Short-time airport departure passenger volume prediction method based on flight and security check integrated information

Abstract

The invention discloses a short-time airport departure passenger volume prediction method based on flight and security check integrated information, and belongs to the technical field of civil aviation transportation and intelligent transportation. Comprising the following steps: acquiring a flight information table and a passenger security check time table, and screening the flight information table and the passenger security check time table to obtain processed data; historical passenger advanced arrival time is collected, and fitting is carried out on the historical passenger advanced arrival time through a logarithmic normal distribution curve to obtain fitting parameters; predicting the advanced arrival time of the passenger based on the fitting parameters and the historical processing data; performing parameter adjustment on the fitting parameter to obtain an adjusted fitting parameter; obtaining a plurality of similar days of the to-be-predicted day based on the weather type and given features, and calculating an average value of adjustment fitting parameters of the similar days; a first parameter matrix is constructed based on the adjustment fitting parameters of each hour segment, and a second matrix is constructed based on the processing data; and generating a random number through the first parameter matrix and the second matrix, and obtaining a passenger arrival prediction amount counted by an hour set within one day based on the random number.

Description

Technical Field

[0001] The invention belongs to the field of civil aviation transportation and intelligent traffic technology, and in particular relates to a method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information. Background Art

[0002] With the vigorous development of my country's civil aviation industry, the demand for civil aviation passenger travel has continued to rise, and the number of passengers in airport terminals has also continued to grow. This demand growth has brought tremendous pressure on airports and airlines in terms of internal personnel and related resource allocation, as well as responding to emergencies. Many operational decisions, including personnel arrangements, resource allocation, and aircraft scheduling, rely on accurate short-term demand forecasts. Accurately predicting the flow of security check passengers in the terminal is crucial for resource planning, service quality improvement, and operational safety.

[0003] Existing studies mainly use the following methods to predict the number of departing passengers at airports: time series models, machine learning models, hybrid models, and distribution models. As a classic method in the field of prediction, the time series model relies on analyzing the trend, seasonal changes, and periodic fluctuations of historical data to predict future passenger flow, but its prediction accuracy significantly decreases when it faces interference from factors such as changes in flight information; machine learning models (such as support vector machines, random forests, etc.) can handle complex nonlinear relationships and can improve the accuracy of predictions to a certain extent, but they often require a large amount of training data, have high requirements for adjusting parameters, and the interpretability of the model is poor; hybrid models combine time series and machine learning models to improve the prediction effect to a certain extent, but the model complexity is high and it is difficult to apply to actual airport operations; there are also studies that use distribution models to model passenger flow, which can better describe the probability distribution characteristics of data, but if the data does not meet the assumptions, its prediction accuracy will be affected.

[0004] In summary, although the existing airport departure passenger volume forecasting methods have certain advantages, they also have obvious limitations and are difficult to cope with the complexity and real-time requirements of airport passenger volume forecasting. Summary of the invention

[0005] The present invention aims to solve the deficiencies of the prior art and proposes a short-term airport departure passenger volume prediction method based on flight and security check aggregate information. The airport departure passengers are modeled using the log-normal distribution and its characteristics, and their parameters (μ and σ) are divided by hourly segments, thereby generating and analyzing the advance arrival time of flight passengers. The model parameters are adjusted in a two-stage parameter adjustment method. In the first stage, the parameter μ is adjusted, and in the second stage, the parameter σ is adjusted. Finally, the weather is divided into categories. When predicting the future day, it is first determined which weather category it belongs to, and then the two similar days closest to the predicted day are selected from the selected category. The parameters of the similar days are averaged and brought into the model to complete the prediction of the predicted day.

[0006] To achieve the above object, the present invention provides the following scheme: a method for predicting the number of passengers departing from an airport in a short time based on flight and security check aggregate information, comprising the following steps:

[0007] Step 1: Collect flight information tables and passenger security check schedules, and screen and process the flight information tables and passenger security check schedules to obtain processed data;

[0008] Step 2: Collect historical passengers’ early arrival time, and fit the historical passengers’ early arrival time through a log-normal distribution curve to obtain fitting parameters;

[0009] Step 3: Predicting the early arrival time of passengers based on the fitting parameters and historical processing data to obtain an initial prediction result;

[0010] Step 4: performing parameter adjustment on the fitting parameters to obtain adjusted fitting parameters;

[0011] Step 5: Based on the weather type and given characteristics, several similar days of the day to be predicted are obtained, and the average values ​​of the adjusted fitting parameters of the similar days are calculated;

[0012] Step 6: Construct a first parameter matrix based on adjusting the fitting parameters for each hour, and construct a second matrix based on the processed data; generate random numbers that satisfy a log-normal distribution through the first parameter matrix and the second matrix, and obtain the predicted number of passenger arrivals per hour in a day based on the random numbers.

[0013] Further preferably, in step 1, after the data is screened and processed, the flight number, terminal, execution date, planned departure time, and passenger security check time information are retained.

[0014] Further preferably, the fitting parameters include an initial value vector μ (0) and σ (0) ;

[0015] In step 4, the parameter adjustment of the fitting parameters includes: the first stage parameter adjustment is used to adjust the initial value vector μ (0) Make adjustments;

[0016] And the second stage parameter adjustment is used to adjust the initial value vector σ (0) Make adjustments.

[0017] Further preferably, the method for adjusting the parameters in the first stage includes:

[0018] Calculate the weighted mean absolute percentage error vector for each hour of the initial forecast And sort them, and sort the initial value vector μ in descending order. (0) The parameter μ in is adjusted to obtain the initial value vector σ (0) The adjusted parameter vector with the smallest prediction error between

[0019] The adjustment method includes: in the i-th time period, the parameter adjustment value of the j-th round is:

[0020]

[0021] In the formula, μ i represents the upper limit of the parameter μ in the i-th time period; Δμ represents the step size of the parameter μ.

[0022] Further preferably, the method for adjusting the parameters in the second stage includes:

[0023] Based on the adjusted parameter vector and the initial value vector σ (0) Get the second advance arrival time, and calculate the weighted average absolute percentage error vector of each hour of the second advance arrival time In order from large to small, the initial value vector σ (0) The parameter σ in is adjusted to obtain the adjusted parameter vector The adjusted parameter vector with the smallest prediction error between

[0024] The adjustment method includes: in the i-th time period, the parameter adjustment value of the j-th round is:

[0025] σ (i,j) = σ i +(j-1)*Δσ,

[0026] In the formula, σ i represents the upper limit of parameter σ in the i-th time period; Δσ represents the step size of parameter σ.

[0027] Further preferably, the method for calculating the weighted mean absolute percentage error of each hourly segment includes:

[0028]

[0029] Where predicted_value(i) is the predicted value of the i-th time period; actual_value(i) is the actual value of the i-th time period.

[0030] Further preferably, the given features in step 5 include: holidays and the number of days until the holidays;

[0031] Similar days are selected under the same weather category, and the distance between similar days is measured using Manhattan distance:

[0032] d(i,j)=|x′ i,1 -x′ j,1 |+|x′ i,2 -x′ j,2 |,

[0033] Where d(i,j) represents the Manhattan distance between the i-th day and the j-th day; x′ i,1 represents the value of the first feature after quantization on the i-th day; x′ j,1 represents the value of the first feature after quantization on the jth day; x′ i,2 represents the value of the second feature after quantization on the i-th day; x′ j,2 Represents the value of the second feature after quantization on the jth day.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention models the advance arrival time of passengers based on flight data and passenger security check arrival data, and can effectively predict the airport passenger throughput in hours through a two-stage parameter adjustment algorithm. Compared with traditional methods, the present invention avoids the prediction error caused by a single data source or a fixed model structure, and overcomes the problem of insufficient interpretability of existing models. The present invention not only simplifies the model calculation, but also can effectively cope with the impact of different date types and weather changes. The airport passenger short-term arrival forecast results output by the model can help airports optimize resource allocation and improve operational efficiency, thereby ensuring passenger service quality and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0037] Figure 1 A method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to an embodiment of the present invention;

[0038] Figure 2 Schematic diagram of adjusting the parameter μ in the first stage in the parameter adjustment method based on two-stage adjustment according to an embodiment of the present invention;

[0039] Figure 3 Schematic diagram of adjusting the parameter σ in the second stage in the parameter adjustment method based on two-stage adjustment according to an embodiment of the present invention;

[0040] Figure 4 A flow chart of selecting similar days based on features and obtaining parameters of a date to be predicted according to an embodiment of the present invention;

[0041] Figure 5 The present invention is a flowchart of a method for predicting the number of airport passenger arrivals based on lognormal distribution according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Embodiment 1:

[0045] like Figure 1 As shown, this embodiment provides a method for predicting the number of passengers departing from an airport in a short time based on flight and security check aggregate information, which specifically includes the following steps:

[0046] Step 1: Collect flight information tables and passenger security check schedules, and screen and process the flight information tables and passenger security check schedules to obtain processed data.

[0047] After the data is filtered, the flight number, terminal, execution date, scheduled departure time, passenger security check time and other information are retained. The filtering operation can be automated in Excel by executing a python program. The above method is used to process the collected real-time flight information table and passenger security check schedule as well as the historical flight information table and passenger security check schedule.

[0048] Step 2: Collect historical passengers’ early arrival times, and fit the historical passengers’ early arrival times through a log-normal distribution curve to obtain fitting parameters.

[0049] In this embodiment, the advance arrival time of each passenger is calculated by the scheduled departure time of each flight and the arrival time of the passenger, with each flight as the unit. Since the specific time when the passenger arrives at the terminal is difficult to obtain, this embodiment uses the time when the passenger starts the security check service as the advance arrival time of the passenger.

[0050] According to the historical passengers' early arrival time, the log-normal distribution curve is used to fit the passengers' early arrival time in the same day and the same period. After removing the outliers, the parameters in the same period are averaged to obtain the fitting parameters, including: the initial value vector μ (0) and σ (0) ; Output the upper and lower limits of two parameters at the same time:

[0051] Step 3: Predict the early arrival time of passengers based on the fitting parameters and historical processing data to obtain an initial prediction result.

[0052] Through the above fitting parameters and flight information (scheduled departure time, number of passengers on the flight), and taking advantage of the fact that the arrival time of passengers on each flight follows a log-normal distribution, the advance arrival time is generated for the passengers on each flight and aggregated in hours.

[0053] In order to avoid the interference of actual values ​​on the error, this embodiment uses the weighted mean absolute percentage error (WMAPE) to evaluate the accuracy of the passenger volume prediction method, and based on this, obtains the weighted mean absolute percentage error vector for each hour period. The calculation method of the weighted mean absolute percentage error includes:

[0054]

[0055] Where predicted_value(i) is the predicted value of the i-th time period; actual_value(i) is the actual value of the i-th time period.

[0056] Step 4: Adjust the fitting parameters to obtain adjusted fitting parameters. In this embodiment, adjusting the fitting parameters includes: first stage parameter adjustment, which is used to adjust the initial value vector μ (0) Adjustment; and the second stage parameter adjustment, used to adjust the initial value vector σ (0) Make adjustments.

[0057] Among them, the method of adjusting parameters in the first stage includes:

[0058] Calculate the weighted mean absolute percentage error vector for each hour of the initial forecast And sort them, and sort the initial value vector μ in descending order. (0) The parameter μ in is adjusted to obtain the initial value vector σ (0) The adjusted parameter vector with the smallest prediction error between

[0059] Specifically, Figure 2 As shown, the adjustment method includes: the update times of each parameter is: In the i-th time period, the parameter adjustment value of the j-th round is:

[0060]

[0061] In the formula, μ i represents the upper limit of the parameter μ in the i-th time period; Δμ represents the step size of the parameter μ.

[0062] The method for adjusting the parameters in the second stage includes:

[0063] Based on the adjusted parameter vector Initial value vector σ (0) The arrival time of passengers on each flight follows the log-normal distribution, and the early arrival time of passengers on each flight is obtained, that is, the second early arrival time, and the weighted average absolute percentage error vector of the hourly segment of the second early arrival time is calculated. In order from large to small, the initial value vector σ (0) The parameter σ in is adjusted to obtain the adjusted parameter vector The adjusted parameter vector with the smallest prediction error between

[0064] Specifically, Figure 3 As shown, the adjustment method includes: the update times of each parameter is: In the i-th time period, the parameter adjustment value of the j-th round is:

[0065] σ (i,j) = σ i +(j-1)*Δσ,

[0066] In the formula, σ i represents the upper limit of parameter σ in the i-th time period; Δσ represents the step size of parameter σ.

[0067] For each time period, when the time period is used as the time period to be adjusted, consider adjusting the parameters of the next time period to reduce the error of the current time period. When adjusting each parameter μ, it is necessary to calculate the current parameter vector μ (i,j) The error between the predicted number of passengers and the actual value is adjusted to minimize the prediction error. When adjusting each parameter σ, the current parameter vector needs to be calculated. With σ (i,j) The predicted error between the number of passengers predicted and the actual value, and the adjustment target is the parameter vector With parameter vector Minimize the prediction error for the current time period.

[0068] Step 5: Based on the weather type and given characteristics, several similar days of the day to be predicted are obtained, and the average values ​​of the adjusted fitting parameters of the similar days are calculated.

[0069] like Figure 4 As shown, in this embodiment, the given features include: holidays and the number of days from holidays; the similar days closest to the predicted day are matched by the above features, two closest similar days are selected, and the parameter average vector of the two closest similar days is calculated. and

[0070] Similar days are selected under the same weather category, and the distance between similar days is measured using Manhattan distance:

[0071] d(i,j)=|x′ i,1 -x′ j,1 |+|x′ i,2 -x′ j,2 |,

[0072] Where d(i,j) represents the Manhattan distance between the i-th day and the j-th day; x′ i,1 represents the value of the first feature after quantization on the i-th day; x′ j,1 represents the value of the first feature after quantization on the jth day; x′ i,2 represents the value of the second feature after quantization on the i-th day; x′ j,2 Represents the value of the second feature after quantization on the jth day.

[0073] Step 6: Construct a first parameter matrix based on adjusting the fitting parameters for each hour, and construct a second matrix based on the processed data; generate random numbers that satisfy a log-normal distribution through the first parameter matrix and the second matrix, and obtain the predicted number of passenger arrivals per hour in a day based on the random numbers.

[0074] Combination Figure 5 In the process of predicting the predicted date, it is necessary to combine the flight plan and the parameter mean vector and Taking advantage of the fact that the arrival time of passengers on each flight follows a normal distribution, we generate an advance arrival time for passengers on each flight and aggregate it in hours to obtain the predicted arrival amount of passengers in hours. and Construct the first parameter matrix At the same time, the flight scheduled departure time and the number of passengers on the flight form the second matrix F P,T For each flight on each day, use With F P,T (i,1) (represents the first number in the i-th row), generates a random number that satisfies the log-normal distribution as the early arrival time P of the passengers on this flight i (Random number P i Satisfy the range of 30-180), and then calculate the actual arrival time F of passengers on flight i P,T (i,2)-P i Finally, passengers are counted hourly to obtain the predicted number of passenger arrivals per hour in a day.

[0075] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information, characterized in that: The following steps are involved: Step 1: Collect flight information tables and passenger security check schedules, and screen and process the flight information tables and passenger security check schedules to obtain processed data; Step 2: Collect historical passengers’ early arrival time, and fit the historical passengers’ early arrival time through a log-normal distribution curve to obtain fitting parameters; Step 3: Predicting the early arrival time of passengers based on the fitting parameters and historical processing data to obtain an initial prediction result; Step 4: performing parameter adjustment on the fitting parameters to obtain adjusted fitting parameters; Step 5: Based on the weather type and given characteristics, several similar days of the day to be predicted are obtained, and the average values ​​of the adjusted fitting parameters of the similar days are calculated; Step 6: Construct a first parameter matrix based on adjusting the fitting parameters for each hour, and construct a second matrix based on the processed data; generate random numbers that satisfy a log-normal distribution through the first parameter matrix and the second matrix, and obtain the predicted number of passenger arrivals per hour in a day based on the random numbers.

2. According to claim 1, a method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information, characterized in that: In step 1, after the data is screened and processed, the flight number, terminal, execution date, planned departure time, and passenger security check time information are retained.

3. The method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to claim 1, characterized in that: The fitting parameters include an initial value vector μ (0) and σ (0) ; In step 4, the parameter adjustment of the fitting parameters includes: the first stage parameter adjustment is used to adjust the initial value vector μ (0) Make adjustments; And the second stage parameter adjustment is used to adjust the initial value vector σ (0) Make adjustments.

4. The method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to claim 3, characterized in that: The method for adjusting the parameters in the first stage includes: Calculate the weighted mean absolute percentage error vector for each hour of the initial forecast And sort them, and sort the initial value vector μ in descending order. (0) The parameter μ in is adjusted to obtain the initial value vector σ (0) The adjusted parameter vector with the smallest prediction error between The adjustment method includes: in the i-th time period, the parameter adjustment value of the j-th round is: m (i,j) = μ i +(j-1)*Dm, In the formula, μ i represents the upper limit of the parameter μ in the i-th time period; Δμ represents the step size of the parameter μ.

5. The method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to claim 4, characterized in that: The method for adjusting the parameters in the second stage includes: Based on the adjusted parameter vector and the initial value vector σ (0) Get the second advance arrival time, and calculate the weighted average absolute percentage error vector of each hour of the second advance arrival time In order from large to small, the initial value vector σ (0) The parameter σ in is adjusted to obtain the adjusted parameter vector The adjusted parameter vector with the smallest prediction error between The adjustment method includes: in the i-th time period, the parameter adjustment value of the j-th round is: s (i,j) = σ i +(j-1)*Δσ, In the formula, σ i represents the upper limit of parameter σ in the i-th time period; Δσ represents the step size of parameter σ.

6. The method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to claim 4, characterized in that: The calculation method for the weighted mean absolute percentage error for each hourly segment is as follows: Where predicted_value(i) is the predicted value of the i-th time period; actual_value(i) is the actual value of the i-th time period.

7. The method for predicting the number of passengers departing from an airport in a short period of time based on flight and security check aggregate information according to claim 1, characterized in that: The given features in step 5 include: holidays and the number of days until the holidays; Similar days are selected under the same weather category, and the distance between similar days is measured using Manhattan distance: d(i,j)=|x′ i,1 -x′ j,1 |+|x′ i,2 -x′ j,2 |, Where d(i,j) represents the Manhattan distance between the i-th day and the j-th day; x′ i,1 represents the value of the first feature after quantization on the i-th day; x′ j,1 represents the value of the first feature after quantization on the jth day; x′ i,2 represents the value of the second feature after quantization on the i-th day; x′ j,2 Represents the value of the second feature after quantization on the jth day.

Images