000), and coordinate transformation is performed on the original geographic data (such as airport coordinates and flight routes). Temporally, atomic clock synchronization technology in the UTC+8 time zone is adopted to ensure that the timestamp error of all data is ≤10ms.

[0063] Secondly, a hierarchical spatiotemporal alignment strategy is implemented. Different alignment logic is adopted for different data types (static data such as airport capacity, dynamic data such as flight fuel consumption). For example, static data can be aligned according to "quarterly updates + grid mapping", where the spatial grid can be set at 1km×1km. Dynamic data can be aligned in real time according to "1-minute sliding window".

[0064] Finally, a data coverage guarantee mechanism is implemented to ensure data coverage. By setting a multi-level data completion strategy, when the data coverage of a single grid is lower than a preset threshold (such as 95%), a three-level completion process of "interpolation of similar grid data in the same area - adaptation of historical data from the same period - supplementation from authoritative data sources" is automatically triggered.

[0065] To help those skilled in the art better understand the above solution, the process is further described below with reference to a specific embodiment: Taking the multi-source data fusion of East China (16 airports including Shanghai, Hangzhou, and Nanjing) in Q1 2025 as an example, the specific implementation steps are as follows: 1. Unification of spatiotemporal reference Spatial reference calibration: The original coordinate data of 16 airports in East China were obtained (some small and medium-sized airports used the WGS84 coordinate system with a deviation of ±50-100m). The coordinate data were uniformly converted to the CGCS2000 coordinate system using ArcGIS's "Coordinate Transformation Tool". Shanghai Pudong Airport (CGCS2000 coordinates: 31°14′N, 121°58′E) was used as the reference control point. The seven-parameter method was used for error correction, and the final coordinate deviation of all airports was ≤15m.

[0066] Time reference synchronization: Deploy Network Time Protocol (NTP) servers in the Flight Operation Control System (FOC), ADS-B trajectory monitoring system, and carbon emission monitoring system to synchronize with the atomic clock of the National Time Service Center. The timestamp is calibrated every 10 minutes to ensure that the timestamp error of flight take-off / landing time, fuel consumption collection time, and carbon emission calculation time is ≤5ms.

[0067] 2. Hierarchical spatiotemporal alignment Static data alignment: Static data such as airport capacity (e.g., Shanghai Hongqiao Airport's maximum number of takeoffs and landings per hour is 45) and aircraft parameters (e.g., the number of seats in an A320neo is 186) are mapped to attributes using a 1km×1km spatial grid. Each grid is associated with airport resources and aircraft type compatibility information within a 3km radius. The grid attributes are updated quarterly to ensure that the matching degree between static data and spatial grid is ≥99%.

[0068] Dynamic data alignment: Flight dynamic data (such as real-time flight altitude and fuel flow) are aligned using a "1-minute sliding window". For example, within the time window of 08:00-08:01, ADS-B trajectory data (1 record every 5 seconds) and fuel consumption data (1 record every 10 seconds) of the flight are collected during this period. The data is then uniformly completed into 1 record per minute through linear interpolation and mapped to a 1km×1km grid of the corresponding flight area to achieve precise correlation between "time window-grid-data".

[0069] 3. Data coverage guarantee Real-time coverage monitoring: A data quality monitoring platform is built to statistically analyze the coverage of various data types (such as flight trajectory data coverage and carbon emission data coverage) in each 1km×1km grid in real time, with a preset threshold of 95%. When the carbon emission data coverage of a certain grid around Hangzhou Xiaoshan Airport (30°15′N, 120°10′E) is found to be only 88%, the data completion process is triggered.

[0070] The specific completion process can achieve three levels of completion execution: Level 1: Using the inverse distance weighted method (IDW), the carbon emission data of five similar grids (flying areas of the same route and aircraft type) with a coverage of ≥98% are used as a benchmark to calculate weighting coefficients for interpolation and completion, thereby increasing the coverage to 92%. Level 2: Retrieve historical carbon emission data for the same period in Q1 2024 for this grid (coverage 97%), and supplement missing period data through time series similarity matching (cosine similarity ≥ 0.92), increasing the coverage to 95.5%; Level 3: If the target is still not met, the system will automatically connect to the Civil Aviation Administration's carbon emission database to supplement the missing key data in the grid, so that the final coverage rate can be stabilized at 99.2%, which is greater than the preset threshold of 95%.

[0071] Through the aforementioned technological expansion and implementation, the spatiotemporal alignment accuracy and data availability of the multi-source data fusion module have been significantly improved: Spatially, airport coordinate deviation has been reduced from ±50-100m to ≤15m, and the accuracy of route segment and grid mapping has reached 99.5%, solving the problem of "spatial misalignment leading to carbon emission calculation deviation" in traditional data; Temporally, data timestamp error is ≤5ms, and dynamic data alignment latency rate at the 1-minute level is <0.3%, meeting the time accuracy requirements for real-time flight operation decisions. Simultaneously, data coverage has been consistently maintained above 95%, and can be improved to above 99% in extreme scenarios through three-level completion, providing highly consistent and complete basic data support for the dynamic optimization configuration model of core elements. This further reduces the error in subsequent fleet configuration and route planning optimization results, effectively ensuring the scientific and accurate nature of civil aviation green transformation decisions.

[0072] Furthermore, in one embodiment, the core element dynamic optimization configuration model includes a dynamic carbon cost, which is calculated by multiplying total carbon emissions by a dynamic carbon price. The dynamic carbon price is determined by coupling a carbon market volatility coefficient and a regional emission reduction policy adjustment coefficient. This will be illustrated below using ARIMA time series model prediction.

[0073] Firstly, regarding the optimization of the ARIMA model and the fusion of multiple factors, at least three related factors—jet fuel prices, GDP growth rate, and carbon quota trading volume—are introduced into the traditional ARIMA model to construct the ARIMA-X multivariate prediction model. Furthermore, the model parameters (p=3, d=1, q=2) are optimized through sliding window cross-validation to improve the prediction accuracy of carbon market volatility coefficient.

[0074] Secondly, the adjustment coefficients for regional emission reduction policies are refined into different levels. A two-dimensional classification system is established based on regional emission reduction pressure (core emission reduction areas, key control areas, and general control areas) and industry contribution (the proportion of civil aviation in regional carbon emissions). The adjustment coefficients are refined into eight gradients within the range of 1.0-1.8, thereby achieving precise policy adaptation.

[0075] Finally, a dynamic carbon cost real-time calibration mechanism will be implemented. Specifically, this can be achieved by connecting to the Shanghai Environment and Energy Exchange's real-time carbon price data, updating the dynamic carbon price hourly, and combining it with real-time flight carbon emission data (collected by the EEMS system every 5 minutes) to establish a minute-level carbon cost calibration model, ensuring that cost accounting is synchronized with actual operations.

[0076] To facilitate a better understanding of the above-described embodiments of the present invention by those skilled in the art, the implementation process is further described below.

[0077] Taking the dynamic carbon cost accounting of an airline in East China (the core emission reduction area) in Q2 2025 as an example, the specific implementation steps are as follows: 1. ARIMA-X model predicts carbon market volatility coefficient Data collection and preprocessing: Data on Shanghai carbon market spot prices (40-80 yuan / ton), jet fuel prices (6-12 yuan / L), GDP growth rate in East China (2.3%-5.8%), and carbon quota trading volume (5 million-12 million tons / quarter) were collected from 2020 to 2024. Z-score standardization was used to eliminate the influence of dimensions, and the moving average method was used to fill in three short-term data gaps.

[0078] Model Training and Parameter Optimization: Using data from 2020-2023 as the training set and 2024 data as the test set, an ARIMA-X model (p=3, d=1, q=2) was constructed, with three factors, including jet fuel prices, as exogenous variables. Parameters were adjusted through 5-fold cross-validation, ultimately reducing the model's prediction error (MAPE) from 8.5% for traditional ARIMA to 5.2%, and outputting a carbon market volatility coefficient of 1.12 for Q2 2025 (based on a benchmark carbon price of 70 yuan / ton).

[0079] 2. Determination of Regional Emission Reduction Policy Adjustment Coefficient Regional classification and coefficient matching: According to the "Implementation Plan for Carbon Emission Reduction in East China", East China is divided into core emission reduction areas (Shanghai, Suzhou, etc.). Civil aviation accounts for 8.2% of the region's carbon emissions, corresponding to the highest gradient in the two-dimensional classification system, and the adjustment coefficient is determined to be 1.5.

[0080] Dynamic adjustment of coefficient: Because the airline exceeded its carbon emission reduction target by 12% in 2024 (5% higher than the industry average), the policy incentive mechanism was triggered, and the adjustment coefficient was reduced by 0.1, ultimately set at 1.4.

[0081] 3. Dynamic carbon pricing and carbon cost accounting Dynamic carbon price calculation: The benchmark carbon price in Q2 2025 is 70 yuan / ton. Combining the carbon market volatility coefficient of 1.12 and the regional adjustment coefficient of 1.4, the dynamic carbon price = 70 × 1.12 × 1.4 = 110.88 yuan / ton.

[0082] Carbon emission calculation: Select the Shanghai-Beijing route of this airline (operated by A320neo), and collect flight data on May 10, 2025 through the EEMS system: flight time 2.5 hours, carbon emission factor per hour 5.8 tons CO2 / hour, total carbon emissions = 2.5 × 5.8 = 14.5 tons.

[0083] Dynamic carbon cost calibration: The Shanghai carbon market real-time price is synchronized hourly (the price at 10:00 on May 10 was 72 yuan / ton). The updated dynamic carbon price = 72 × 1.12 × 1.4 = 113.47 yuan / ton, and the real-time carbon cost = 14.5 × 113.47 ≈ 1645.32 yuan, which is 2.3% lower than the initial calculation value of the day, meeting the calibration accuracy requirement (≤3%).

[0084] The above-described implementation scheme significantly improves the accuracy and timeliness of dynamic carbon cost accounting. By reducing the prediction error of carbon market volatility coefficients, it avoids cost accounting biases caused by traditional static pricing. Refining the regional adjustment coefficient to eight gradients improves policy adaptability by 60%, balancing emission reduction requirements with airline operational realities. Hourly updates to dynamic carbon prices and minute-level calibration of carbon costs achieve real-time synchronization with actual operations. Ultimately, the cost calculation error of the core element optimization model can be reduced from 12% to below 5%, providing accurate cost basis for fleet configuration and route planning, effectively balancing "emission reduction targets - economic costs," and enhancing the scientific rigor and feasibility of civil aviation green transformation decisions.

[0085] Furthermore, in one embodiment, the core element dynamic optimization configuration model includes multi-dimensional constraints, which at least include: passenger flow balance constraints, route capacity supply and demand constraints, fleet monthly utilization rate constraints, crew available flight capability constraints, and regional carbon quota constraints.

[0086] Firstly, based on the data, multi-source fusion data (such as 1km×1km grid spatiotemporally aligned data) is accessed and correlated with dynamic carbon costs (such as ARIMA-X model prediction coupled with policy adjustment coefficients).

[0087] Secondly, the multi-dimensional constraint system includes at least passenger flow balance constraints, route capacity supply and demand constraints, fleet monthly utilization rate constraints, crew available flight capability constraints, and regional carbon quota constraints.

[0088] In addition, multi-objective weighted optimization modeling can be used. Specifically, the weights of the three-dimensional objectives of "maximizing carbon emission reduction, minimizing operating costs, and maximizing capacity utilization" (30%, 25%, and 45%) can be determined by the AHP method, and the objective function and constraint boundary can be quantified.

[0089] In addition, the improved GA-PSO hybrid algorithm can be used to solve the problem: it adopts a switching mechanism of "GA global search + PSO local optimization", introduces adaptive parameters and constraint penalty terms, and improves the solution efficiency and the quality of the optimal solution.

[0090] In addition, the original multidimensional constraints can be decomposed into "static benchmark constraints + dynamic real-time constraints" through a constraint dynamic adaptation system, which clarifies the quantification standards, data sources and trigger adjustment rules of each constraint.

[0091] In addition, a minute-level iterative adjustment mechanism can be established. By establishing a 30-minute closed-loop process of "data synchronization - constraint update - model solving - solution verification - system integration", real-time synchronization between the solution and operations can be achieved.

[0092] The following is a complete implementation process for optimizing 12 core routes (Shanghai-Beijing, Shanghai-Guangzhou, etc.) of an airline in East China (20 A320neo, 15 B737MAX, 10 ARJ21) in Q3 2025: 1. Regarding data access and multidimensional constraint definition Multi-source fusion data access: By connecting to the spatiotemporal fusion database described in the previous embodiment through RESTful API, 286 1km×1km grid data (airport capacity, flight trajectory, carbon emission basic data) covering 12 flight routes are obtained, ensuring that the timestamp error is ≤5ms and the spatial deviation is ≤15m.

[0093] Dynamic carbon cost access: Synchronize with the calculation results of the above embodiment of the invention. The airline's Q3 dynamic carbon price benchmark value is 110.88 yuan / ton, and it receives the real-time price of the Shanghai carbon market every hour (e.g., 115 yuan / ton at 13:00 on August 15).

[0094] Data association mapping: Establish a "grid data-route-aircraft type-dynamic carbon cost" association table. For example, the Shanghai-Beijing route corresponds to 12 core grids, and associates the unit carbon emission factor of 5.8 tons / hour of the A320neo aircraft with the real-time carbon cost.

[0095] In the multi-dimensional constraint settings, the passenger flow balance constraint can be set as follows: the daily average passenger flow fluctuation of a single route ≤ ±5%, and the transfer passenger flow connection time between routes ≥ 45 minutes; the route capacity supply and demand constraint can be set as the route capacity deployment ≥ daily average predicted passenger flow × 1.05 (to ensure load factor) and ≤ daily average predicted passenger flow × 1.2 (to avoid capacity waste); the fleet monthly utilization rate constraint can be set as follows: A320neo ≥ 270 hours / month, B737MAX ≥ 260 hours / month, ARJ21 ≥ 240 hours / month; the crew available flight capacity constraint can be set as the monthly flight time of a single crew ≤ 100 hours, and the continuous flight time per day ≤ 8 hours; furthermore, the regional carbon quota constraint can be set as the airline's Q3 carbon quota limit in East China is 5,000 tons, and the monthly carbon emissions of a single route ≤ 450 tons.

[0096] Furthermore, in one implementation scenario, a multi-objective weighted optimization model is implemented. The weights can be determined by inviting 10 industry experts to score using the 1-9 scale, constructing a judgment matrix, calculating the weight vector [0.3, 0.25, 0.45] (i.e., carbon emission reduction of 30%, cost reduction of 25%, and transportation capacity of 45%), and verifying that the consistency test CR=0.04 (<0.1), indicating that the weights are valid.

[0097] The objective function is quantified and calculated as follows: Carbon emission reduction target: =|Optimized carbon intensity -4.2| (unit: tons of CO2 / kilometer). Cost target: =Fuel ​​cost × 0.6 + Dynamic carbon cost × 0.3 + Operation and maintenance cost × 0.1 (Target ≤ 0.8 yuan / passenger-kilometer).

[0098] Capacity target: =0.4 × occupancy rate + 0.6 × daily utilization rate (target: occupancy rate ≥ 82%, daily utilization rate ≥ 9.5 hours).

[0099] Therefore, the final overall goal is: F = (Normalization process).

[0100] In one embodiment, the model variables can be defined as follows: decision variables are "aircraft type-route" matching relationship (i.e., which aircraft type flies which route), flight frequency (3-10 flights / day), and capacity deployment (number of seats × frequency).

[0101] Furthermore, in one implementation scenario, an improved GA-PSO hybrid algorithm is used to solve the problem. The specific process can be as follows: Algorithm initialization: Based on the airline's Q2 operating data, generate 100 initial solutions (e.g., 7 daily B737MAX flights on the Shanghai-Guangzhou route). The fitness function F' = F - penalty term (0.2F is deducted when any multidimensional constraint is violated).

[0102] Parameter configuration: Population size 100, number of iterations 150, crossover probability linearly decreasing from 0.8 to 0.6 in the GA stage (first 75 iterations), mutation probability decreasing from 0.03 to 0.01; inertia weight decreasing from 0.9 to 0.4 in the PSO stage (last 75 iterations).

[0103] Speed ​​update formula: This formula allows for adjusting the optimization direction and step size of candidate solutions for core element configurations, adapting to high-dimensional decision-making scenarios involving civil aviation fleets, routes, and capacity. parameter The next iteration velocity of the particle represents the adjustment magnitude and direction of the next round of the candidate scheme (particle) for the i-th core element configuration (such as the increase or decrease of flight frequency, the intensity of aircraft type allocation). 0.7 is the inertia weight (w), which controls the degree to which the particle retains the adjustment velocity of the previous round, balancing global search (exploring new schemes) and local optimization (optimizing existing schemes). 1.4 represents the current iteration speed of the particle and the current adjustment range of the i-th candidate solution (e.g., the adjustment speed for a certain flight route is "1 aircraft / iteration"). 1.4 represents the individual learning factor. This controls the intensity of particle learning from its own historical best solutions, reinforcing the influence of individual experience on current adjustments; the next 1.4 is the social learning factor (…). This controls the intensity of particle learning towards the global optimal solution, thereby enhancing the sharing and diffusion of optimal experience among the group. The number of [0,1] intervals is randomly generated as an individual random factor to increase the randomness of individual learning and avoid getting trapped in local optima. This is a social random factor used to randomly generate the number of intervals [0,1], increasing the randomness of group learning and improving search diversity. The best position in the history of the particle is the configuration that the i-th candidate solution has performed best in all previous iterations (such as the best combination of a certain aircraft type and route, or the best flight frequency). The optimal position for the particle swarm is the optimal configuration of core elements among all candidate schemes (the optimal solution that satisfies the three-dimensional objectives of carbon emission reduction, cost reduction, and capacity utilization). Given the current position of the particle, the current configuration status of the core elements of the i-th candidate solution (such as the current aircraft type-route matching relationship, flight frequency, and capacity deployment).

[0104] Solution results: The algorithm converged in 48 iterations (40% less than the traditional GA), the optimal solution had an F value of 0.92, and the initial configuration schemes, such as 8 daily A320neo flights on the Shanghai-Beijing route (275 hours of monthly utilization) and 6 daily ARJ21 flights on the Shanghai-Xiamen route (245 hours of monthly utilization), all satisfied the initial multidimensional constraints.

[0105] Furthermore, in one implementation scenario, the process of implementing the constraint dynamic adaptation system can be as follows: Detailed configuration of static constraints: The monthly fleet utilization constraint can be combined with the aircraft maintenance plan. The A320neo reserves 2 days of maintenance period per month, which is equivalent to 276 hours of available flight time per month. The constraint threshold is adjusted to 270-276 hours / month. The regional carbon quota constraint can be allocated according to the carbon emission ratio of the route. The monthly quota for the Shanghai-Beijing route is 450 tons, and the monthly quota for the Shanghai-Ningbo route is 180 tons.

[0106] Real-time response to dynamic constraints: Passenger flow balance constraints, for example, on August 15th, data from Ctrip and TravelSky showed that the average daily passenger flow on the Shanghai-Shenzhen route increased by 12% compared to the forecast, triggering adjustments to capacity supply and demand constraints (capacity deployment ≥ increased passenger flow × 1.05). Crew available flight capacity constraints, for example, if 3 crews took unforeseen leave that day, reducing available flight time by 15 hours, the model automatically limited the frequency of flights on the Shanghai-Chengdu and Shanghai-Chongqing routes operated by the affected crews. Dynamic weather constraints, for example, when a thunderstorm warning was issued by the meteorological system, the takeoff and landing capacity of Hongqiao Airport was reduced to 33 aircraft / hour (from 48 aircraft / hour), superimposed on the route capacity supply and demand constraint adjustment plan.

[0107] Furthermore, in one implementation scenario, the minute-level iterative adjustment mechanism can be implemented as follows: The 30-minute iteration process can include: 0-10 minutes: Synchronizing real-time carbon price (115 yuan / ton), Hongqiao Airport capacity data, passenger flow forecast updates, and crew leave information. 10-20 minutes: Updating model constraint parameters (Hongqiao capacity coefficient 0.7, Shanghai-Shenzhen passenger flow growth coefficient 1.12, crew availability coefficient 0.85), and resolving. 20-30 minutes: Verifying the solution (satisfying all multidimensional constraints, F-value 0.89 ≥ 0.85 standard), and pushing it to the airline's OMS system.

[0108] The final plan can be adjusted as follows: Aircraft type and frequency: The Hongqiao-Shenzhen route will be replaced by an A320neo instead of a B737MAX (carbon emissions will be reduced by 18%), and the frequency will be reduced from 7 flights to 5 flights (to meet crew capacity constraints); the Pudong-Shenzhen route will be reduced from 6 flights to 9 flights (to meet passenger flow growth constraints).

[0109] Fleet allocation: The ARJ21 has added 2 flights on the Shanghai-Ningbo route, with a monthly utilization rate of 252 hours (in line with the 240-260 hour constraint).

[0110] Carbon quota adaptation: After optimization, the monthly carbon emissions of the Shanghai-Beijing route will be 432 tons (≤450 tons of quota), and the airline's total carbon emissions in Q3 are expected to be 4,850 tons (≤5,000 tons of quota).

[0111] Through the embodiments provided by this invention, supported by multi-source fusion data and dynamic carbon costs, combined with a clear multi-dimensional constraint system, multi-objective weighted modeling, improved hybrid algorithms, and minute-level iterative mechanisms, the efficiency of core element optimization configuration is comprehensively improved. The convergence speed of the solution can be increased by 40%, avoiding local optima and constraint violations, and improving the overall achievement rate of the three-dimensional objectives. The dynamic response delay of multi-dimensional constraints is ≤30 minutes, adapting to unexpected scenarios such as passenger flow, crew changes, and weather conditions. Iterative adjustments synchronize the solution with real-time data, improving route load factors and reducing carbon emissions and operating costs per unit of capacity. Thus, it ensures both constraint compliance and efficiently balances emission reduction targets and operational benefits, significantly improving the scientific and practical nature of civil aviation core resource allocation.

[0112] Furthermore, in one embodiment, the model solving module integrates an adaptive genetic algorithm, which is coupled with mixed-integer dynamic programming to improve the efficiency of solving high-dimensional discrete decision variables; wherein, the adaptive genetic algorithm includes crossover probability and / or mutation probability, which are dynamically adjusted according to the fitness value of the model solution based on the core element dynamic optimization configuration.

[0113] In the Adaptive Genetic Algorithm (AGA), the crossover probability (Pc) and mutation probability (Pm) are dynamically adjusted according to the individual's fitness value, avoiding the shortcomings of traditional genetic algorithms where "fixed probabilities lead to premature convergence or slow convergence." Its key formulas and parameter definitions are as follows: 1. Definition of Fitness Function. First, the fitness function quantifies the quality of an individual (solution). The formula is: Where F' is the adjusted individual fitness value (0-1 points, the higher the better) including constraints and penalties. The original comprehensive objective function value (obtained by weighting carbon emission reduction, cost reduction, and capacity utilization). The number of constraint types (e.g., carbon allowances, crew flight time, etc., in this example $n=5$); For the first Penalty weights for class constraints (set according to constraint importance, hard constraints) Larger); - : No. Class constraint violation flag (when violated) When compliant ).

[0114] 2. The dynamic adjustment formula for crossover probability (Pc) adjusts Pc using a piecewise function based on the interval of an individual's fitness value, achieving "fewer crossovers for high-quality solutions and more crossovers for low-quality solutions":

[0115] In the above formula: Pc is the crossover probability of the current individual (value 0-1); This represents the maximum crossover probability (an empirical value of 0.9, ensuring sufficient recombination of inferior solutions); This represents the minimum crossover probability (an empirical value of 0.5 to avoid damaging high-quality solutions). This represents the maximum fitness value in the current population; This represents the average fitness value of the current population; This represents the current fitness value of the individual.

[0116] 3. The dynamic adjustment formula for the mutation probability (Pm) is consistent with the logic of Pc, adjusting Pm in segments to balance "local exploration" and "global convergence":

[0117] In the above formula, Pm is the mutation probability of the current individual (with a value of 0-0.03 to avoid excessive mutation); This represents the maximum mutation probability (empirical value 0.03, for rapid reconstruction of inferior solutions). This represents the minimum mutation probability (empirical value 0.005, a small improvement for a high-quality solution). , , The meaning is the same as the crossover probability formula mentioned above.

[0118] Furthermore, the core of the coupling technique between AGA and Mixed Integer Dynamic Programming (MIDP) is "MIDP preprocessing + AGA global optimization + cooperative triggering restart". MIDP's strong constraint handling capability reduces the solution complexity of AGA, while AGA's global search capability improves the solution space coverage of MIDP. The specific mechanism is as follows: 1. Core Formula of Coupled Architecture (1) MIDP Solution Space Filtering Formula: MIDP first performs hard constraint filtering on high-dimensional discrete variables to retain compliant solution space: ,in, : The effective solution space after filtering; High-dimensional decision variable vector (including aircraft type-route matching, flight frequency, etc.) , (Number of variables); - : No. Hard constraint functions (such as fleet monthly utilization constraints) ,Right now (hours / months); - Number of hard constraint types (in this example) (Including aircraft performance, airport capacity, and carbon quota constraints).

[0119] (2) Cooperative triggering condition formula: When AGA convergence stalls, MIDP is triggered to repartition the solution space and restart the optimization: ,in,- :continuous The improvement in the optimal solution of the next iteration; : No. The optimal fitness value in the next iteration; : No. The optimal fitness value in the next iteration ( (Trigger threshold, empirical value 30); - Minimum lift threshold (empirical value 0.001, i.e., lift less than 0.1% is considered convergence stagnation).

[0120] 2. Coupled execution process 1) MIDP preprocessing stage: Substituting hard constraint functions ,filter Invalid solutions outside the scope will reduce the solution space from compressed to Magnitude ( (Number of variables).

[0121] 2). AGA Global Optimization Phase: In Within this process, the current optimal solution is output through iterative optimization using the dynamic Pc / Pm formula described above.

[0122] 3) Cooperative Triggering Phase: If the following conditions are met... MIDP breaks down the original problem into Each sub-problem (such as grouping by route) is solved separately and then fed back to AGA to restart global optimization.

[0123] The following is a detailed description of the specific implementation process, taking the optimization of core elements (58 high-dimensional discrete variables, 5 types of constraints) of an airline in East China in Q3 2025 as an example. The implementation steps are as follows: First, initialize the AGA core parameters. 1. Set basic parameters: , , , , , .

[0124] 2. Calculate initial fitness: Based on airline operation data, the initial population consists of 100 individuals. , .

[0125] Secondly, perform dynamic PC / Pm calculation and execution. 1. Select a high-quality individual ( ), calculate Pc and Pm: ;

[0126] Perform crossover / mutation: crossover with other individuals with a probability of 0.52, and random mutation with a probability of 0.0065, preserving superior genes.

[0127] 2. Select a substandard individual ( ), calculate Pc and Pm: - , - Perform crossover / mutation: high probability of crossover recombination of genes, high probability of mutation to escape local optima.

[0128] Secondly, the coupling of AGA and MIDP is implemented. 1. MIDP preprocessing: Substituting hard constraint functions Filtering out violations such as "ARJ21 operating Shanghai-Urumqi (exceeding range)" and "A320neo monthly utilization rate of 280 hours (exceeding the 276-hour limit)" reduces the available solutions. Compress to .

[0129] 2. AGA Iterative Optimization: Iterate 60 times within the effective solution space. On the 30th iteration... During the 60th iteration , This triggers collaboration.

[0130] 3. MIDP Subproblem Decomposition: The 12 routes are divided into two subproblems: "Main Line to Beijing, Shanghai and Guangzhou" and "Surrounding Branch Lines". MIDP is used to solve the optimal solution of the branch line subproblem (ARJ21 will prioritize the deployment of short-haul routes such as Shanghai-Ningbo), and the solution is fed back to AGA.

[0131] 4. AGA Restart Optimization: Combining the optimal solutions to subproblems, after 20 iterations... This meets the target requirements.

[0132] Finally, constraint penalties and fitness calibration are performed on an individual. However, violating carbon quota constraints ( , After calibration: This individual falls into the intermediate solution range, according to... , Participate in subsequent iterations.

[0133] By employing the dynamic probability adjustment formula of the adaptive genetic algorithm in the above embodiments, precise control over solutions of varying quality is achieved, avoiding premature convergence and ineffective iterations. Combined with solution space filtering and subproblem decomposition using mixed-integer dynamic programming (MIDP), the complexity of solving high-dimensional variables is significantly reduced. This coupled architecture leverages both the constraint handling advantages of MIDP and the global optimization capabilities of AGA, enhancing the algorithm's adaptability to multi-constraint, high-dimensional decision-making scenarios in civil aviation (such as aircraft-route matching, flight frequency, and capacity deployment tiers). This solution ensures a stable and efficient solution process, significantly improving the compliance and optimality of the solutions, providing robust algorithmic support for the dynamic optimization of core elements, and fully meeting the practical needs of complex decision-making in civil aviation operations.

[0134] Furthermore, in another embodiment, the model solving module may further include: High-dimensional variable block encoding strategy: Block encoding is performed according to the type of decision variable (aircraft type allocation, route frequency, capacity deployment) to reduce chromosome length and decoding complexity and improve iteration efficiency.

[0135] Fitness stratification dynamic adjustment rule: Divide the fitness value into three categories: "high-quality solution, medium-quality solution, and low-quality solution". Set a differentiated probability adjustment formula for each category to avoid premature convergence or slow convergence caused by a single rule.

[0136] MIDP-AGA collaborative triggering mechanism: MIDP preprocesses high-dimensional variables to filter out solution spaces that violate hard constraints, and then AGA performs global optimization; if AGA convergence stalls during iteration, MIDP is triggered to repartition the solution space to achieve collaborative optimization.

[0137] Constraint-adaptive fitness function: A constraint penalty term is embedded in the fitness function to apply gradient penalty to solutions that violate multidimensional constraints (such as carbon quotas and crew flight time), thereby improving the feasibility of the solution.

[0138] The following is a detailed explanation of the above-described implementation scheme. Taking the optimization of core elements of an airline in East China in Q3 2025 as an example (involving 12 routes, 3 aircraft types, a total of 58 high-dimensional decision variables, including 36 aircraft type-route combinations, 12 flight frequency variables, and 10 capacity deployment tier variables), the implementation steps are as follows: 1. Basic Coupling Setup of AGA and MIDP Define the set of high-dimensional discrete decision variables: including aircraft type-route matching (3 aircraft types × 12 routes = 36 variables), route frequency (12 variables, with values ​​of 3-10 flights / day), and capacity deployment tiers (10 variables, divided into 5 tiers according to the number of seats: 1=150 seats, 2=168 seats, ..., 5=186 seats).

[0139] Construct a basic coupled architecture: Integrate AGA and MIDP algorithms in the solution module. MIDP is responsible for decomposing the high-dimensional optimization problem into two sub-problems: "machine type allocation" and "frequency optimization". AGA is responsible for global optimization across sub-problems and synchronizes intermediate solution results in real time through internal data interfaces.

[0140] 2. Implementation of high-dimensional variable block coding Using a binary-decimal hybrid encoding method, the chromosome is designed in three parts according to variable type: The first block (aircraft type code): 36-bit binary, with each 3 bits corresponding to the aircraft type of one route (001=A320neo, 010=B737MAX, 100=ARJ21). The second block (frequency code): 12-bit decimal, with each bit directly corresponding to the flight frequency of one route (values ​​from 3 to 10). The third block (capacity code): 10-bit decimal, with each bit corresponding to one tier of capacity deployment. After encoding, the chromosome length is reduced from 232 bits in the traditional single-code approach to 58 bits, significantly improving decoding efficiency.

[0141] 3. Implementation of Fitness Stratification and Dynamic Probability Adjustment Fitness intervals are defined: based on the F-value of the comprehensive objective function (0-1 points), three types of intervals are defined: excellent solution (F≥0.85), medium solution (0.6≤F<0.85), and poor solution (F<0.6).

[0142] Define the probability rules for differentiation: High-quality solution: Pc=0.5-0.7, Pm=0.005-0.01 (low crossover, low variation, preserving high-quality genes); Medium solution: Pc = 0.7-0.9, Pm = 0.01-0.02 (medium crossover and mutation, promoting gene recombination); Poor solution: Pc=0.9, Pm=0.02-0.03 (high crossover and high mutation, invalid gene reconstruction).

[0143] Dynamic adjustment: Every 20 iterations, the population fitness distribution is recalculated, and the corresponding probability intervals are automatically matched. For example, if 30% of individuals are high-quality solutions in a certain iteration, their Pc is automatically set to 0.6 and Pm to 0.008.

[0144] 4. MIDP-AGA collaborative triggering implementation, including: MIDP Preprocessing: In the initial iteration stage, MIDP is used to filter out solutions that violate hard constraints (such as "ARJ21 operates ultra-long-haul routes" or "single route frequency exceeds 10 flights"), thus reducing the initial solution space from... Level compressed to 10 8 This reduces the magnitude of invalid iterations.

[0145] Collaborative triggering execution: If AGA fails to update the optimal solution for 30 consecutive iterations (convergence stalls), MIDP is triggered to repartition the solution space. For example, the "Shanghai-Beijing route frequency optimization" is split into independent subproblems, solved, and fed back to AGA to restart global optimization.

[0146] 5. Implementation of Constrained Fit Fitness Function Design the fitness function: F' = F - gradient penalty term, where the penalty rule is: Violation of carbon quota constraints: deduct 0.2F; Violation of crew flight time constraints: deduct 0.15F; Violation of airport capacity constraints: deduct 0.1F. For example, a solution with an original F=0.8, but violating carbon quota constraints, will have its F'=0.64 after penalty, and will be automatically classified into the medium solution interval for subsequent iterations.

[0147] The high-dimensional variable block encoding in the above embodiments simplifies the solution complexity, and the fitness hierarchical strategy achieves accurate probabilistic adaptation. Combined with the collaborative mechanism of MIDP and AGA and a constraint-adaptive fitness function, it effectively solves problems such as numerous invalid iterations, unstable convergence, and poor solution compliance in solving high-dimensional discrete decision variables. This technical solution enhances the algorithm's adaptability to multi-constraint scenarios in civil aviation operations, ensures efficient and stable solution processes, provides reliable algorithmic support for the dynamic optimization and configuration of core elements, and fully meets the accuracy and feasibility requirements of complex decision-making in civil aviation operations.

[0148] Furthermore, in one embodiment, the scenario simulation and iteration module supports scenario parameters including at least: carbon tax rate adjustment, changes in total carbon allowances, passenger demand fluctuations, and fuel efficiency improvement rate of new aircraft models. Among these, carbon tax rate adjustment refers to changes in the tax standards levied on carbon emissions at the national or regional level, a core policy variable affecting airlines' carbon costs, used to simulate operational decision adaptation under different policy intensities. Changes in total carbon allowances represent the adjustment ratio (increase, decrease, or maintenance) of the total regional carbon emission allowances obtained by airlines, directly related to emission reduction pressure, and can be used to simulate the impact of quota constraints on capacity allocation. Passenger demand fluctuations represent the positive / negative fluctuation ratio of the average daily passenger volume on a route compared to a benchmark value, reflecting changes in market demand, and used to simulate the impact of supply and demand balance on route frequency and aircraft model selection. The fuel efficiency improvement rate of new aircraft models represents the percentage reduction in fuel consumption of newly added or replaced aircraft models compared to existing models, directly related to carbon emission intensity and operating costs, and used to simulate the empowering effect of aircraft model updates on green operations.

[0149] Furthermore, in one embodiment scenario, to address the issues of incomplete scenario coverage, low simulation efficiency, and iterative lag, the following technical solutions can also be included, thereby forming an enhanced technical system of "parameter systematization - computational efficiency - iterative dynamization - adaptation closed-loop": Scene parameter layering and refinement technology: In addition to the basic parameters, a "derived parameter layer" is constructed to supplement secondary variables that are strongly related to the basic parameters, forming a two-level system of "basic core parameters + derived adaptation parameters" to improve the accuracy of scene simulation.

[0150] Multi-scenario parallel computing technology: Adopting a multi-process / multi-thread parallel processing architecture, massive scenario combinations are distributed to multiple computing units for synchronous execution, which greatly shortens the overall simulation time and improves the decision response speed.

[0151] Dynamic iterative triggering technology: Set parameter deviation thresholds and triggering rules, compare simulated parameters with actual operational data in real time, and automatically start scenario updates and re-simulation when the deviation exceeds the threshold to avoid the simulation from reality being out of sync.

[0152] Simulation-Optimization Reverse Adaptation Technology: Establish a linkage mechanism between scenario simulation results and core element optimization models, input simulation conclusions from different scenarios into the optimization model, dynamically adjust model parameters and constraints, and form a closed loop of "simulation-optimization-verification-iteration".

[0153] The following is an example of scenario simulation and iteration for an airline in East China in Q3 2025. The specific implementation steps are explained below: (I) Scene parameter system configuration. Specific parameter configurations can be as follows: Carbon tax rate adjustment: 3 levels (50 yuan / ton, 80 yuan / ton, 120 yuan / ton, with a benchmark of 80 yuan / ton).

[0154] Changes in total carbon allowances: 3 tiers (-10%, 0, +10%, with a base allowance of 5,000 tons).

[0155] Passenger demand fluctuation range: 4 levels (-15%, -5%, +5%, +15%, with a base daily average passenger flow of 23,000). New engine fuel efficiency improvement rates: 2 levels (15% and 25%, corresponding to A321neo and B787-9 models).

[0156] Further parameter configurations may include: Policy implementation cycle: There are three categories after the carbon tax adjustment: "immediate implementation", "3-month transition period" and "6-month transition period".

[0157] Regional differentiation coefficient: 1.2 for core cities in East China (Shanghai and Hangzhou), and 0.9 for surrounding cities.

[0158] Fuel price linkage coefficient: For every 10% fluctuation in fuel prices, passenger demand fluctuations are adjusted by ±2%.

[0159] Aircraft introduction schedule: New aircraft models will be introduced in two categories: "quarterly batch introduction" and "half-year centralized introduction".

[0160] (ii) Parallel simulation execution in multiple scenarios Scenario combination generation: By permuting all basic and derived parameters, 3×3×4×2×3×2×2=864 complete scenario combinations are generated. Parallel architecture deployment: 16 parallel processes are launched using the Python multiprocessing library, each process is assigned 54 scenarios, and intermediate results are synchronized between processes through queues. Simulation of core logic: Each scenario is substituted into the core element optimization model to calculate the three-dimensional target values ​​of "carbon emission reduction, operating cost, and capacity utilization rate," and outputs a scenario report including the optimal configuration scheme and risk warning thresholds.

[0161] (III) Dynamic Iteration Triggering and Execution Deviation threshold setting: The deviation threshold between the actual value and the simulated value of each parameter is uniformly set to ±8%.

[0162] Real-time data monitoring: Connecting with airline OMS systems and the Civil Aviation Administration's data platform, on September 10, it was monitored that: actual passenger demand increased by 12% compared to the benchmark value (exceeding the 8% threshold), and the actual implementation standard of Shanghai carbon tax was 100 yuan / ton (exceeding the original set level).

[0163] Iterative update operation: The module automatically adds scenarios such as "carbon tax of 100 yuan / ton" and "passenger demand +12%", recalculates the adaptation plan (such as increasing the number of A321neo aircraft on the Shanghai-Guangzhou route and increasing the frequency of Pudong-Shenzhen flights), and pushes it to the decision-making system.

[0164] (iv) Simulation-Optimization Reverse Adaptation By linking the simulation results of 864 scenarios with the core element optimization model, a "scenario parameter-optimal configuration" mapping library was established: High carbon tax (≥100 RMB / ton) + high fuel efficiency (25%) scenario: Suitable for the "prioritize the deployment of new aircraft models + reduce routes with high energy consumption aircraft models" solution. High passenger demand growth (≥+10%) + carbon quota reduction (-10%) scenario: Suitable for the "increase the frequency of trunk line flights + optimize the allocation of regional line capacity" solution. Long transition period (6 months) + phased introduction of new aircraft models scenario: Suitable for the "gradual retirement of old aircraft models + phased deployment of new aircraft models" solution.

[0165] The technical solutions described above utilize basic scenario parameters to construct a core simulation framework. Combined with the solution's layered parameters, parallel computing, dynamic iteration, and reverse adaptation technologies, the comprehensiveness, efficiency, and adaptability of scenario simulation are significantly improved. The solution achieves full coverage of uncertainties across multiple dimensions, including carbon policies, market demands, and aircraft updates. Parallel computing drastically reduces simulation time, dynamic iteration ensures real-time synchronization between simulation and actual operations, and the reverse adaptation mechanism makes optimization more targeted. This technology effectively supports airlines in responding to various operational scenario changes, enhancing the flexibility and resilience of green operation decisions, and providing solid scenario simulation and iterative support for the sustainable development of the civil aviation industry.

[0166] Furthermore, in one embodiment, the spatiotemporal visualization module further includes: using (Python) programming to integrate a plotting tool (SigmaPlot) to generate carbon emission comparison curves and cost-emission reduction benefit scatter plots for different aircraft types and routes within each planning period.

[0167] In specific implementation scenarios, a collaborative mode of "Python data processing + SigmaPlot visualization rendering" can be adopted. Python is responsible for cleaning, aggregating, and calculating multi-source data (such as total carbon emissions of each aircraft type / route, cost-emission reduction benefit value). Standardized data is transmitted to SigmaPlot through data interfaces (CSV / JSON file interaction or COM component calls), and SigmaPlot completes the professional-grade chart drawing.

[0168] The visualization output can include the following formats: Carbon emission comparison curves: Using "planning period (day / week / month)" as the X-axis and "carbon emissions (tons of CO2)" as the Y-axis, multi-curve comparison charts are generated by classifying by aircraft type (A320neo / B737MAX / ARJ21) or route (Shanghai-Beijing / Shanghai-Guangzhou, etc.) to intuitively show the time-series changes of carbon emissions in different dimensions.

[0169] Cost-Emission Reduction Benefit Scatter Plot: With "carbon emission reduction (tons of CO2)" as the X-axis and "operating cost (ten thousand yuan)" as the Y-axis, each scatter point represents the benefit value of a single aircraft type / route within the planning period, which can clearly show the relationship between cost and emission reduction.

[0170] The above embodiment uses Python to read multi-source fused data → data standardization processing (unifying units, filling in missing values) → aggregation calculation by planning cycle / aircraft type / route → output standardized data file → call SigmaPlotAPI to load data → configure chart style (axis, legend, color scheme) → generate visualization chart, thus completing the visualization output process.

[0171] Furthermore, the potential problems of "single visualization dimensions, weak interactivity, and insufficient real-time performance" can be addressed by constructing a full-featured visualization system that integrates "multi-dimensional overlay, interactive operation, real-time updates, templated configuration, and linked analysis," as detailed below: Multi-temporal and spatial dimension layer overlay: On the basis of the original curve / scatter plot, a 1km×1km spatiotemporal grid layer, an airport geographic information layer, and a flight trajectory layer are overlaid to achieve the integrated visualization of "carbon emission data + geospatial information" and intuitively locate high emission areas / routes.

[0172] Interactive visualization: Develop mouse interaction functions (zoom, box selection, drill-down), support clicking on scatter / curve nodes to view detailed data (such as carbon emission composition and cost details for a certain route and time period), and support quick filtering of visualization content by aircraft type / route / cycle.

[0173] Real-time data-driven updates: By connecting to the airline's real-time operations database and setting a minute-level data synchronization cycle, Python automatically triggers the SigmaPlot chart refresh when carbon emission / cost data is updated, enabling real-time dynamic presentation of visualization results.

[0174] Customizable visualization templates: Multiple visualization templates are preset for different user roles (management level / scheduling level / accounting level) and planning cycles (daily / quarterly / yearly) (such as management focusing on overall benefits and scheduling level focusing on carbon emissions of a single route). One-click template switching is supported to generate adapted charts.

[0175] Visualized results linked analysis: The visualization charts are linked with the core element optimization model. Clicking on the "high cost, low emission reduction" scatter points in the chart will automatically trigger the model to recalculate the optimization scheme for that aircraft type / route, realizing a closed loop of "visualized analysis - optimization decision".

[0176] The following is a case study of the implementation of the spatiotemporal visualization module for an airline in East China (12 core routes, 3 aircraft types) in Q3 2025. The specific process is described below: (a) Data preparation Basic data: Extract carbon emission data for each aircraft type / route in Q3 (average monthly carbon emission of A320neo per route is 432 tons), operating cost data (including fuel and carbon costs, average monthly cost of RMB 850,000 per route), and planning cycle data (divided by week, a total of 13 planning weeks).

[0177] Extended data: supplemented with geographic coordinate data of 286 1km×1km grids covering 12 routes, geographic information (latitude and longitude, capacity) of airports such as Shanghai, Hangzhou and Nanjing, and real-time flight trajectory data (updated every 5 minutes).

[0178] (II) Generation of core charts Python data preprocessing: Read multi-source data, unify the units (carbon emissions: tons of CO2, cost: 10,000 yuan), and use linear interpolation to complete the missing cost data for three routes. Aggregate the data by "week-aircraft type-route" dimension to generate a standardized CSV file (containing fields: planning week, aircraft type, route, carbon emissions, operating cost, and emission reduction).

[0179] Python integration with SigmaPlot: The CSV file mentioned above is loaded by calling the SigmaPlot COM interface through the pywin32 library. Chart parameters are configured: For the carbon emission comparison curve, the X-axis is set to "Planning Week (1-13)" and the Y-axis to "Carbon Emissions (tons)", with curve colors set according to aircraft type (blue for A320neo, red for B737MAX, and green for ARJ21); for the cost-emission reduction scatter plot, the X-axis is set to "Emission Reduction (tons)" and the Y-axis to "Operating Costs (RMB 10,000)", with the scatter plot size mapping to passenger traffic on the route. Basic charts are generated: the "Q3 Weekly Carbon Emission Comparison Curve for Each Aircraft Type" and the "Q3 Cost-Emission Reduction Benefit Scatter Plot for Each Route" are output and saved as PNG and PDF formats.

[0180] (III) Further Enhancement of Visualization Multi-dimensional layer overlay: Import the geographic base map of East China (SHP format) into SigmaPlot and overlay a 1km×1km grid layer (with grid IDs labeled). Link carbon emission data to the corresponding grids and label the grid carbon emission intensity with a color gradient (green-yellow-red) to generate a "Spatiotemporal Heat Map of Carbon Emissions from Air Routes in East China", which visually shows that the grids around Shanghai Hongqiao Airport have the highest carbon emission intensity.

[0181] Interactive feature development: A visual interactive interface was built based on the Python Tkinter framework, embedding charts generated by SigmaPlot.

[0182] Develop interactive features: use the mouse wheel to zoom in and out of the heatmap, and click on a scatter plot of the Shanghai-Beijing route to display a pop-up window showing the route's "Weekly Carbon Emission Details (45% during takeoff and landing, 55% during cruising)" and "Cost Composition (60% fuel cost, 30% carbon cost)".

[0183] Real-time configuration updates: A Python script is written to run every 30 minutes, connecting to the airline's real-time operations database to synchronize the latest carbon emission / cost data. SigmaPlot automatic refresh rules are configured: When the data update volume is ≥5%, the chart is automatically re-rendered. For example, if carbon emissions for the Shanghai-Shenzhen route increase by 8% is synchronized at 14:00 on September 15th, the chart will update the curve trend in real time.

[0184] Customizable template application: Three preset template types are available: Management layer template (showing an overview of the entire airline's quarterly costs and emissions reductions), Dispatch layer template (showing daily carbon emission curves for a single route), and Accounting layer template (showing scatter plots of carbon cost details for each aircraft type). Dispatchers can select the "Shanghai-Xiamen route + daily cycle" template to generate a daily carbon emission comparison curve for that route in September with a single click.

[0185] Linked analysis execution: Clicking on the Shanghai-Chongqing route scatter plot marked "High Cost, Low Emission Reduction" in the scatter plot will automatically invoke the core element optimization model and recalculate the aircraft configuration scheme for this route (replacing the B737MAX with the A320neo). After the model outputs the optimized scheme, the visualization module automatically generates a "Carbon Emissions-Cost Comparison Curve Before and After Optimization," intuitively displaying the optimization benefits (12% emission reduction, 7% cost reduction).

[0186] The embodiments provided by this invention provide a basic visualization of carbon emission comparison curves and cost-emission reduction scatter plots. Combined with extended multi-dimensional layer overlay, interactive operation, and real-time updates, the richness of spatiotemporal visualization dimensions and the ease of interaction are significantly enhanced. Multi-layer overlay achieves the fusion of carbon emission data and geospatial data; interactive functions support refined data drill-down; real-time updates ensure that visualization results are synchronized with actual operations; and linked analysis establishes a closed loop of "visual analysis - optimization decision-making." This solution allows airline operators to intuitively grasp the carbon emission and cost-effectiveness characteristics of different aircraft types and routes, improving the intuitiveness and accuracy of green operation decisions and providing efficient visual support for civil aviation carbon emission management.

[0187] Furthermore, in one embodiment, the decision variables of the core element dynamic optimization configuration model include: the frequency of flights operated by aircraft type a on route i to j in year t. The fleet size of aircraft type a in year t The number of new models a in year t and number of retirements By quantifying the configuration parameters along the "aircraft type-route-time" dimension, a clear decision object and boundary are provided for model solving. The technical definition, value logic, and constraint relationships are as follows: Specifically, among the decision variables, flight frequency refers to the flight frequency of aircraft type a operating on route i to j in year t (denoted as...). The unit is "shift / day", and the value is a non-negative integer. This indicates that aircraft type A does not operate on this route, reflecting the matching of aircraft type and route, and operational intensity. The fleet size is the fleet size of aircraft type A in year t (denoted as...). The unit is "frame", and the value is a positive integer, equal to the first frame. The difference between the annual fleet size and the number of new additions and retirements in that year is the fundamental constraint on capacity allocation. The decision variable "number of new additions" is the number of new additions of aircraft type a in year t (denoted as ). The unit is "frames," and the value is a non-negative integer, subject to constraints such as the procurement cycle of new aircraft models and budget limitations. The decision variable "retirement quantity" is: the number of aircraft model a retired in year t (denoted as ). The unit is "frame", and the value is a non-negative integer, subject to the age of the existing model and maintenance costs (machines with an age of ≥15 years are given priority for retirement).

[0188] Regarding constraints, the fleet size balance constraint is as follows: Flight frequency and capacity matching constraints: ( This refers to the number of seats per shift for aircraft type A. (Daily average utilization rate of the machine model); Non-negative constraints: , , , .

[0189] The above-mentioned decision-making loop of "fleet size - aircraft type scheduling - route operation" provides the model with quantifiable and optimizable core operational objects, which are directly related to the three major goals of carbon emission reduction, cost reduction and capacity utilization.

[0190] In one implementation scenario, a comprehensive decision variable system encompassing "segmented dimensions, dynamic linkage, refined constraints, and sensitivity adaptation" can be constructed to address the issues of "single variable dimensions, weak dynamic adaptability, and failure to consider refined operational needs." The specific details are as follows: Flight frequency will be broken down by time of day: Expanded to time-segmented frequency ( 's' represents the time period: morning peak 6-9 am, off-peak 9-18 pm, evening peak 18-21 pm, and nighttime 21-6 am, adapting to passenger flow fluctuations and carbon emission differences at different times. Fleet size and capacity tiering: [The text abruptly ends here, likely due to an incomplete translation or source material.] Divided into multiple tiers based on the number of seats ( K represents the class: small (less than 150 seats), medium (150-200 seats), and large (more than 200 seats), precisely matching the passenger flow demand of the route.

[0191] In addition, decision variables may also include: the number of aircraft types allocated ( ): The number of aircraft types transferred from type a to type b in year t, to address the short-term capacity gap. (Route capacity deployment ratio) Aircraft type A on the route The capacity share is used to constrain the monopoly of a single aircraft type on route operations. Dynamic constraint linkage: Establishing a linkage relationship between variables and real-time operational data (such as... Adjustment in conjunction with real-time passenger flow deviations Linked to the delivery cycle of new models). Multi-cycle rolling optimization: Expand "year t" into a three-level cycle of "annual-quarter-month" ( =Annual, =Quarterly, =Monthly), variable nesting optimization by period (monthly) Adjustment → Quarterly Adjustment → Annual Adjustment). Variable sensitivity analysis: By increasing the variable sensitivity coefficient ( (v is the decision variable), which quantifies the impact of each unit change in the variable on the three-dimensional target, supporting the prioritization of decision-making.

[0192] The following is based on 2026 ( Taking the optimization of core elements of an airline in East China (current aircraft types: 20 A320neo, 15 B737MAX, 10 ARJ21) as an example, the implementation process is described in detail below: First, initialize the basic decision variables (based on 2025 operational data): fleet size ( ): A320neo = 20 aircraft, B737MAX = 15 aircraft, ARJ21 = 10 aircraft; Number of new / retired aircraft in 2026 ( ): Five new A320neo aircraft added ( ), 0 retired ( ). No new B737 MAX aircraft were added, and 3 were retired (aircraft age ≥ 15 years); 3 new ARJ21 aircraft were added, and 1 was retired. Fleet size in 2026 ( A320neo= Frame, B737MAX ARJ21 = Frame. Flight frequency configuration ( A320neo aircraft operated the Shanghai-Beijing route. Shanghai, Beijing): Flight / Day; Shanghai-Guangzhou: Flights / Days; B737 MAX operates flights between Shanghai and Shenzhen: Class / Day; Hangzhou-Guangzhou: Flights / Days; ARJ21 operates flights from Shanghai to Xiamen: Flight / Day; Nanjing-Ningbo: Flights / Day. Constraint Check: Fleet Size Balance: A320neo (Complies with constraints). Capacity matching: A320neo single shift $S_a=186$ seats, average daily utilization rate Available transport capacity per hour per year × ×365× =25×9.5×365×186≈1.58× Seats; Total flight frequency and annual capacity demand × =(8+7)×365×186≈1.01× Seat (satisfies constraints).

[0193] Furthermore, based on the above configuration, the time-segmented flight frequency configuration will be further refined and optimized. Shanghai-Beijing route (A320neo): Morning peak =Class 4, Pingfeng =Class 2, Evening Rush Hour =Class 2, Night =0 flights; Shanghai-Xiamen route (ARJ21): Morning peak =Class 2, Pingfeng =Class 2, Evening Rush Hour =Class 2, Night =0 classes; Constraint: The frequency of a single aircraft type operating the same route during the same time period. ≤3 shifts (to avoid scheduling conflicts). Fleet capacity tiered allocation ( A320neo: Mid-size (186 seats) 20 aircraft, large-scale storage (240 seats) 5 aircraft; B737MAX: Mid-range (178 seats) 12 aircraft; ARJ21: Small-sized (90 seats) 12 aircraft; Route matching: Shanghai-Beijing (high passenger flow) is allocated 2 A320neo large aircraft, with 1 flight each during the morning and evening peak hours.

[0194] Continue configuring other variables: number of aircraft types to be allocated ( ): Passenger traffic on the Shanghai-Shenzhen route is expected to increase in Q2 of 2026, with two aircraft being redeployed from B737MAX to A320neo. ); Route capacity ratio ( ): Capacity percentage of single aircraft type on the Shanghai-Guangzhou route ≤60%, the original configuration of 7 A320neo flights / day (accounting for 70%) has been adjusted to 6 A320neo flights and 4 B737MAX flights. =60%).

[0195] Multi-cycle rolling optimization: Monthly frequency adjustment: Passenger traffic on the Shanghai-Beijing route will decrease by 10% in January 2026, with nighttime routes adjusted accordingly. Frequency adjusted from 0 flights to 1 flight (off-peak diversion); Quarterly fleet adjustment: During the rainy season in East China in Q3, 3 ARJ21s were transferred from regional to mainline routes (short-haul, high-punctuality demand); Annual additions / retirements: Based on Q3 operational data, 2 additional A320neo aircraft were added. (From 5 to 7).

[0196] Variable sensitivity analysis: Calculation of sensitivity coefficients: A320neo flight frequency For every additional 1 flight / day ( Carbon emission reduction increased by 0.02 ( Operating costs increased by 0.01 ( ); Decision priority: Prioritize increasing the frequency of the A320neo Shanghai-Beijing route (optimal overall sensitivity), and postpone the new B737MAX plans.

[0197] Furthermore, by substituting other decision variables into the core element optimization model and solving it using the AGA-MIDP coupled algorithm, the optimal configuration scheme for 2026 is output: Time-based frequency: Peak hours on trunk routes. Frequency accounts for 60%, off-peak 30%, nighttime 10%; Fleet configuration: 27 A320neo aircraft ( ), 10 B737MAX aircraft ( ), ARJ2112 aircraft; derivative variables: 3 aircraft type adjustments per quarter, capacity share of a single route. ≤60%.

[0198] Through the above-described embodiments of the present invention, the basic decision variables ( , This solution constructs a core optimization framework, combining techniques such as time-segmentation, capacity tiering, and derived variables to significantly improve the granularity and dynamic adaptability of decision variables. The solution achieves configuration optimization from "coarse-grained annual" to "fine-grained time-segment," avoiding resource waste through variable linkage and constraint refinement, and clarifying decision priorities through sensitivity analysis, making the optimization scheme more aligned with actual operational needs. This technical solution effectively supports airlines in the scientific allocation of core decisions such as fleet size, aircraft type scheduling, and route frequency, balancing carbon emission reduction, cost reduction, and capacity utilization goals, and improving the accuracy and operability of core civil aviation resource allocation.

[0199] Furthermore, in one embodiment, the system further includes a carbon quota adaptation module, used to acquire regional carbon quota update data in real time, automatically adjust the carbon quota constraint parameters in the model, and trigger the model to be solved again. Specifically, it may include: Data acquisition submodule: Real-time acquisition of updated regional carbon quota data (including annual total adjustments, quarterly quota allocation changes, etc.) through the API interface of the Civil Aviation Administration's carbon quota management platform, with a data synchronization cycle of ≤1 hour.

[0200] Parameter Adjustment Submodule: Maps the acquired carbon quota update data to the carbon quota constraint parameters in the core element optimization model. This will automatically overwrite the original parameter values ​​of the model.

[0201] Solution trigger submodule: When the carbon quota parameter adjustment range is ≥5%, the model solution module (i.e., the AGA-MIDP coupled algorithm mentioned above) is automatically called to start the re-solution process.

[0202] The aforementioned carbon quota adaptation module uses "real-time data synchronization → automatic parameter mapping → on-demand model re-solution" as its core link to solve the problem of "statically fixed carbon quota constraints and inability to adapt to dynamic policy adjustments" in traditional models, ensuring that the optimization scheme always meets the requirements of the latest carbon policy.

[0203] Furthermore, the carbon quota adaptation module can further address the problems of the original technical solution, namely "single data source, crude adjustment strategy, and insufficient linkage," by constructing an enhanced carbon quota adaptation system that integrates "multi-source data fusion, hierarchical adjustment strategy, cross-module linkage, and quota early warning verification." Specifically, this includes: Multi-source carbon quota data integration: In addition to regional carbon quotas, three new data sources have been added: industry carbon quota benchmarks, airlines' own remaining quotas, and carbon trading market quota prices, forming a multi-dimensional data input of "policy + market + self".

[0204] The tiered adjustment strategy categorizes adjustments based on the magnitude of carbon quota updates (small: 5%-10%, medium: 10%-20%, large: ≥20%), corresponding to different model re-solution strategies. Specifically, for small adjustments: only carbon quota constraint parameters are updated, and the model is called for rapid solution (reducing the number of iterations by 50%); for medium adjustments: the decision variable range of claim 8 is simultaneously linked (e.g., reducing the number of new high-energy-consuming models), and standard solution is initiated; for large adjustments: the scenario simulation module of claim 6 is linked to generate a quota adjustment plan, and then a full solution is initiated.

[0205] Cross-module linkage and adaptation: It links with the spatiotemporal visualization module in the above embodiment to generate a comparison curve of "quota adjustment - carbon emission change" in real time; it links with the decision variable module in the above embodiment to automatically lock the adjustment range of high carbon emission decision variables.

[0206] Quota consumption early warning and verification: The quota consumption prediction submodule can predict the quota consumption trend for the next 30 days based on current operating data, and trigger an early warning when the quota is predicted to be exceeded; the solution result verification submodule can ensure that the carbon emissions of the re-solved solution do not exceed the latest quota constraints.

[0207] The following is a detailed explanation of the implementation process using an airline in East China in Q2 of 2026 (original regional carbon allowance = 1300 tons / quarter): First, real-time carbon quota adaptation data was obtained: On May 10, 2026, the Civil Aviation Administration of China's Q2 carbon quota adjustment notice for East China was obtained through the API interface: the total amount was reduced from 1300 tons to 1200 tons (an adjustment of 7.7%). Automatic parameter adjustment: The module updated the carbon quota constraint parameter \(E^{\text{quota}}(2026Q2)\) in the model from 1300 tons to 1200 tons. Triggering model re-solution: Because the adjustment magnitude was ≥5%, the AGA-MIDP coupled algorithm was automatically invoked to start the solution, outputting the adjustment plan: reducing the daily frequency of the Shanghai-Shenzhen route using B737MAX aircraft by one flight, reducing the total carbon emissions to 1180 tons.

[0208] Furthermore, through enhanced carbon quota adaptation and multi-source data fusion: Industry carbon quota benchmark was obtained: the average quota for the East China civil aviation industry in Q2 was 1250 tons per airline; airline's remaining quota was obtained: as of May 10, airlines had consumed 650 tons of quota in Q2, with 550 tons remaining; carbon trading price was obtained: the Shanghai carbon market quota price on that day was 120 yuan / ton (a 15% increase compared to the previous month). A tiered adjustment strategy was implemented: adjustment level determination: quota reduction of 7.7% (moderate adjustment). Linked decision variables: the number of new B737 MAX aircraft was simultaneously locked (adjusted from 0 to -2, i.e., 2 aircraft retired early). Standard solution was initiated: the model combined with the adjusted variable range to solve, outputting a solution: 2 B737 MAX aircraft retired early + 1 reduction in daily frequency on the Shanghai-Shenzhen route, resulting in a total carbon emission of 1150 tons. Cross-module integration: Integration with the spatiotemporal visualization module: Generates a carbon emission comparison curve from "before quota adjustment (1300 tons) - after adjustment (1200 tons)," visually demonstrating the contribution of the B737MAX adjustment to emission reduction. Integration with the scenario simulation module: Simulates a contingency plan of "further 10% reduction in quota," outputting an alternative plan of "accelerated introduction of new aircraft models." Quota warning and verification: Consumption prediction: Based on the current plan, predicts that the remaining quota in Q2 can support the economy until June 25th, with no risk of exceeding the quota; Result verification: Verifies that the new plan's carbon emissions of 1150 tons ≤ 1200 tons comply with the latest quota constraints.

[0209] The technical solutions described above enable real-time synchronization of carbon quota data and automatic model adaptation. Combined with the extended solution's multi-source data fusion, hierarchical adjustment, cross-module linkage, and early warning verification, the accuracy, flexibility, and foresight of carbon quota adaptation are significantly improved. Multi-source data allows for more comprehensive adjustment decisions, the hierarchical strategy balances solution efficiency and solution quality, cross-module linkage enhances the synergy of adjustments, and early warning verification avoids the risk of quota violations. This technology ensures that airlines consistently comply with carbon policy constraints while minimizing the impact on operational efficiency through refined adjustments, effectively improving the efficiency of carbon quota utilization and the dynamic adaptability of core element optimization solutions.

[0210] In summary, the technical solutions provided by the present invention, as described above, complete the construction of a full-chain support system for green civil aviation operations, encompassing "data fusion, cost accounting, optimized configuration, scenario simulation, visualization adaptation, and dynamic quota adjustment." Through precise spatiotemporal alignment of multi-source data and refined dynamic carbon cost accounting, a solid foundation of decision-making data is laid. The core element optimization model, combined with multi-objective weighted modeling, improved intelligent algorithms, and multi-dimensional flexible constraints, achieves the scientific allocation of fleet, routes, and capacity. Multi-scenario parallel simulation and minute-level iteration mechanisms adapt to dynamic changes in policies, markets, and aircraft types. Spatiotemporal visualization and cross-module linkage enhance the intuitiveness and synergy of decision-making. The multi-source data fusion and hierarchical adjustment of the carbon quota adaptation module ensure policy compliance and operational flexibility. The overall solution effectively balances carbon emission reduction targets and operational benefits, significantly improving the accuracy of core resource allocation, decision-making response efficiency, and risk resistance capabilities, providing comprehensive technical support for the green transformation and efficient operation of the civil aviation industry.

[0211] Figure 2 An exemplary structural block diagram of an electronic device 200 according to an embodiment of the present invention is shown.

[0212] like Figure 2 As shown, the electronic device 200 specifically includes a memory 201, a processor 202, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can implement the method described in any of the embodiments of method 100 described above. It is understood that the electronic device is a specific implementation of the aforementioned method 100, therefore the foregoing text is combined with... Figure 1 The characteristics described can be similarly applied here, and will not be repeated here.

[0213] While numerous embodiments of the invention have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of the invention. The appended claims are intended to define the scope of protection of the invention and therefore cover equivalents or alternatives within the scope of these claims.

Claims

1. A green transition decision support system based on the carbon emission characteristics of civil aviation, characterized in that, include: The multi-source data fusion module is used to integrate core element data of China's civil aviation, geographic data and carbon emission data. It achieves data alignment through spatial interpolation and temporal smoothing, and outputs the data in the form of a distance matrix. The core element dynamic optimization configuration model module minimizes the net present value of the total system cost during the planning period by embedding a dynamic optimization configuration model of civil aviation core elements from the perspective of carbon cost. The dynamic optimization configuration model of civil aviation core elements adopts a method of coupling graph theory network modeling and mixed integer dynamic programming. The total cost includes investment cost, operating cost and dynamic carbon cost. The model solving module uses mathematical modeling system tools to solve the dynamic optimization configuration model of the core elements, and outputs the fleet configuration, route flight frequency and carbon emission results for each planning period; The spatiotemporal visualization module uses geographic information system tools to visualize the spatial distribution of civil aviation elements and the spatiotemporal characteristics of carbon emissions. The scenario simulation and iteration module automatically updates the model and provides feedback on optimization results by supporting parameter adjustments for different carbon policies and demand fluctuation scenarios.

2. The green transformation decision support system according to claim 1, characterized in that, The alignment process of the multi-source data fusion module includes: matching the multi-source data according to a spatial grid of a set size and a timestamp of a set time level to achieve consistency in the spatial-temporal dimension, so that the data coverage is greater than a preset threshold.

3. The green transformation decision support system according to claim 1, characterized in that, The core element dynamic optimization configuration model includes dynamic carbon cost, which is calculated by multiplying total carbon emissions by dynamic carbon price. The dynamic carbon price is determined by coupling carbon market volatility coefficient and regional emission reduction policy adjustment coefficient.

4. The green transformation decision support system according to claim 1, characterized in that, The core element dynamic optimization configuration model includes multi-dimensional constraints, which include at least: passenger flow balance constraints, route capacity supply and demand constraints, fleet monthly utilization rate constraints, crew available flight capability constraints, and regional carbon quota constraints.

5. The green transformation decision support system according to claim 1, characterized in that, The model solving module integrates an adaptive genetic algorithm, which is coupled with mixed-integer dynamic programming to improve the efficiency of solving high-dimensional discrete decision variables. The adaptive genetic algorithm includes crossover probability and / or mutation probability, which are dynamically adjusted according to the fitness value of the model solution based on the core elements.

6. The green transformation decision support system according to claim 1, characterized in that, The scenario simulation and iteration module supports scenario parameters including at least: carbon tax rate adjustment, changes in total carbon quota, fluctuations in passenger demand, and fuel efficiency improvement rate of new aircraft models.

7. The green transformation decision support system according to claim 1, characterized in that, The spatiotemporal visualization module also includes: integrating plotting tools through programming to generate carbon emission comparison curves and cost-emission reduction benefit scatter plots for different aircraft types and routes within each planning period.

8. The green transformation decision support system according to claim 1, characterized in that, The decision variables of the dynamic optimization configuration model of the core elements include: the frequency of flights operated by aircraft type a on route i to j in year t. The fleet size of aircraft type a in year t The number of new models a in year t and number of retirements .

9. The green transformation decision support system according to claim 1, characterized in that, The system also includes a carbon quota adaptation module, which is used to acquire regional carbon quota update data in real time, automatically adjust the carbon quota constraint parameters in the model, and trigger the model to be solved again.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the functions of the green transition decision support system based on the carbon emission characteristics of civil aviation as described in any one of claims 1-9.

Images