CN115269758A - Passenger-guidance-oriented road network passenger flow state deduction method and system - Google Patents
Passenger-guidance-oriented road network passenger flow state deduction method and system Download PDFInfo
- Publication number
- CN115269758A CN115269758A CN202210922005.4A CN202210922005A CN115269758A CN 115269758 A CN115269758 A CN 115269758A CN 202210922005 A CN202210922005 A CN 202210922005A CN 115269758 A CN115269758 A CN 115269758A
- Authority
- CN
- China
- Prior art keywords
- passenger flow
- time
- passenger
- data
- station
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000001556 precipitation Methods 0.000 claims description 4
- 238000007621 cluster analysis Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 2
- 235000019580 granularity Nutrition 0.000 description 9
- 230000006870 function Effects 0.000 description 6
- 238000003064 k means clustering Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000007423 decrease Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000005206 flow analysis Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2462—Approximate or statistical queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- 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"
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Human Resources & Organizations (AREA)
- Probability & Statistics with Applications (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Software Systems (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- Development Economics (AREA)
- Fuzzy Systems (AREA)
- Computational Linguistics (AREA)
- Marketing (AREA)
- Remote Sensing (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- General Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a road network passenger flow state deduction method and system oriented to passenger guidance, wherein the method comprises the following steps: acquiring historical passenger flow data of each station of urban rail transit; based on historical passenger flow volume data and weather data of each station, relevant factors influencing passenger flow are analyzed and determined by utilizing Pearson correlation coefficients; performing clustering analysis on historical passenger flow volume data of all the sites to obtain classification results of all the sites, and analyzing the passenger flow volume distribution conditions of different types of the sites; and (3) building a short-time passenger flow prediction model, collecting real-time passenger flow data of various stations, and inputting the real-time passenger flow data of various stations into the short-time passenger flow prediction model to obtain a short-time passenger flow prediction result. The invention improves the accuracy, real-time property and adaptability to emergency.
Description
Technical Field
The invention relates to the technical field of rail transit, in particular to a road network passenger flow state deduction method and system oriented to passenger guidance.
Background
The existing urban rail transit passenger flow dynamic prediction method is based on the historical passenger flow OD data, and urban rail transit passenger flow prediction is carried out by using a probability tree, a traditional machine learning model and the like. The method comprises the specific implementation steps of obtaining historical passenger flow OD data of a traffic network, wherein the historical passenger flow OD data comprises historical time and time-period passenger flow OD data corresponding to the date of the historical time, and the historical time comprises the date and the time period; and performing cluster analysis on the time-interval passenger flow OD data, and determining a plurality of categories corresponding to the time-interval passenger flow OD data. And determining the category of the predicted passenger flow OD data corresponding to the time to be predicted, and determining the predicted passenger flow OD data of the time to be predicted according to the category.
Most of the existing urban rail transit passenger flow prediction methods use time series models, the models generally have certain assumptions and high model complexity, and the methods are difficult to apply to actual production practice and have the problems of low accuracy and low efficiency.
The existing urban rail transit passenger flow prediction method mainly depends on historical passenger flow OD data, the real-time performance of the data is weak, and the performance is poor when unpredictable situations such as emergency situations, short-time changes and the like are met.
The existing urban rail transit passenger flow prediction method does not carry out scientific classification on all stations of a road network, lacks personalized time and space analysis on key stations, and loses certain precision in actual production practice.
Disclosure of Invention
In view of this, the invention provides a passenger-oriented road network passenger flow state deduction method and system, which improve the accuracy, real-time performance and adaptability to emergency situations of prediction.
The invention provides a passenger-oriented road network passenger flow state deduction method, which comprises the following steps: acquiring historical passenger flow data of each station of urban rail transit; based on historical passenger flow volume data and weather data of each station, relevant factors influencing passenger flow are analyzed and determined by using Pearson correlation coefficients; performing clustering analysis on historical passenger flow data of each station to obtain classification results of various stations, and analyzing passenger flow distribution conditions of different types of stations; and (3) building a short-time passenger flow prediction model, collecting real-time passenger flow data of various stations, and inputting the real-time passenger flow data of various stations into the short-time passenger flow prediction model to obtain a short-time passenger flow prediction result.
Further, the passenger flow volume data at least comprises the station number of the passenger entering and exiting the station, the station name, the time of entering and exiting the station and the card type.
Further, the step of determining relevant factors influencing passenger flow by using Pearson correlation coefficient analysis comprises: dividing the historical passenger flow volume data of each station into passenger flow volume data of a plurality of time periods; calculating a corresponding Pearson correlation coefficient by using the passenger flow volume data and the weather data of each time period; comparing the passenger flow volume data of each time period with the Pearson correlation coefficient corresponding to the weather data, and determining related passenger flow volume factors and related weather factors which influence the passenger flow; and selecting daily passenger flow data of part of the sites in one month and average passenger flow data of different time periods of each day in one week for analysis, and determining two factors of whether the work day and the passenger flow peak period influence the passenger flow.
Further, the calculation formula of the Pearson correlation coefficient is as follows:
wherein, X i Indicating the amount of traffic in the ith time slot,
and S X Respectively representing the mean value and the variance of the passenger flow; y is i Indicates the corresponding influence factor value of the ith time period,
and S Y Respectively representing the mean and variance of the corresponding influence factors; and r is Pearson correlation coefficient.
Further, the relevant factors influencing the passenger flow include temperature, humidity, visibility, precipitation, cloud cover, previous time period passenger flow/number, whether working day and whether peak time period.
Further, the step of performing cluster analysis on the historical passenger flow volume data of each site includes: acquiring the category number of each site and a site passenger flow data set; solving the maximum value and the minimum value of passenger flows of all stations; taking a K +1 equant point between the maximum value and the minimum value as an initial clustering center; dividing all the sites into the clustering categories with the closest passenger flow; accumulating and summing the difference value between the passenger flow of each vector station in the cluster and the cluster center to serve as a convergence judgment condition of the algorithm; finding the mean value of each class of passenger flow as a new clustering center, and finishing iteration when the difference of clustering structures of two consecutive times is smaller than a set threshold value; and outputting a site clustering result, wherein the site clustering result comprises a high passenger flow site, a medium passenger flow site and a low passenger flow site.
Further, the step of analyzing the passenger flow distribution conditions of different types of sites includes: the distribution rules of the historical passenger flow data of the high passenger flow station, the medium passenger flow station and the low passenger flow station in time and space are analyzed respectively, and different passenger flow peak time periods of various stations are determined.
Further, the step of building a short-time passenger flow prediction model comprises the following steps: constructing an input layer, a hidden layer and an output layer of a short-time passenger flow prediction model; dividing historical passenger flow volume data of each station into a training set and a test set according to a certain proportion, collecting real-time passenger flow volume data of each station, and adding the real-time passenger flow volume data into the test set; training the short-time passenger flow prediction model by using a training set; and verifying the trained short-time passenger flow prediction model by using a test set.
Further, the step of inputting the real-time passenger flow volume data of each station into the short-time passenger flow prediction model to obtain the short-time passenger flow prediction result comprises the following steps: preprocessing the real-time passenger flow data of each station, and dividing the preprocessed real-time passenger flow data of each station by selecting different time granularities; and carrying out predictive analysis on the real-time passenger flow data in different time to obtain the passenger flow state of each station.
In a second aspect of the present invention, a passenger-oriented guidance-oriented road network passenger flow state deduction system includes: a memory for storing a computer program; a processor for implementing the steps of the passenger-oriented guidance road network passenger flow state deduction method when executing the computer program.
According to the passenger guidance oriented road network passenger flow state deduction method and system, pearson correlation coefficients are used for analysis, and whether factors such as weather conditions, historical passenger flow, whether the factors are in peak time periods or not, whether the factors are in working days and the like are determined as short-time passenger flow influence factors of rail transit; dividing actual rail transit stations into three types of stations of high, medium and low passenger flow by using an improved K-means clustering algorithm, analyzing the distribution rule of the passenger flow of each type of stations in time and space, and determining different passenger flow peak time periods of each type of stations; the urban rail transit short-time passenger flow prediction method based on the long-time memory neural network and the gate control circulation unit is adopted to carry out prediction analysis on the passenger flow of different passenger flow type stations in different time, so that the prediction accuracy, the real-time performance and the adaptability to emergency situations are improved.
Drawings
For purposes of illustration and not limitation, the present invention will now be described in accordance with its preferred embodiments, particularly with reference to the accompanying drawings, in which:
fig. 1 is a flowchart of a method for deducing passenger traffic state of a road network oriented to passenger guidance according to an embodiment of the present invention;
FIG. 2 is a flow chart of analyzing passenger flow characteristics based on an improved K-means clustering algorithm;
FIG. 3 is a schematic diagram of high, medium and low traffic stations corresponding to peak hours;
FIG. 4 is a schematic diagram of a short-time passenger flow prediction model;
fig. 5 is a schematic structural diagram of a passenger-oriented guidance road network passenger flow state deduction system according to another embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a detailed description of the present invention will be given below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention, and the described embodiments are merely a subset of the embodiments of the present invention, rather than a complete embodiment. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
Fig. 1 is a flowchart of a method for deducing passenger-oriented road network traffic state according to an embodiment of the present invention. Specific implementations of the present application are described in detail below with reference to the accompanying drawings.
Referring to fig. 1, the passenger-oriented guidance-oriented road network passenger flow state deduction method includes the following steps:
s100, AFC data of each station of the urban rail transit in actual operation are obtained.
AFC data is historical passenger flow data of passengers entering and leaving a station in the whole network collected by an automatic fare collection system for urban rail transit, and the passenger flow data generally comprises attributes such as passenger station number, station name, time of entering and leaving the station, card type and the like. The data has the characteristics of high precision, strong quasi-real-time performance and the like in passenger flow records, and is often applied to the research on passenger travel behaviors, passenger flow distribution characteristics, flow characteristics and the like. The quasi-real-time performance of AFC data is mainly embodied in that after a passenger punches a card in a station to go in and out, the system uploads the time of the station to go in and out, the station to go in and out and the like according to a certain time interval, so that the AFC data is used as a data basis for carrying out passenger flow OD dynamic prediction, the prediction result is more accurate, and the real-time performance of passenger flow can be more embodied.
In this embodiment, the dynamic OD data in the AFC data is mainly acquired, where the dynamic OD data includes time and time-period passenger flow OD data corresponding to the time, and the time includes date and time period.
And S200, based on AFC data and weather data, determining relevant factors influencing passenger flow by utilizing Pearson correlation coefficient analysis.
In this embodiment, the Pearson correlation coefficient calculation formula is as follows:
wherein, X i Indicating the amount of traffic in the ith time period,
and S X Respectively representing the mean value and the variance of the passenger flow; y is i Indicates the corresponding influence factor value for the ith period,
and S Y Respectively, mean and variance of the respective influencing factors. The Pearson correlation coefficient r has a value range of [ -1,1]A smaller absolute value indicates a lower correlation of the influencing factor with the passenger flow volume. The positive and negative values of r represent different correlations, with positive values representing positive correlations and negative values representing negative correlations.
Weather has a great influence on urban rail transit passenger flow, and particularly, some passengers who go out for shopping and play can flexibly select travel modes and time according to weather conditions. In this embodiment, passenger flow influence analysis is mainly performed on five weather conditions, namely, temperature, humidity, visibility, precipitation and cloudiness, and the obtained Pearson correlation coefficients are all within a 99% confidence interval, as shown in table 1.
TABLE 1 Pearson correlation coefficient of correlation factors affecting passenger flow
Note: * Indicating significant correlation at the 0.01 level (99% confidence); significance (two-tailed) refers to the differential check value
Intuitively, rail transit passenger flow is influenced by whether the rail transit passenger flow works or not and whether the rail transit passenger flow is in a rush hour or not, and the rail transit passenger flow fluctuates. And selecting daily passenger flow data of part of the stations in one month and average passenger flow data of different time periods in one week for analysis, determining whether two factors, namely working days and whether the peak time of the passenger flow is influenced on the rail passenger flow, and using the two factors as influence factors for subsequent passenger flow prediction.
According to the analysis, factors influencing the rail transit passenger flow change have diversity, and part or all of 8 factors are used as parameter indexes for short-time passenger flow analysis and prediction according to actual conditions, and specific parameters are shown in table 2.
TABLE 2 relevant factors influencing passenger flow
As can be seen from table 2, relevant factors affecting the passenger flow include temperature, humidity, visibility, precipitation, cloudiness, previous-period passenger flow/number, whether the day of work is on, and whether the peak period is on.
In the embodiment, for AFC data of each station of the rail transit in actual operation, a Pearson correlation coefficient is used for analysis, and weather conditions, historical passenger flow volume, whether the time period is a peak time period, whether the time period is a working day and the like are determined as short-time passenger flow influence factors of the rail transit.
And S300, analyzing the passenger flow characteristics based on the improved K-means clustering algorithm.
Based on K-means algorithm improvement, a station clustering algorithm (SK-means) for rail transit is provided for clustering station passenger flow, the passenger flow classification standard of various Stations is determined, and the distribution condition of different types of passenger flow Stations is analyzed. Based on the general rule of passenger flow distribution of rail transit stations, the SK-means algorithm improves the selection of the initial clustering center. As the passenger flow of most stations of the rail transit in the same city is uniformly distributed between the maximum value and the minimum value of the passenger flow of the stations, the K +1 equal division point between the maximum value and the minimum value of the passenger flow is selected as an initial clustering central point by the algorithm. The stations are divided into three types of high, medium and low passenger flow.
Fig. 2 is a flowchart of a station clustering algorithm for rail transit. Referring to fig. 2, the station clustering algorithm for rail transit includes the following steps:
s301, acquiring the class number K of each site and a site passenger flow data set N;
s302, solving the maximum value and the minimum value of passenger flows of N stations, wherein the maximum value and the minimum value are as follows: max, min = getMaxAndMin (N);
s303, taking a K +1 bisector between the maximum value and the minimum value as an initial clustering center, wherein the initial clustering center is centlist = initcentlist (max, min, K + 1);
s304, dividing the N sites into the clustering classes with the closest passenger flow, wherein the closest clustering class is clusterDict = minDistance (N, centerList);
s305, accumulating and summing the difference value between the passenger flow of each vector station in the cluster and the cluster center, and taking the sum as a convergence judgment condition of the algorithm; wherein nVar = getVar (clusterDict, centerList);
s306, finding the mean value of each type of passenger flow to serve as a new clustering center, and ending iteration when the difference of the clustering structures of two consecutive times is less than 0.0001.
while abs(nVar-oVar)>＝0.0001；
centerList＝getcenterList(clusterDict)；
clusterDict＝minDistance(N,centerList)；
oVar＝nVar；
nVar＝getVar(clusterDict,centerList).
And S307, outputting the clustering result clusterDict of the sites.
An example of the partitioning result is as follows:
s308, analyzing the distribution rule of the passenger flow of various sites in time and space, and determining different passenger flow peak time periods of various sites.
In the rail transit operation time period, the passenger flow of each station is influenced by the time of residents going to and from work and the time of going to and from school, and the regular change of the peak in the morning and at night is presented. And the morning and evening peak hours may be different for different types of sites in different cities. The corresponding peak periods are set based on a characterization of historical AFC data.
Taking the average station passenger flow volume of the weekdays (except the working day of rest) from 9 to 12 months in 2019 of a certain city and the days of monday (except the holidays) as examples, the passenger flow time distribution characteristics of high, medium and low passenger flow stations (respectively selecting 2 stations) are analyzed, as shown in fig. 3.
Setting the peak time period of working days of various stations as 7;
setting the non-working day peak time period of the high passenger flow rate station as 14;
and setting the non-working day peak time period of the medium passenger flow station as 17.
In this embodiment, a K-means clustering algorithm (K-means clustering, K-means) is improved to divide actual rail transit stations into three types of stations, namely high, medium and low passenger flow rates, and the distribution rules of the passenger flow rates of the various stations in time and space are analyzed to determine different passenger flow peak time periods of the various stations
S400, building a short-time passenger flow prediction model.
In this embodiment, an LSTM or GRU model is used as a basic processing unit, and a seven-layer Short-term Passenger Flow prediction model structure (spf) is built, and the model structure is shown in fig. 4.
As shown in fig. 4, the SPFF model includes an input layer, a hidden layer, and an output layer. The input layer and the output layer of the SPFF model are all connected layers. The input layer meets the input requirement of neurons in the hidden layer by carrying out primary processing on AFC sample data; the output layer maps a plurality of actual results of the hidden layer into expected results of the model through a full-connection network again. The hidden layer of the SPPF model comprises seven layers of neural network layers and four layers of Dropout layers which are mutually interpenetrated, wherein the Dropout layers can temporarily remove the neurons from the network according to a set probability, so that overfitting is effectively prevented; the neurons of the neural network layer are unified as LSTM or GRU. In fig. 4, the subscripts t-1 and t represent the last time and the current time, x and h represent the input and output data vectors, respectively, C represents the state vector of the hidden layer,
a candidate value vector created by a tanh function (as shown in equation 2) is represented, σ represents a sigmoid function (as shown in equation 3), and f, i, and o are calculation results of a plurality of σ functions (as shown in equation 4, where W is a weight matrix and b is a bias value).
The expression of the output of the SPFF model is:
the calculated relationship between other intermediate quantities and output quantities in the SPFF model is as follows:
in the SPFF model, the hyperbolic tangent function tanh is used as the activation function (as shown in equation 2), the Adma is used as the optimizer, and the Mean Square Error (MSE) is used as the loss function. Furthermore, there are 3 other main parameters in the SPFF model as follows:
(1) Time granularity (Tg)
The time granularity Tg refers to a time interval for collecting data, and in this embodiment refers to a time interval for extracting passenger flow volume data. For short-term traffic prediction, the time granularity is typically selected to be no more than 15 minutes. The amount of information contained in different time granularities can be different, the larger the time granularity is, the more the information amount is, the more accurate the prediction is, but the larger time interval can cause the slight change of the passenger flow to be not easy to be perceived. In order to research the optimal model for short-time passenger flow prediction, different time granularity analysis can be selected to compare prediction results.
(2) Time step (Ts)
The time step Ts refers to the sequence length of a single sample, i.e. the number of data periods contained in a single sample. Different Ts will influence the length of training time, and although the iteration speed is accelerated by over-small Ts, the problem of easy falling into local optimal solution also exists; too large Ts also loses prediction accuracy due to reduced correlation of data in the same sample. There are studies that yield better predictions for short-term traffic predictions of different granularity, with Ts values of 3 or 6.
(3) Iteration time (It)
The number of iterations It is related to the model convergence speed. As the number of iterations increases, the model gradually converges and the loss value of the model gradually decreases, but when the loss value decreases to a lower value, the loss value cannot decrease again with the iteration of the model. Therefore, the excessive number of iterations cannot make the model loss value always decrease, and the excessive training time is consumed.
In order to improve the prediction effect of the SPPF model on short-time passenger flow, prediction practice is respectively carried out on high, medium and low passenger flow stations, AFC passenger flow data of each station are correspondingly divided into a training set and a test set according to the proportion of 8. The time granularity Tg is set to be not less than 5 minutes, the time step Ts is set to be three types of 1/3/6, and the iteration number It is set to be 800.
S500, collecting real-time passenger flow data of various stations, inputting the real-time passenger flow data of each station into a short-time passenger flow prediction model to predict short-time passenger flow, and obtaining a short-time passenger flow prediction result.
The short-time passenger flow prediction model is used for predicting and analyzing the passenger flow of real-time passenger flow data of various stations in different time periods, and the short-time passenger flow prediction result generated by prediction is synchronously uploaded to a corresponding module server for urban rail transit station group passenger flow analysis, so that accurate road network passenger flow state is provided for personalized travel of passengers, and accurate travel time information is provided for dynamic path planning.
The passenger guidance oriented road network passenger flow state deduction method utilizes Pearson correlation coefficients to analyze and determine whether factors such as weather conditions, historical passenger flow, whether the factors are in peak time periods, whether the factors are in working days and the like are determined as short-time passenger flow influence factors of rail transit; dividing actual rail transit stations into three types of stations of high, medium and low passenger flow by using an improved K-means clustering algorithm, analyzing the distribution rule of the passenger flow of each type of stations in time and space, and determining different passenger flow peak time periods of each type of stations; two urban rail transit short-time passenger flow prediction methods respectively based on a long-time memory neural network and a gate control circulation unit are adopted to carry out prediction analysis on the passenger flow of different passenger flow type stations in different time. Compared with the traditional road network passenger flow prediction method based on historical OD data, the accuracy, the real-time performance and the adaptability to emergency situations are improved.
In correspondence to the above method embodiment, referring to fig. 5, fig. 5 is a schematic diagram of a passenger-oriented guidance road network passenger flow state deduction system according to another embodiment of the present invention, where the system 100 may include:
a memory 101 for storing a computer program;
the processor 102, when executing the computer program stored in the memory 101, may implement the following steps:
acquiring historical passenger flow data of each station of urban rail transit; based on historical passenger flow volume data and weather data of each station, relevant factors influencing passenger flow are analyzed and determined by using Pearson correlation coefficients; performing clustering analysis on historical passenger flow data of each station to obtain classification results of various stations, and analyzing passenger flow distribution conditions of different types of stations; and (3) building a short-time passenger flow prediction model, collecting real-time passenger flow data of various stations, and inputting the real-time passenger flow data of various stations into the short-time passenger flow prediction model to obtain a short-time passenger flow prediction result.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A passenger-oriented guidance-oriented road network passenger flow state deduction method is characterized by comprising the following steps:
acquiring historical passenger flow data of each station of urban rail transit;
based on historical passenger flow volume data and weather data of each station, relevant factors influencing passenger flow are analyzed and determined by using Pearson correlation coefficients;
performing clustering analysis on historical passenger flow data of each station to obtain classification results of various stations, and analyzing passenger flow distribution conditions of different types of stations;
and (3) building a short-time passenger flow prediction model, acquiring real-time passenger flow data of various stations, and inputting the real-time passenger flow data of the various stations into the short-time passenger flow prediction model to obtain a short-time passenger flow prediction result.
2. The passenger-oriented guidance-oriented road network passenger flow state deduction method as claimed in claim 1, wherein the passenger flow data at least comprises passenger station number, station name, station time and card type.
3. The passenger-oriented guidance road network passenger flow state deduction method according to claim 1, wherein the step of determining relevant factors influencing the passenger flow by using Pearson correlation coefficient analysis comprises:
dividing historical passenger flow data of each station into passenger flow data of a plurality of time periods;
calculating a corresponding Pearson correlation coefficient by using the passenger flow volume data and the weather data of each time period;
comparing the passenger flow volume data of each time period with the Pearson correlation coefficient corresponding to the weather data, and determining related passenger flow volume factors and related weather factors which influence the passenger flow;
and selecting daily passenger flow data of part of the sites in one month and average passenger flow data of different time periods of each day in one week for analysis, and determining two factors of whether the work day and the passenger flow peak period influence the passenger flow.
4. The passenger-oriented guidance road network passenger flow state deduction method according to claim 3, wherein the Pearson correlation coefficient is calculated as follows:
wherein, X i Indicating the amount of traffic for the ith time period,
and S X Respectively representing the mean value and the variance of the passenger flow; y is i Indicates the corresponding influence factor value of the ith time period,
and S Y Respectively representing the mean and variance of the corresponding influence factors; r is Pearson correlation coefficient.
5. The passenger-oriented guidance-oriented road network passenger flow state deduction method as claimed in claim 3, wherein the relevant factors influencing the passenger flow include temperature, humidity, visibility, precipitation, cloud cover, previous period passenger flow/number, whether working day and whether peak time period.
6. The passenger-oriented guidance road network passenger flow state deduction method as claimed in claim 1, wherein the step of performing cluster analysis on the historical passenger flow data of each station comprises:
acquiring the category number of each site and a site passenger flow data set;
solving the maximum value and the minimum value of passenger flows of all stations;
taking a K +1 equant point between the maximum value and the minimum value as an initial clustering center;
dividing all the sites into the clustering categories with the closest passenger flow;
accumulating and summing the difference value between the passenger flow of each vector station in the cluster and the cluster center to serve as a convergence judgment condition of the algorithm;
finding the mean value of each class of passenger flow as a new clustering center, and finishing iteration when the difference of clustering structures of two consecutive times is smaller than a set threshold value;
and outputting a site clustering result, wherein the site clustering result comprises a high passenger flow site, a medium passenger flow site and a low passenger flow site.
7. The passenger-oriented guidance road network passenger flow state deduction method as claimed in claim 6, wherein the step of analyzing the passenger flow distribution of different types of stations comprises:
the distribution rules of the historical passenger flow data of the high passenger flow station, the medium passenger flow station and the low passenger flow station in time and space are analyzed respectively, and different passenger flow peak time periods of various stations are determined.
8. The passenger-oriented guidance road network passenger flow state deduction method according to claim 1, wherein the step of building a short-time passenger flow prediction model comprises:
constructing an input layer, a hidden layer and an output layer of a short-time passenger flow prediction model;
dividing historical passenger flow volume data of each station into a training set and a test set according to a certain proportion, collecting real-time passenger flow volume data of each station, and adding the real-time passenger flow volume data into the test set;
training a short-time passenger flow prediction model by using a training set;
and verifying the trained short-time passenger flow prediction model by using a test set.
9. The passenger-oriented guidance road network passenger flow state deduction method as claimed in claim 1, wherein the step of inputting the real-time passenger flow data of each station into the short-time passenger flow prediction model to obtain the short-time passenger flow prediction result comprises:
the real-time passenger flow data of each station is preprocessed,
selecting real-time passenger flow data of each station after different time granularity division preprocessing;
and performing predictive analysis on the real-time passenger flow data in different time to obtain the passenger flow state of each station.
10. A passenger-oriented guidance-oriented road network passenger flow state deduction system, comprising:
a memory for storing a computer program;
processor for implementing the steps of the passenger oriented guidance road network traffic state deduction method according to any of claims 1 to 9 when executing said computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210922005.4A CN115269758A (en) | 2022-08-02 | 2022-08-02 | Passenger-guidance-oriented road network passenger flow state deduction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210922005.4A CN115269758A (en) | 2022-08-02 | 2022-08-02 | Passenger-guidance-oriented road network passenger flow state deduction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115269758A true CN115269758A (en) | 2022-11-01 |
Family
ID=83747535
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210922005.4A Pending CN115269758A (en) | 2022-08-02 | 2022-08-02 | Passenger-guidance-oriented road network passenger flow state deduction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115269758A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116778739A (en) * | 2023-06-20 | 2023-09-19 | 深圳市中车智联科技有限公司 | Public transportation scheduling method and system based on demand response |
| CN117408393A (en) * | 2023-12-06 | 2024-01-16 | 华中科技大学 | Prediction method and system for comprehensive passenger transportation hub traffic flow under abnormal event |
-
2022
- 2022-08-02 CN CN202210922005.4A patent/CN115269758A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116778739A (en) * | 2023-06-20 | 2023-09-19 | 深圳市中车智联科技有限公司 | Public transportation scheduling method and system based on demand response |
| CN117408393A (en) * | 2023-12-06 | 2024-01-16 | 华中科技大学 | Prediction method and system for comprehensive passenger transportation hub traffic flow under abnormal event |
| CN117408393B (en) * | 2023-12-06 | 2024-03-19 | 华中科技大学 | Prediction method and system for comprehensive passenger transportation hub traffic flow under abnormal event |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110570651B (en) | Road network traffic situation prediction method and system based on deep learning | |
| WO2023056696A1 (en) | Urban rail transit short-term passenger flow forecasting method based on recurrent neural network | |
| CN107610464A (en) | A kind of trajectory predictions method based on Gaussian Mixture time series models | |
| Wang | Predicting tourism demand using fuzzy time series and hybrid grey theory | |
| CN115269758A (en) | Passenger-guidance-oriented road network passenger flow state deduction method and system | |
| CN113096388B (en) | Short-term traffic flow prediction method based on gradient lifting decision tree | |
| CN107978153B (en) | A kind of multimode traffic factors influencing demand analysis method based on space vector autoregression model | |
| CN113159364A (en) | Passenger flow prediction method and system for large-scale traffic station | |
| CN110390349A (en) | Bus passenger flow volume based on XGBoost model predicts modeling method | |
| CN111860989B (en) | LSTM neural network short-time traffic flow prediction method based on ant colony optimization | |
| CN111582559B (en) | Arrival time estimation method and device | |
| CN109376906B (en) | Travel time prediction method and system based on multi-dimensional trajectory and electronic equipment | |
| CN112183605B (en) | Civil aviation control sector classification method based on operation characteristics | |
| CN116681176B (en) | Traffic flow prediction method based on clustering and heterogeneous graph neural network | |
| CN113380071B (en) | Parking space level parking demand state evaluation method for large complex | |
| CN112884222B (en) | Time-period-oriented LSTM traffic flow density prediction method | |
| CN113205223A (en) | Electric quantity prediction system and prediction method thereof | |
| CN112149922A (en) | Method for predicting severity of accident in exit and entrance area of down-link of highway tunnel | |
| Zekić-Sušac et al. | Cluster analysis and artificial neural networks in predicting energy efficiency of public buildings as a cost-saving approach | |
| CN113674524A (en) | LSTM-GASVR-based multi-scale short-time traffic flow prediction modeling and prediction method and system | |
| CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
| CN110490365B (en) | Method for predicting network car booking order quantity based on multi-source data fusion | |
| CN113553350B (en) | Traffic flow partition model for similar evolution mode clustering and dynamic time zone division | |
| CN117912235B (en) | Planning data processing method and system for smart city | |
| CN114117903B (en) | Short-time passenger flow prediction method for rail transit based on bp neural network |
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 |