CN110210648A - Control zone strategy method for predicting based on grey shot and long term memory network - Google Patents
Control zone strategy method for predicting based on grey shot and long term memory network Download PDFInfo
- Publication number
- CN110210648A CN110210648A CN201910360715.0A CN201910360715A CN110210648A CN 110210648 A CN110210648 A CN 110210648A CN 201910360715 A CN201910360715 A CN 201910360715A CN 110210648 A CN110210648 A CN 110210648A
- Authority
- CN
- China
- Prior art keywords
- value
- air traffic
- year
- traffic flow
- flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000007787 long-term memory Effects 0.000 title claims abstract description 15
- 238000012216 screening Methods 0.000 claims abstract description 15
- 238000010219 correlation analysis Methods 0.000 claims abstract description 10
- 239000013598 vector Substances 0.000 claims description 58
- 238000012549 training Methods 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 21
- 230000015654 memory Effects 0.000 claims description 20
- 238000012360 testing method Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 230000006403 short-term memory Effects 0.000 claims description 11
- 230000001186 cumulative effect Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 9
- 238000012847 principal component analysis method Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000009825 accumulation Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000005316 response function Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008676 import Effects 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 238000013277 forecasting method Methods 0.000 claims 7
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000005457 optimization Methods 0.000 abstract description 2
- 230000004907 flux Effects 0.000 abstract 2
- 230000001105 regulatory effect Effects 0.000 abstract 2
- 238000007726 management method Methods 0.000 description 10
- 238000013439 planning Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G06Q50/40—
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of control zone strategy method for predicting based on grey shot and long term memory network, belong to air traffic flow management technical field.The present invention includes the following steps: step 1: reading data；Step 2: data prediction；Step 3: utilizing grey correlation analysis preliminary screening influence factor；Step 4: extracting main feature using principal component analytical method；Step 5: establishing gray strategy flux prediction model；Step 6: establishing shot and long term memory network strategy flux prediction model；Step 7: establishing grey shot and long term memory network combination forecasting.The airspace structures optimizations such as the present invention can delineate for the sector of regulatory area, air route adjusts provide scientific basis, it realizes the effective use of airspace resource, while providing foundation for resource requirements distribution such as regulatory area future personnel investment, finance investment and fixed assets investments.
Description
Technical Field
The invention relates to a grey long-short term memory network-based strategic flow prediction method for controlled airspace, and belongs to the technical field of air traffic flow management.
Background
In recent years, China's civil aviation industry is rapidly developed, and the demand of air transportation is gradually vigorous. At present, the air traffic flow management service in China is always in a state of lag and low efficiency, and the contradiction between the air traffic flow management service and the air traffic flow which is increased at a high speed is increasingly prominent. Unreasonable planning of airspace, low degree of systemization and automation of flow management, low guarantee capability of an airspace management system and randomness of flow control lead the existing airspace resources and management means to be difficult to adapt to rapid increase of air traffic flow, and the situations of traffic congestion, flight conflict and the like occur in airports, terminal areas, airway intersections and the like, so that a bottleneck of an air traffic network is formed, and the situations of ground waiting before flight, air waiting, diversion, yawing in flight and the like are directly caused, thereby affecting flight safety, increasing flight fuel consumption and reducing normal flight. Therefore, a scientific and effective air traffic flow management system needs to be established to scientifically manage the air traffic flow, and the premise and basis for flow management is to accurately count and predict the distribution and development trend of the air traffic flow.
The prediction of the number of aircrafts in a certain airspace and a certain time period is part of the prediction of the air traffic flow, is an important basis and a decision basis for air traffic planning and management, and provides a basis for improving the operation efficiency of the whole country and the region. According to the statistics of the national and regional distribution of air traffic flow in the early stage, areas and waypoints which are possibly crowded in a period of time in the future are predicted, control personnel are enabled to carry out corresponding strategic deployment, a deployment scheme is made in advance, and effective control measures are taken in the later flight operation stage, so that the problem of flight delay caused by air traffic congestion is greatly relieved, flight safety is guaranteed, the reasonable utilization rate of airspace resources is improved, and the operation efficiency and the economic benefit of an airline company are greatly improved. In addition, the flow prediction can also improve the rationality of flight scheduling, eliminate a series of unreasonable and unsafe factors caused by traffic jam and promote the reasonable distribution and effective monitoring of flight times. The air traffic flow prediction also has guiding effect on the construction and planning of facilities such as airports and the like, so that resources such as airspace, air routes and the like are effectively and fully utilized, the future China aviation resource allocation is promoted to better adapt to the continuously increased traffic transportation demand, powerful support is provided for the effective smooth operation of air traffic, and the sustainable development of China civil aviation transportation industry is facilitated.
Air traffic flow forecasts can be divided into strategic forecasts and tactical forecasts. The strategic prediction focuses on the analysis of the medium-and-long-term flow space-time variation trend, and has an important guiding function for adjusting the national airspace structure, relieving airspace congestion, controlling personnel and equipment allocation of units and the like. The tactical prediction considers the conditions of airspace structures, control rules, airway configuration and the like, and counts and predicts the number of aircrafts in a period of time in the future for airspace units such as airway points, airway sections, sectors and the like according to the flight plan of the day and real-time updated airspace data, telegraph, radar, weather and the like, so that a controller can respond to the air traffic conditions at ease, and simultaneously provides data support for an air traffic management decision maker.
Currently, most air traffic flow prediction studies are directed to tactical short-term flow prediction.
Disclosure of Invention
In order to accurately predict the air traffic flow, the invention provides a strategic flow prediction method of a controlled airspace based on a grey long-short term memory network, which can provide scientific basis for the optimization of airspace structures such as sector planning, airway adjustment and the like of a controlled area, realize the effective utilization of airspace resources and provide basis for the allocation of resource requirements such as future personnel investment, financial investment, fixed asset investment and the like of the controlled area.
The invention adopts the following technical scheme for solving the technical problems:
an air traffic flow strategy prediction method based on a grey long-short term memory neural network comprises the following steps:
step 1: reading data
Reading an air traffic flow data set which comprises a tower, approaching and regional annual flying frames, and then reading a factor data set which influences strategic flow in corresponding years, wherein the factor data set comprises various indexes of national and regional economy, population, consumption level, various traffic mode transportation volumes, transportation employment personnel and fixed asset investment, tourism, import and export amount, airline number and airplane frame number;
step 2: data pre-processing
Merging the air traffic flow data set and the influence factor data set according to the year, and then processing missing values and abnormal values of the flow data set and the influence factor data set;
and step 3: primarily screening influence factors by utilizing grey correlation analysis;
and 4, step 4: extracting main characteristics by using a principal component analysis method;
and 5: establishing a grey strategic flow prediction model;
step 6: establishing a strategic flow prediction model of the long-term and short-term memory network;
and 7: and establishing a grey long-short term memory network combined prediction model.
In the step 2, for missing values in the data set, filling the vacant data records by using an averaging method; and replacing the abnormal value in the data by utilizing an interpolation method.
The specific process of step 3 is as follows:
step 3.1: firstly, taking a flow data set as a reference vector and all influence factor data sets as comparison vectors, and carrying out non-dimensionalization processing on the reference vector and the comparison vectors by adopting initialization transformation;
step 3.2, calculating a gray correlation coefficient vector ξ between the reference vector and any comparison vector, wherein the calculation formula is as follows:
wherein, Δ is the absolute value of the difference between all the comparison vectors and the reference vector, M is the minimum value in Δ, M is the maximum value in Δ, and rho belongs to [0,1] as the resolution coefficient;
step 3.3: calculating the average value of elements in the grey correlation coefficient vector to obtain the grey correlation degree between the reference vector and all the comparison vectors; and finally, screening the influence factors according to the relevance degree to obtain the preliminarily screened influence factors.
The specific process of step 4 is as follows:
step 4.1, carrying out zero-mean standardization on the influence factors obtained by the primary screening;
step 4.2: calculating a correlation coefficient matrix of the screened influence factors;
step 4.3: calculating the eigenvalue and the eigenvector of the correlation coefficient matrix, wherein the number of the eigenvalue and the eigenvector is the same as the number of the influence factors obtained by screening;
step 4.4: calculating the principal component contribution rate and the accumulated contribution rate according to the characteristic value: the contribution rate of each principal component is the ratio of the corresponding characteristic value to the sum of all the characteristic values, and the accumulated contribution rate is the sum of the accumulated principal component contribution rates.
The specific process of step 5 is as follows:
step 5.1: the air traffic flow sequence in the raw data sequence is:
wherein Is a raw data sequence of air traffic flow,is the air traffic flow value of the first year,is the air traffic flow value of the second year,the air traffic flow value of the nth year;
the main characteristic sequences of the influencing factors are as follows:
wherein ：is the original data sequence of the first main feature,is firstThe value of the first year's key feature of the individual key feature,the value of the second year key feature of the first key feature,the value of the first key feature in the nth year,is the original data sequence of the (N-1) th main feature,is the value of the leading feature of the first year of the (N-1) th leading feature,is the value of the second year key feature of the (N-1) th key feature,is the value of the N-th main characteristic of the N-1 st main characteristic;
step 5.2: performing primary accumulation generation processing on the original data sequence to obtain a generation sequence as follows:
wherein ,for the generation of sequences of air traffic flow,accumulated values for the air traffic flow of the previous year,accumulating the numerical values for the air traffic flow of the previous two years,accumulating the numerical value for the air traffic flow of the previous n years,for the generation sequence of the first main feature,the numerical value is accumulated for the key feature of the year before the first key feature,the numerical values are accumulated for the key features of the first two years,the numerical value is accumulated for the key features n years before the first key feature,for the N-1 st main feature generation sequence,the accumulated value for the key feature of the year preceding the N-1 key feature,the accumulated value is the main feature of the first two years of the N-1 st main feature,accumulating numerical values for the main characteristics of the previous n years;
step 5.3: the series of fitted differential equations is generated using a first accumulation, namely:
wherein The generated sequence representing air traffic flow is derived over time t,generating a sequence representing the second main feature by obtaining the parameter sequences a, b by least squares2,b3,…,bN(ii) a Solving the differential equation to obtain a time response function, namely:
wherein ：represents the accumulated forecast of air traffic flow for the first year,cumulative predicted value representing air traffic flow in the second year, bjRepresenting the solved jth parameter in the differential equation,represents the cumulative value of the key features two years before the j-1 st key feature,represents the cumulative value of the key features n years before the j-1 st key feature,represents the air traffic of the nth yearAccumulating and predicting the flow;
step 5.4: the time response function is derived and reduced to obtain a prediction equation, and finally, a predicted value of the air traffic flow of the historical year is obtained
wherein ：a predicted value representing the air traffic flow in the first year,a predicted value representing the air traffic flow in the second year,represents the accumulated predicted value of the air traffic flow in the (n-1) th year,and (4) representing the predicted value of the air traffic flow in the nth year.
The specific process of step 6 is as follows:
step 6.1: initializing long and short term memory network parameters
Step 6.1.1: input layer and output layer arrangement
When an input layer and an output layer of the flow prediction neural network are constructed, the number of samples of each batch of training is initially set to be 1; time step, initial setting is 1; inputting a characteristic; outputting the characteristics; initialized input layer weights and biases; weights and biases of the initialized output layers;
step 6.1.2: long-short term memory network layer setup
When a long-term and short-term memory network layer is built, initializing an activation function and uniformly selecting a tanh function and a sigmoid function; the number of the initialized network layers is set to be 1, and the number of the network nodes is set to be 10; initializing a network layer weight and bias;
step 6.1.3: loss function setting
Selecting a square error loss function with higher convergence rate in a regression model in a flow prediction neural network;
step 6.2: setting network input layer characteristics and output characteristics
Step 6.2.1: constructing an input sample vector
Inputting the flight flow of a sample at a certain moment and the information of auxiliary variables obtained according to grey correlation analysis and a principal component analysis method;
step 6.2.2: constructing an output sample vector
The corresponding output sample vector is the information of the flight flow at a later moment;
step 6.2.3: data normalization
Performing zero-mean standardization on the input sample vector and the output sample vector to generate a dimensionless training data set;
step 6.3: model training and evaluation
Dividing samples according to 80% of samples as a training set and 20% of samples as a testing set, and then training a neural network according to different sample numbers, time steps, training cycle numbers, hidden layer numbers and node numbers of each batch for input sample vectors and output sample vectors of the training set: and finally, evaluating the model prediction effect according to the test set, and selecting the average absolute percentage error as an evaluation index.
The specific process of step 7 is as follows:
step 7.1: firstly, respectively predicting by adopting a gray prediction model and a neural network model to obtain a prediction result of the models on a test set, and evaluating the prediction effect of the prediction models by using average absolute percentage errors to obtain errors of the two models on the test set;
step 7.2: and (3) performing weighted combination on the two models, taking the weighted sum of the errors of the two models on the test set in the step 6.1 as an objective function, solving by taking the sum of the weights as 1 as a constraint condition to obtain an optimal weighting coefficient, wherein the prediction result of the combined model is the weighted sum of the predicted values of the two models.
The invention has the following beneficial effects:
1. the invention provides a method for combining gray correlation analysis and a principal component analysis method, which is characterized in that the gray correlation analysis method is firstly utilized to primarily screen influence factors of flow, and the principal component analysis method is utilized to extract main characteristics, so that auxiliary variables influencing strategic flow prediction are analyzed and obtained.
2. The invention applies the grey long-short term memory network combined prediction model, integrates the advantages of the long-short term memory network and the grey prediction model, establishes a new strategic flow prediction model, and realizes the high-precision prediction of the strategic flow of the control area.
3. The invention perfects the research of the civil aviation industry in the aspect of air traffic flow strategic prediction in China, promotes the development of air traffic intellectualization, and provides scientific basis for strategic planning of future control areas.
Drawings
Fig. 1 is a flow chart of a strategic traffic prediction method based on a grey long-short term memory neural network.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings.
A method for predicting strategic traffic of airspace control based on a grey long-short term memory network is shown in fig. 1, and includes the following steps:
step 1: reading data:
reading an air traffic flow data set which comprises a tower, approaching and regional annual flying frames, and then reading a factor data set which influences strategic flow in corresponding years, wherein the factor data set comprises various indexes of national and regional economy, population, consumption level, transportation volume of various traffic modes, employment personnel and fixed asset investment, tourism, import and export amount, airline number and airplane frame number.
Step 2: data preprocessing:
and merging the air traffic flow data set and the influence factor data set according to the year, and then processing the missing value and the abnormal value of the flow data set and the influence factor data set. And for missing values in the data set, filling the vacant data records by using an averaging method. And replacing the abnormal value in the data by utilizing an interpolation method.
And step 3: the method for preliminarily screening the influence factors by utilizing grey correlation analysis comprises the following specific steps:
step 3.1: firstly, a flow data set is used as a reference vector, all influence factor data sets are used as comparison vectors, and the reference vector and the comparison vectors are subjected to non-dimensionalization processing by adopting initialization transformation.
Step 3.2, calculating a gray correlation coefficient vector ξ between the reference vector and any comparison vector, wherein the calculation formula is as follows:
wherein, Δ is the absolute value of the difference between all comparison vectors and reference vectors, M is the minimum value of Δ, M is the maximum value of Δ, ρ ∈ [0,1] is the resolution coefficient, and the resolution coefficient of the patent is 0.5 according to the minimum information principle.
Step 3.3: and calculating the average value of the elements in the gray correlation coefficient vector to obtain the gray correlation degree between the reference vector and all the comparison vectors. Finally, screening the influence factors according to the relevance degree to obtain the preliminarily screened influence factors
And 4, step 4: the main characteristics are extracted by using a principal component analysis method, and the specific method comprises the following steps:
and (3) carrying out principal component analysis on the influence factors obtained by screening by the grey correlation analysis method in the step (2), and extracting main characteristics according to the analysis result, wherein the specific method comprises the following steps:
and 4.1, carrying out zero-mean standardization on the influence factors obtained by the primary screening.
Step 4.2: and calculating and screening the obtained correlation coefficient matrix of the influencing factors.
Step 4.3: and calculating the eigenvalue and the eigenvector of the correlation coefficient matrix. The number of the characteristic values and the number of the characteristic vectors are the same as the number of the influence factors obtained by screening.
Step 4.4: and calculating the principal component contribution rate and the accumulated contribution rate according to the characteristic value. The contribution rate of each principal component is the ratio of the corresponding eigenvalue to the sum of all eigenvalues. The accumulated contribution rate is the sum of the accumulated principal component contribution rates. The patent takes the principal component corresponding to the characteristic value with the cumulative contribution rate exceeding 95% as the main characteristic of the influence factor.
And 5: establishing a grey strategic flow prediction model, wherein the specific method comprises the following steps:
step 5.1: the air traffic flow sequence in the raw data sequence is:
wherein Is a raw data sequence of air traffic flow,is the air traffic flow value of the first year,is the air traffic flow value of the second year,is the air traffic flow value of the nth year.
The main characteristic sequences of the influencing factors are as follows:
wherein ：is the original data sequence of the first main feature,the value of the first key feature for the first year,the value of the second year key feature of the first key feature,the value of the first key feature in the nth year,is the original data sequence of the (N-1) th main feature,is the value of the leading feature of the first year of the (N-1) th leading feature,is the value of the second year key feature of the (N-1) th key feature,is the value of the N-1 th main characteristic of the year.
Step 5.2: performing primary accumulation generation processing on the original data sequence to obtain a generation sequence as follows:
wherein ,for the generation of sequences of air traffic flow,accumulated values for the air traffic flow of the previous year,accumulating the numerical values for the air traffic flow of the previous two years,accumulating the numerical value for the air traffic flow of the previous n years,for the generation sequence of the first main feature,the numerical value is accumulated for the key feature of the year before the first key feature,the numerical values are accumulated for the key features of the first two years,the numerical value is accumulated for the key features n years before the first key feature,for the N-1 st main feature generation sequence,the accumulated value for the key feature of the year preceding the N-1 key feature,the accumulated value is the main feature of the first two years of the N-1 st main feature,the numerical values are accumulated for the main characteristics of the previous n years.
Step 5.3: the series of fitted differential equations is generated using a first accumulation, namely:
wherein The generated sequence representing air traffic flow is derived over time t,generating a sequence representing the second main feature by obtaining the parameter sequences a, b by least squares2,b3,…,bN。
Solving the differential equation to obtain a time response function, namely:
wherein ：represents the accumulated forecast of air traffic flow for the first year,cumulative predicted value representing air traffic flow in the second year, bjRepresenting the solved jth parameter in the differential equation,represents the cumulative value of the key features two years before the j-1 st key feature,represents the cumulative value of the key features n years before the j-1 st key feature,and (4) representing the accumulated predicted value of the air traffic flow in the nth year.
Step 5.4: the time response function is derived and reduced to obtain a prediction equation, and finally, a predicted value of the air traffic flow of the historical year is obtained
wherein ：a predicted value representing the air traffic flow in the first year,a predicted value representing the air traffic flow in the second year,represents the accumulated predicted value of the air traffic flow in the (n-1) th year,and (4) representing the predicted value of the air traffic flow in the nth year.
Step 6: the method for establishing the strategic flow prediction model of the long-term and short-term memory network comprises the following specific steps:
step 6.1: initializing long and short term memory network parameters
Step 6.1.1: and an input layer and an output layer are arranged. When constructing the input and output layers of the traffic-predicting neural network, the settings of the following parameters are considered: the number of samples of each batch of training is initially set to 1; time step, initial setting is 1; inputting a characteristic; outputting the characteristics; initialized input layer weights and biases; the weights and biases of the initialized output layers.
Step 6.1.2: setting a long-term and short-term memory network layer. When a long-term and short-term memory network layer is built, initializing an activation function and uniformly selecting a tanh function and a sigmoid function; when the number of network layers and the number of nodes in each layer are considered, the number of initialized network layers is set to be 1 layer, and the number of network nodes is set to be 10; network layer weights and biases are initialized.
Step 6.1.3: and setting a loss function. The loss function is the main basis of model parameter correction, and the square error loss function with higher convergence rate in the regression model is selected in the flow prediction neural network.
Step 6.2: setting network input layer characteristics and output characteristics
Step 6.2.1: an input sample vector is constructed. The input sample is flight flow at a certain moment and the information of the auxiliary variable obtained according to grey correlation analysis and principal component analysis.
Step 6.2.2: an output sample vector is constructed. The corresponding output sample vector is information of the flight flow at a later time.
Step 6.2.3: and (6) standardizing data. And (4) performing zero-mean standardization on the input sample vector and the output sample vector to generate a dimensionless training data set.
Step 6.3: and (5) training and evaluating the model. The samples were divided according to 80% of the samples as training set and 20% of the samples as testing set. Then, training the neural network according to different sample numbers of each batch, time steps, training cycle numbers, hidden layer numbers and node numbers for the input sample vectors and the output sample vectors of the training set: the number of samples of each batch is selected from the {1, 2, 3, 4, 5, 6} set for setting, the time step is selected from the {1, 2, 3, 4} set for setting, the training period number is selected from the {50, 100, 150, 200, 250} set for setting, the number of hidden layers is selected from the {1, 2, 3, 4} set for setting, and the number of hidden layers is selected from the {10, 20, 30, 40} set for setting. By minimizing the loss function formed by the predicted output and the actual output, various weights and offset values are continuously updated, so that the model prediction effect is better. And finally, evaluating the model prediction effect according to the test set, and selecting the average absolute percentage error as an evaluation index.
And 7: establishing a combined prediction model of the grey long-short term memory network, which comprises the following specific steps:
the invention mainly adopts a parallel combination mode, processes through a grey model and a strategic flow prediction model of a long-term and short-term memory network respectively, and combines the processing results, and the specific method comprises the following steps:
step 7.1: firstly, a grey prediction model and a neural network model are adopted to respectively predict to obtain the prediction results of the models on a test set, and the average absolute percentage error is used to evaluate the prediction effect of the prediction models to obtain the errors of the two models on the test set.
Step 7.2: and (3) performing weighted combination on the two models, taking the weighted sum of the errors of the two models on the test set in the step 6.1 as an objective function, solving by taking the sum of the weights as 1 as a constraint condition to obtain an optimal weighting coefficient, so that the average absolute error and the mean square error predicted by the weighted combination model are minimized, and the optimal prediction effect is achieved. The prediction result of the combined model is the weighted sum of the prediction values of the two models.
Claims (7)
1. A strategic flow forecasting method for airspace control based on a grey long-short term memory network is characterized by comprising the following steps:
step 1: reading data
Reading an air traffic flow data set which comprises a tower, approaching and regional annual flying frames, and then reading a factor data set which influences strategic flow in corresponding years, wherein the factor data set comprises various indexes of national and regional economy, population, consumption level, various traffic mode transportation volumes, transportation employment personnel and fixed asset investment, tourism, import and export amount, airline number and airplane frame number;
step 2: data pre-processing
Merging the air traffic flow data set and the influence factor data set according to the year, and then processing missing values and abnormal values of the flow data set and the influence factor data set;
and step 3: primarily screening influence factors by utilizing grey correlation analysis;
and 4, step 4: extracting main characteristics by using a principal component analysis method;
and 5: establishing a grey strategic flow prediction model;
step 6: establishing a strategic flow prediction model of the long-term and short-term memory network;
and 7: and establishing a grey long-short term memory network combined prediction model.
2. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 1, wherein in step 2, for the missing values in the data set, a mean value method is adopted to fill up the missing data records; and replacing the abnormal value in the data by utilizing an interpolation method.
3. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 1, wherein the specific process of step 3 is as follows:
step 3.1: firstly, taking a flow data set as a reference vector and all influence factor data sets as comparison vectors, and carrying out non-dimensionalization processing on the reference vector and the comparison vectors by adopting initialization transformation;
step 3.2, calculating a gray correlation coefficient vector ξ between the reference vector and any comparison vector, wherein the calculation formula is as follows:
wherein, Δ is the absolute value of the difference between all the comparison vectors and the reference vector, M is the minimum value in Δ, M is the maximum value in Δ, and rho belongs to [0,1] as the resolution coefficient;
step 3.3: calculating the average value of elements in the grey correlation coefficient vector to obtain the grey correlation degree between the reference vector and all the comparison vectors; and finally, screening the influence factors according to the relevance degree to obtain the preliminarily screened influence factors.
4. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 1, wherein the specific process of step 4 is as follows:
step 4.1, carrying out zero-mean standardization on the influence factors obtained by the primary screening;
step 4.2: calculating a correlation coefficient matrix of the screened influence factors;
step 4.3: calculating the eigenvalue and the eigenvector of the correlation coefficient matrix, wherein the number of the eigenvalue and the eigenvector is the same as the number of the influence factors obtained by screening;
step 4.4: calculating the principal component contribution rate and the accumulated contribution rate according to the characteristic value: the contribution rate of each principal component is the ratio of the corresponding characteristic value to the sum of all the characteristic values, and the accumulated contribution rate is the sum of the accumulated principal component contribution rates.
5. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 1, wherein the specific process of step 5 is as follows:
step 5.1: the air traffic flow sequence in the raw data sequence is:
wherein For air traffic flowThe original sequence of data of the volume is,is the air traffic flow value of the first year,is the air traffic flow value of the second year,the air traffic flow value of the nth year;
the main characteristic sequences of the influencing factors are as follows:
wherein ：is the original data sequence of the first main feature,the value of the first key feature for the first year,the value of the second year key feature of the first key feature,the value of the first key feature in the nth year,is the original data sequence of the (N-1) th main feature,is the value of the leading feature of the first year of the (N-1) th leading feature,is the value of the second year key feature of the (N-1) th key feature,is the value of the N-th main characteristic of the N-1 st main characteristic;
step 5.2: performing primary accumulation generation processing on the original data sequence to obtain a generation sequence as follows:
wherein ,for the generation of sequences of air traffic flow,accumulated values for the air traffic flow of the previous year,accumulating the numerical values for the air traffic flow of the previous two years,accumulating the numerical value for the air traffic flow of the previous n years,for the generation sequence of the first main feature,is the first main characteristicThe accumulated value of the main characteristics of the previous year is proved,the numerical values are accumulated for the key features of the first two years,the numerical value is accumulated for the key features n years before the first key feature,for the N-1 st main feature generation sequence,the accumulated value for the key feature of the year preceding the N-1 key feature,the accumulated value is the main feature of the first two years of the N-1 st main feature,accumulating numerical values for the main characteristics of the previous n years;
step 5.3: the series of fitted differential equations is generated using a first accumulation, namely:
wherein The generated sequence representing air traffic flow is derived over time t,generating sequences representing the second main feature by least squaresBy finding the parameter sequences a, b2,b3,…,bN(ii) a Solving the differential equation to obtain a time response function, namely:
wherein ：represents the accumulated forecast of air traffic flow for the first year,cumulative predicted value representing air traffic flow in the second year, bjRepresenting the solved jth parameter in the differential equation,represents the cumulative value of the key features two years before the j-1 st key feature,represents the cumulative value of the key features n years before the j-1 st key feature,the accumulated predicted value of the air traffic flow in the nth year is represented;
step 5.4: the time response function is derived and reduced to obtain a prediction equation, and finally, a predicted value of the air traffic flow of the historical year is obtained
wherein ：represents the firstThe predicted value of the annual air traffic flow,a predicted value representing the air traffic flow in the second year,represents the accumulated predicted value of the air traffic flow in the (n-1) th year,and (4) representing the predicted value of the air traffic flow in the nth year.
6. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 1, wherein the specific process of step 6 is as follows:
step 6.1: initializing long and short term memory network parameters
Step 6.1.1: input layer and output layer arrangement
When an input layer and an output layer of the flow prediction neural network are constructed, the number of samples of each batch of training is initially set to be 1; time step, initial setting is 1; inputting a characteristic; outputting the characteristics; initialized input layer weights and biases; weights and biases of the initialized output layers;
step 6.1.2: long-short term memory network layer setup
When a long-term and short-term memory network layer is built, initializing an activation function and uniformly selecting a tanh function and a sigmoid function; the number of the initialized network layers is set to be 1, and the number of the network nodes is set to be 10; initializing a network layer weight and bias;
step 6.1.3: loss function setting
Selecting a square error loss function with higher convergence rate in a regression model in a flow prediction neural network;
step 6.2: setting network input layer characteristics and output characteristics
Step 6.2.1: constructing an input sample vector
Inputting the flight flow of a sample at a certain moment and the information of auxiliary variables obtained according to grey correlation analysis and a principal component analysis method;
step 6.2.2: constructing an output sample vector
The corresponding output sample vector is the information of the flight flow at a later moment;
step 6.2.3: data normalization
Performing zero-mean standardization on the input sample vector and the output sample vector to generate a dimensionless training data set;
step 6.3: model training and evaluation
Dividing samples according to 80% of samples as a training set and 20% of samples as a testing set, and then training a neural network according to different sample numbers, time steps, training cycle numbers, hidden layer numbers and node numbers of each batch for input sample vectors and output sample vectors of the training set: and finally, evaluating the model prediction effect according to the test set, and selecting the average absolute percentage error as an evaluation index.
7. The strategic flow forecasting method for airspace control based on the grey long-short term memory network as claimed in claim 6, wherein the specific process of step 7 is as follows:
step 7.1: firstly, respectively predicting by adopting a gray prediction model and a neural network model to obtain a prediction result of the models on a test set, and evaluating the prediction effect of the prediction models by using average absolute percentage errors to obtain errors of the two models on the test set;
step 7.2: and (3) performing weighted combination on the two models, taking the weighted sum of the errors of the two models on the test set in the step 6.1 as an objective function, solving by taking the sum of the weights as 1 as a constraint condition to obtain an optimal weighting coefficient, wherein the prediction result of the combined model is the weighted sum of the predicted values of the two models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910360715.0A CN110210648B (en) | 2019-04-30 | 2019-04-30 | Gray long-short term memory network-based control airspace strategic flow prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910360715.0A CN110210648B (en) | 2019-04-30 | 2019-04-30 | Gray long-short term memory network-based control airspace strategic flow prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110210648A true CN110210648A (en) | 2019-09-06 |
| CN110210648B CN110210648B (en) | 2023-05-23 |
Family
ID=67785406
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910360715.0A Active CN110210648B (en) | 2019-04-30 | 2019-04-30 | Gray long-short term memory network-based control airspace strategic flow prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110210648B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111191842A (en) * | 2019-12-30 | 2020-05-22 | 中国民用航空飞行学院 | Civil aviation transport airport fire station site selection optimization method |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111428932A (en) * | 2020-03-27 | 2020-07-17 | 中国民航大学 | Medium-and-long-term air traffic flow prediction method based on wavelet transformation and gray prediction |
| CN112419131A (en) * | 2020-11-20 | 2021-02-26 | 中南大学 | Method for estimating traffic origin-destination demand |
| CN112787882A (en) * | 2020-12-25 | 2021-05-11 | 国网河北省电力有限公司信息通信分公司 | Internet of things edge traffic prediction method, device and equipment |
| CN117434486A (en) * | 2023-12-20 | 2024-01-23 | 智联信通科技股份有限公司 | DC shunt metering error analysis processing method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106251002A (en) * | 2016-07-18 | 2016-12-21 | 华南理工大学 | Association analysis method for the meteorological big data of overhead transmission line load-bearing capacity assessment |
| CN107045785A (en) * | 2017-02-08 | 2017-08-15 | 河南理工大学 | A kind of method of the short-term traffic flow forecast based on grey ELM neutral nets |
| CN108694476A (en) * | 2018-06-29 | 2018-10-23 | 山东财经大学 | A kind of convolutional neural networks Stock Price Fluctuation prediction technique of combination financial and economic news |
| CN109215349A (en) * | 2018-10-26 | 2019-01-15 | 同济大学 | Traffic flow forecasting method when long based on deep learning |
| CN109508812A (en) * | 2018-10-09 | 2019-03-22 | 南京航空航天大学 | A kind of aircraft Trajectory Prediction method based on profound memory network |
| CN109615169A (en) * | 2018-11-08 | 2019-04-12 | 国家电网有限公司 | A kind of distribution network reliability evaluation method based on MEA-IElman neural network |
-
2019
- 2019-04-30 CN CN201910360715.0A patent/CN110210648B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106251002A (en) * | 2016-07-18 | 2016-12-21 | 华南理工大学 | Association analysis method for the meteorological big data of overhead transmission line load-bearing capacity assessment |
| CN107045785A (en) * | 2017-02-08 | 2017-08-15 | 河南理工大学 | A kind of method of the short-term traffic flow forecast based on grey ELM neutral nets |
| CN108694476A (en) * | 2018-06-29 | 2018-10-23 | 山东财经大学 | A kind of convolutional neural networks Stock Price Fluctuation prediction technique of combination financial and economic news |
| CN109508812A (en) * | 2018-10-09 | 2019-03-22 | 南京航空航天大学 | A kind of aircraft Trajectory Prediction method based on profound memory network |
| CN109215349A (en) * | 2018-10-26 | 2019-01-15 | 同济大学 | Traffic flow forecasting method when long based on deep learning |
| CN109615169A (en) * | 2018-11-08 | 2019-04-12 | 国家电网有限公司 | A kind of distribution network reliability evaluation method based on MEA-IElman neural network |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111191842A (en) * | 2019-12-30 | 2020-05-22 | 中国民用航空飞行学院 | Civil aviation transport airport fire station site selection optimization method |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111237134B (en) * | 2020-01-14 | 2022-04-01 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111428932A (en) * | 2020-03-27 | 2020-07-17 | 中国民航大学 | Medium-and-long-term air traffic flow prediction method based on wavelet transformation and gray prediction |
| CN111428932B (en) * | 2020-03-27 | 2022-12-06 | 中国民航大学 | Medium-and-long-term air traffic flow prediction method based on wavelet transformation and gray prediction |
| CN112419131A (en) * | 2020-11-20 | 2021-02-26 | 中南大学 | Method for estimating traffic origin-destination demand |
| CN112419131B (en) * | 2020-11-20 | 2022-07-08 | 中南大学 | Method for estimating traffic origin-destination demand |
| CN112787882A (en) * | 2020-12-25 | 2021-05-11 | 国网河北省电力有限公司信息通信分公司 | Internet of things edge traffic prediction method, device and equipment |
| CN117434486A (en) * | 2023-12-20 | 2024-01-23 | 智联信通科技股份有限公司 | DC shunt metering error analysis processing method |
| CN117434486B (en) * | 2023-12-20 | 2024-03-08 | 智联信通科技股份有限公司 | DC shunt metering error analysis processing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110210648B (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110210648B (en) | Gray long-short term memory network-based control airspace strategic flow prediction method | |
| Yang et al. | A network traffic forecasting method based on SA optimized ARIMA–BP neural network | |
| WO2020248228A1 (en) | Computing node load prediction method in a hadoop platform | |
| WO2021082393A1 (en) | Airport surface variable slide-out time prediction method based on big data deep learning | |
| CN106951976A (en) | A kind of bus passenger flow Forecasting Methodology based on pattern classification | |
| CN102200759A (en) | Nonlinear kernelled adaptive prediction method | |
| CN104992244A (en) | Airport freight traffic prediction analysis method based on SARIMA and RBF neural network integration combination model | |
| CN109858700A (en) | BP neural network heating system energy consumption prediction technique based on similar screening sample | |
| Wu et al. | An improved svm model for flight delay prediction | |
| CN114444660A (en) | Short-term power load prediction method based on attention mechanism and LSTM | |
| CN110097100A (en) | A kind of Bridge State Assessment method based on discrete dynamic Bayesian network | |
| CN112232561A (en) | Power load probability prediction method based on constrained parallel LSTM quantile regression | |
| CN111311905A (en) | Particle swarm optimization wavelet neural network-based expressway travel time prediction method | |
| CN110489844A (en) | One kind being suitable for the uneven large deformation grade prediction technique of soft rock tunnel | |
| CN108830405B (en) | Real-time power load prediction system and method based on multi-index dynamic matching | |
| CN116109212B (en) | Airport operation efficiency evaluation index design and monitoring method | |
| CN109887290B (en) | Traffic flow prediction method based on balance index smoothing method and stack type self-encoder | |
| CN115983487B (en) | Airport terminal passenger space-time distribution prediction method based on chi-square distribution | |
| CN115271186B (en) | Reservoir water level prediction and early warning method based on delay factor and PSO RNN Attention model | |
| CN114463978B (en) | Data monitoring method based on track traffic information processing terminal | |
| CN116384807A (en) | Method for evaluating congestion situation of airway network based on fuzzy comprehensive evaluation model | |
| Yang et al. | Forecasting model for urban traffic flow with BP neural network based on genetic algorithm | |
| Shao et al. | Air transportation delay propagation analysis with uncertainty in coloured–timed Petri nets | |
| CN106652443A (en) | Longitudinal and transverse similar highway short-time traffic prediction method | |
| Ji et al. | Train delays prediction based on feature selection and random forest |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |