CN109215344B - Method and system for urban road short-time traffic flow prediction - Google Patents

Method and system for urban road short-time traffic flow prediction Download PDF

Info

Publication number
CN109215344B
CN109215344B CN201811133463.XA CN201811133463A CN109215344B CN 109215344 B CN109215344 B CN 109215344B CN 201811133463 A CN201811133463 A CN 201811133463A CN 109215344 B CN109215344 B CN 109215344B
Authority
CN
China
Prior art keywords
learning
student
traffic flow
students
prediction
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.)
Active
Application number
CN201811133463.XA
Other languages
Chinese (zh)
Other versions
CN109215344A (en
Inventor
张鹏翔
曹扬
洒科进
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CETC Big Data Research Institute Co Ltd
Original Assignee
CETC Big Data Research Institute Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by CETC Big Data Research Institute Co Ltd filed Critical CETC Big Data Research Institute Co Ltd
Priority to CN201811133463.XA priority Critical patent/CN109215344B/en
Publication of CN109215344A publication Critical patent/CN109215344A/en
Application granted granted Critical
Publication of CN109215344B publication Critical patent/CN109215344B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0112—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from the vehicle, e.g. floating car data [FCD]
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G08G1/0145—Measuring and analyzing of parameters relative to traffic conditions for specific applications for active traffic flow control

Abstract

The invention discloses a method and a system for predicting short-term traffic flow of urban roads. The method specifically comprises the following steps: carrying out data preprocessing; selecting an LSSVM kernel function; setting parameters such as problem solving dimension d, maximum iteration number Mit, student group n and the like; optimally ordering the students of the upper generation and the lower generation according to the adaptive values; learning in the 'teaching' and 'learning' stages; if the maximum iteration time Mit is reached and a prediction error condition is met, setting a punishment parameter c and a kernel width parameter sigma of an LSSVM (least squares support vector machine) according to the optimal student subject score, and constructing a prediction model based on the LSSVM to predict the short-term traffic flow of the urban road; carrying out reverse normalization output; and performing performance evaluation according to the prediction evaluation index.

Description

Method and system for urban road short-time traffic flow prediction
Technical Field
The invention relates to the technical field of traffic flow prediction of intelligent traffic systems, in particular to a method and a system for predicting short-term traffic flow of urban roads.
Background
With the continuous and rapid development of economy, the quantity of retained urban automobiles is high, and the accompanying urban traffic congestion becomes a global challenge which puzzles human lives. Traffic pressure is relieved by means of traffic induction and traffic control, and as one of core contents of modern intelligent traffic systems, accurate traffic flow prediction is the basis and key for solving traffic jam and constructing a smart city traffic management system.
The traditional short-time traffic flow prediction method has a time sequence method, an autoregressive model, a grey theory and the like. The method is relatively mature in technology and simple in structure, but is basically established in a linear rule, so that the traffic flow sequence with complex and nonlinear urban traffic and large fluctuation of influencing factors is difficult to predict. In recent years, intelligent technology is developed vigorously, and fuzzy reasoning, neural networks, deep learning and the like are well applied to short-time traffic flow prediction. For example, prior art patent application nos.: in the patent application of CN2018100451199, an artificial neural network particle swarm algorithm is adopted to predict traffic road congestion, a first receiving module is arranged on a cloud server and used for receiving the first traffic data and the second traffic data, and a particle swarm algorithm program module arranged on the cloud server and used for processing data received by the first receiving module to obtain first predicted data; the transmission module is arranged on the cloud server and used for transmitting the first prediction data to equipment on the terminal side; the device on the terminal side comprises a second receiving module for receiving data returned by the cloud server; the storage device is used for storing the data received by the second receiving module; and the updating module is arranged on the cloud server and used for acquiring the first traffic data and the second traffic data in real time, taking the first traffic data and the second traffic data as new input variables, continuously learning based on the particle swarm algorithm program module, and continuously optimizing the first predicted data, wherein the artificial neural network method has low convergence speed, the prediction precision is influenced by the initial connection weight and the threshold parameter, and meanwhile, the artificial neural network method can converge to a local optimal point. The patent application numbers are: the CN201510478215 provides a traffic flow prediction method based on a deep learning neural network structure in the prior art, which collects various traffic flow data, trains the collected various traffic flow data by using a deep automatic encoder model, adjusts the deep automatic encoder model in the training process, and finally predicts short-term traffic flow by using the adjusted deep automatic encoder model, wherein the deep learning model has superior performance, but the network level is deep and is a black box model, so that the prediction is more complex and uncertain. In the prior art, patent publication No. CN105389980B provides a short-term traffic flow prediction method based on a long-term and short-term memory recurrent neural network, and according to the prediction time interval of the short-term traffic flow, the input historical traffic flow data is aggregated; preprocessing the aggregated historical traffic flow; setting reasonable parameters for a long-time memory recurrent neural network; training the neural network prediction model by using the preprocessed data; a prediction model is called to predict the traffic flow of a specified time interval and estimate the prediction error, the method also belongs to a neural network model, short-time prediction is carried out through parameter setting, although the calculation time and the precision are improved, the characteristic that the convergence speed of an artificial neural network method is slow cannot be avoided. The Least Square Support Vector Machine (LSSVM) algorithm is based on a VC dimension theory of a statistical learning theory and a structure risk minimum principle, and is a novel Machine learning method. The LSSVM regression prediction method has the advantages of high accuracy, low complexity, good robustness, capability of overcoming dimension disaster and the like. However, when the traditional LSSVM predicts the traffic flow, the penalty parameter c and the kernel parameter σ are generally selected by a trial and error method, so that the prediction process is not intelligent and has large and uncontrollable error. Therefore, the researchers can improve the selection of the parameters of the least square support vector machine. And a particle swarm optimization least square support vector machine, a genetic algorithm improved least square support vector machine and a drosophila algorithm optimized least square support vector machine are provided. And a group intelligent optimization algorithm is introduced, so that the core parameter selection of the LSSVM is faster, more intelligent and more accurate, and the traffic flow prediction precision is improved. However, the optimization LSSVM algorithm is complex, the iterative convergence speed is slow, and the parameters are difficult to achieve global optimization.
Disclosure of Invention
At least one of the objectives of the present invention is to overcome the above problems in the prior art, and to provide a method and a system for predicting short-term traffic flow in urban roads, which have strong global optimization capability, simple steps, and high prediction accuracy.
In order to achieve the above object, the present invention adopts the following aspects.
A method for urban road short-time traffic flow prediction, the method comprising: data preprocessing, kernel function selection, algorithm parameter initialization, random data generation, adaptive value calculation, optimal sequencing, stage learning, iteration termination judgment, LSSVM (least squares support vector machine) based optimal short-term traffic flow prediction, inverse normalization output and prediction evaluation;
the data preprocessing specifically comprises the following steps: carrying out data normalization processing on the selected historical traffic flow data;
the algorithm parameter initialization specifically comprises the following steps: selecting an improved teaching and learning algorithm, wherein the improved teaching and learning algorithm comprises the settings of teachers and students, setting parameters, specifically setting the problem dimension d of the improved teaching and learning algorithm, the maximum iteration number Mit and the student group n, setting the upper limit value and the lower limit value of the punishment parameter c and the nuclear parameter sigma of the least square support vector machine LSSVM, and initializing the parameters of individual students;
the adaptive value calculation specifically comprises: calculating the adaptive value of the student score by using the LSSVM, and taking the highest adaptive value as a teacher;
the optimal sequence specifically includes: optimally ordering students of the next generation according to the adaptive values;
the iteration termination judgment specifically comprises the following steps: and judging according to the maximum iteration times and the prediction error condition.
The algorithm parameter initialization comprises the following steps: randomly generating initial solution X of individual subject scores of students0(ii) a Random initialization learning step length ri。
And (3) setting a punishment parameter c and a kernel width parameter sigma of the LSSVM according to the optimization of the improved teaching and learning algorithm, taking the adaptive value calculated by the LSSVM as the subject performance of the student, and taking the adaptive value with the highest adaptive value as a teacher.
The improved teaching and learning algorithm comprises the following steps: and (3) associating the students of the previous generation with the current generation, solving the excellent students of the two generations to be sorted according to the mapping, performing teaching and learning stage learning, and calculating and storing the corresponding optimal adaptive value through the LSSVM.
Learning in the teaching phase, characterised by being based on formulas
Figure GDA0003019766240000041
And updating the conditions to evolve the subject scores of the students when
Figure GDA0003019766240000042
Adapted value of greater than
Figure GDA0003019766240000043
And updating the subject scores of the students.
The Difference is the Difference between the average performance of the current students of the subject and the level of the teacher, and is represented by a pass formula
Figure GDA0003019766240000044
Calculating TF as teaching factor, and calculating by the formula TF as round [1+ rand (0,1)]And (6) obtaining.
Learning in learning stage, characterized by randomly picking two students XpAnd Xq：
When X is presentp<XqTime, show student XqScore comparing student XpIs excellent, so XpTo XqLearning according to a formula
Figure GDA0003019766240000045
Updating the student subject scores;
when X is presentp>XqTime, show student XpScore comparing student XqIs excellent, so XpTo XqLearning according to a formula
Figure GDA0003019766240000046
And updating the subject scores of the students.
The dimension d of the learning subject vector is 2.
The maximum number of iterations Mit is 100 and the number of student populations is 30.
The prediction evaluation specifically includes: selecting the average absolute percentage error, the relative error and the root mean square error as the prediction performance evaluation indexes, wherein the specific formula is as follows:
Figure GDA0003019766240000047
Figure GDA0003019766240000048
Figure GDA0003019766240000049
in the formula: y isiThe real value of the traffic flow is obtained; y'iIndicating traffic flowMeasuring; n is the number of traffic flow collection points,
the technical scheme of the invention also comprises a system for predicting the short-time traffic flow of urban roads, which comprises the following steps: the system comprises a display, an input and output device, at least one processor, a memory in communication connection with the at least one processor, and a power supply device for supplying power;
wherein, the display is used for displaying the prediction result; the input and output equipment is used for inputting initialization parameters; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the above method.
In summary, due to the adoption of the technical scheme, the invention at least has the following beneficial effects:
(1) the short-time traffic flow prediction is carried out by optimizing a least square support vector machine through an improved 'teaching and learning' algorithm, so that the prediction precision is effectively improved, and the calculation processing time is obviously shortened.
(2) The method and the system have the application scene that the urban road with strong randomness of short-term traffic flow is adopted, and the prediction precision is still high, so the popularization and application prospect is optimistic.
(3) The urban road short-time traffic flow prediction method and system provided by the invention are convenient for traffic management departments to scientifically and reasonably dredge traffic flow, and alleviate urban congestion to a certain extent. Meanwhile, technical support is provided for constructing an intelligent traffic management system.
Drawings
Fig. 1 is a flowchart of a method for urban road short-time traffic flow prediction according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a system for urban road short-time traffic flow prediction according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and embodiments, so that the objects, technical solutions and advantages of the present invention will be more clearly understood. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Fig. 1 illustrates a method for urban road short-time traffic flow prediction according to an embodiment of the present invention. The method of this embodiment comprises the steps of:
step 101 data preprocessing
Specifically, for the selected historical traffic flow data, the weather influence factors and other feature vectors have different dimensions and larger numerical value difference, in order to prevent the characteristic factors from being submerged, data normalization processing is uniformly performed. The preprocessing formula is as follows:
Figure GDA0003019766240000061
in the formula: x' is a normalized value; x is an original sample value; x is the number ofmaxRepresents the maximum value of the sample; x is the number ofminRepresents the sample minimum;
step 102 kernel function selection
The introduction of the kernel function enables prediction under the nonlinear condition of LSSVM calculation to be more convenient, so that the solution of the decision function problem is converted into how to select a proper and effective kernel function.
Step 103: algorithm parameter initialization
Specifically, for improving the "Teaching-Learning" algorithm (Improvement Learning-Based Optimization algorithm, ITLBO): setting a problem solving dimension d (2), a maximum iteration number Mit (100) and a student population n (30); and setting a penalty factor c of a Least Square Support Vector Machine (LSSVM) and upper and lower limit values ub and lb of a kernel width parameter sigma (the upper and lower limit values of c are respectively 0 and 100, and the upper and lower limit values of sigma are respectively 0 and 100).
Step 104: randomly generating student score
For example, in [0,1 ]]Interval random generation of initial solution X of individual subject scores of students0(ii) a In (0,1)]Interval random initialization learning step length ri。
Step 105: calculating adaptive value by LSSVM
For example, the fitness value calculated by the LSSVM is regarded as the subject performance of "student", and the highest fitness value is regarded as "teacher".
Step 106: best sequencing for students of upper and lower generations
By adopting the improved 'teaching and learning' algorithm strategy provided by the invention, the students in the previous generation are associated with the current generation, excellent students in the two generations are obtained according to the mapping sequence, and the corresponding optimal adaptive value is stored through the calculation of the LSSVM.
Step 107: learning in the "teaching" stage
Specifically, the individual achievements of the students are evolved according to the formula (2) and the updating conditions thereof.
Figure GDA0003019766240000071
Figure GDA0003019766240000072
TF＝round[1+rand(0,1)] (4)
In the formula: r isiRepresents a learning step size of [0,1 ]]An interval random number;
Figure GDA0003019766240000073
represents the level of the teacher of j subjects; TF is a teaching factor and is randomly determined to be 1 or 2 by a formula (4); difference is the Difference between the average performance of the current students of the subject and the level of the teacher;
Figure GDA0003019766240000074
is the average of all students. When in use
Figure GDA0003019766240000075
Adapted value of greater than
Figure GDA0003019766240000076
And updating the student score. Namely:
Figure GDA0003019766240000077
Figure GDA0003019766240000078
End
step 108: learning in the "learning" stage
Specifically, two students X can be randomly selectedpAnd XqAnd (5) evolving the individual student according to the formula (5-6) and the updating condition thereof.
When X is presentp<XqThen, it indicates "student" XqScore comparison "student" XpIs excellent, so XpTo XqLearning, updating the formula as follows:
Figure GDA0003019766240000079
when X is presentp>XqThen, it indicates "student" XpScore comparison "student" XqIs excellent, so XpTo XqLearning, updating the formula as follows:
Figure GDA0003019766240000081
step 109: algorithm iteration termination determination
Specifically, if the maximum iteration time Mit is reached and the prediction error condition is met, the current optimal student subject performance, namely the optimal least square support vector machine punishment factor c and the nuclear parameter sigma combination, is saved. If the requirement is not met, the method goes to step 105 to recalculate the current adaptive value of the student individual and the subsequent steps of the adaptive value.
Step 110, LSSVM optimal short-time traffic flow prediction
And when the maximum iteration time Mit is reached and the prediction error condition is met, substituting the optimal parameter combination obtained in the step 109 into an LSSVM prediction model to predict the short-term traffic flow.
111, outputting the inverse normalization
Specifically, the output of the prediction algorithm of the ITLBO-LSSVM is denormalized by formula (7).
x＝xmin+(xmax-xmin)x' (7)
Step 112, predictive evaluation
And reasonably evaluating the prediction precision in order to verify the prediction accuracy of each algorithm. The method selects an average Absolute percentage Error (MAPE), a Relative Error (RE) and a Root Mean Square Error (RMSE) as prediction performance evaluation indexes. The description of the relative error and the average relative error to the error is more intuitive, and the root mean square relative error is mostly used as a comprehensive index for error analysis.
Figure GDA0003019766240000091
Figure GDA0003019766240000092
Figure GDA0003019766240000093
In the formula: y isiThe real value of the traffic flow is obtained; y'iRepresenting a traffic flow predicted value; n (36) is the traffic flow collection point number.
Fig. 2 is a schematic structural diagram illustrating a system for urban road short-time traffic flow prediction according to an embodiment of the present invention.
In the embodiments, the short-term traffic flow prediction method and system based on the improved teaching and learning algorithm optimization least square support vector machine (ITLBO-LSSVM) are provided. The method improves the teaching and learning algorithm, so that the relevance of the students to the generations is enhanced. Particularly, two adjacent generations of excellent student individuals are reserved during the optimal iteration and used for the next iteration, and the optimization iteration time is effectively shortened. The method comprises the steps of combining a penalty factor c of a least square support vector machine needing optimal iteration and a nuclear parameter sigma to serve as subjects learned by students, and using prediction accuracy calculated by the least square support vector machine as current subject scores of the students. And iterating to obtain the best subject performance combination of the students, namely the optimal c and sigma parameters of the least square support vector machine. The ITLBO-LSSVM method is used for predicting the short-term traffic flow in the North route of the Guiyang Jinyang in China, and experimental results show that the ITLBO-LSSVM method has excellent performance, good calculation efficiency and higher prediction precision.
The display 601 is used for displaying the prediction result; the input/output device 602 is used for inputting initialization parameters; the memory 604 stores instructions executable by the at least one processor 603, the instructions being executable by the at least one processor 603, while the system also includes a power supply device 605 for maintaining operation.
The foregoing is merely a detailed description of specific embodiments of the invention and is not intended to limit the invention. Various alterations, modifications and improvements will occur to those skilled in the art without departing from the spirit and scope of the invention.

Claims (8)

1. A method for urban road short-time traffic flow prediction, characterized in that the method comprises: data preprocessing, kernel function selection, algorithm parameter initialization, random data generation, adaptive value calculation, optimal sequencing, stage learning, iteration termination judgment, LSSVM (least squares support vector machine) based optimal short-term traffic flow prediction, inverse normalization output and prediction evaluation;
the data preprocessing specifically comprises the following steps: carrying out data normalization processing on the selected historical traffic flow data;
the algorithm parameter initialization specifically comprises the following steps: selecting an improved teaching and learning algorithm, wherein the improved teaching and learning algorithm comprises the settings of teachers and students, setting parameters, specifically setting the problem dimension d of the improved teaching and learning algorithm, the maximum iteration number Mit and the student group n, setting the upper limit value and the lower limit value of the punishment parameter c and the nuclear parameter sigma of the least square support vector machine LSSVM, and initializing the parameters of individual students;
the adaptive value calculation specifically comprises: calculating the adaptive value of the student score by using the LSSVM, and taking the highest adaptive value as a teacher;
the optimal sequence specifically includes: optimally ordering students of the next generation according to the adaptive values;
the stage learning specifically includes: learning in a teaching stage and learning in a learning stage;
the iteration termination judgment specifically comprises the following steps: judging according to the maximum iteration times and the prediction error condition;
the prediction evaluation specifically includes: selecting the average absolute percentage error MAPE, the relative error RE and the root mean square error RMSE as prediction performance evaluation indexes, wherein the specific formula is as follows:
Figure FDA0003019766230000011
Figure FDA0003019766230000012
Figure FDA0003019766230000013
in the formula: y isiThe real value of the traffic flow is obtained; y'iRepresenting a traffic flow predicted value; n is the traffic flow collection point number;
the improved teaching and learning algorithm comprises the following steps: associating students of the previous generation with the current generation, solving excellent students of the two generations to be ordered according to mapping, carrying out teaching and learning stage learning, and calculating and storing a corresponding optimal adaptive value through an LSSVM;
the teaching stage learning comprises the following steps: according to the formula
Figure FDA0003019766230000021
And the updating conditions thereof optimize the subject performance of the students when
Figure FDA0003019766230000022
Adapted value of greater than
Figure FDA0003019766230000023
When the student needs to be updated, the student subject scores are updated;
wherein, the Difference is the Difference between the average performance of the current students of the subject and the level of a teacher;
the learning stage comprises the following steps: randomly selecting two students XpAnd Xq：
When X is presentp<XqTime, show student XqScore comparing student XpIs excellent, so XpTo XqLearning according to a formula
Figure FDA0003019766230000024
Updating the student subject scores;
when X is presentp>XqTime, show student XpScore comparing student XqIs excellent, so XqTo XpLearning according to a formula
Figure FDA0003019766230000025
Updating the student subject scores;
wherein r isiRepresents a learning step size of [0,1 ]]And (4) interval random numbers.
2. The method of claim 1, wherein the algorithm parameter initialization comprises: randomly generating initial solution X of individual subject scores of students0(ii) a Random initialization learning step length ri。
3. The method according to claim 1, characterized in that it comprises: and (3) optimally setting a punishment parameter c and a kernel parameter sigma of the LSSVM according to an improved teaching and learning algorithm, taking the adaptive value calculated by the LSSVM as the subject performance of the student, and taking the adaptive value with the highest adaptive value as a teacher.
4. The method of claim 1, wherein Difference pass is
Figure FDA0003019766230000026
Calculating TF as teaching factor, and calculating by the formula TF as round [1+ rand (0,1)]Obtaining where riRepresents a learning step size of [0,1 ]]An interval random number;
Figure FDA0003019766230000027
represents the level of the teacher of j subjects;
Figure FDA0003019766230000028
average for all students; difference is the Difference between the average performance of current students in subject and the level of teachers.
5. The method of claim 1, wherein the improved teaching and learning algorithm solves the problem with a dimension d of 2.
6. The method of claim 1, wherein the maximum number of iterations Mit is 100 and the number of student groups is 30.
7. The method according to claim 1, wherein the update condition is:
Figure FDA0003019766230000031
8. a system for urban road short-time traffic flow prediction, characterized in that the system comprises: the system comprises a display, an input and output device, at least one processor, a memory in communication connection with the at least one processor, and a power supply device for supplying power;
wherein, the display is used for displaying the prediction result; the input and output equipment is used for inputting initialization parameters; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 7.
CN201811133463.XA 2018-09-27 2018-09-27 Method and system for urban road short-time traffic flow prediction Active CN109215344B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811133463.XA CN109215344B (en) 2018-09-27 2018-09-27 Method and system for urban road short-time traffic flow prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811133463.XA CN109215344B (en) 2018-09-27 2018-09-27 Method and system for urban road short-time traffic flow prediction

Publications (2)

Publication Number Publication Date
CN109215344A CN109215344A (en) 2019-01-15
CN109215344B true CN109215344B (en) 2021-06-18

Family

ID=64981819

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811133463.XA Active CN109215344B (en) 2018-09-27 2018-09-27 Method and system for urban road short-time traffic flow prediction

Country Status (1)

Country Link
CN (1) CN109215344B (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109785627A (en) * 2019-02-22 2019-05-21 魏一凡 A kind of crossroad access flux monitoring system
CN109816983A (en) * 2019-02-26 2019-05-28 昆明理工大学 A kind of short-term traffic flow forecast method based on depth residual error network
CN109949577B (en) * 2019-04-25 2021-07-27 贵州大学 Road traffic flow prediction method
CN110210664B (en) * 2019-05-29 2020-07-24 东南大学 Deep learning method for short-term prediction of using behaviors of multiple individual vehicles
CN110444022A (en) * 2019-08-15 2019-11-12 平安科技（深圳）有限公司 The construction method and device of traffic flow data analysis model
CN117373263B (en) * 2023-12-08 2024-03-08 深圳市永达电子信息股份有限公司 Traffic flow prediction method and device based on quantum pigeon swarm algorithm

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103150609A (en) * 2013-02-18 2013-06-12 健雄职业技术学院 Modeling method for short time traffic flow predicting model
CN103730006A (en) * 2014-01-26 2014-04-16 吉林大学 Short-time traffic flow combined forecasting method
CN106781465A (en) * 2016-12-06 2017-05-31 广州市科恩电脑有限公司 A kind of road traffic Forecasting Methodology
CN106837678A (en) * 2017-03-15 2017-06-13 大连大学 Based on the turbine-generator units PID governor parameters optimization for improving TLBO algorithms
CN108229352A (en) * 2017-12-21 2018-06-29 上海交通大学 A kind of standing detection method based on deep learning

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180152475A1 (en) * 2016-11-30 2018-05-31 Foundation Of Soongsil University-Industry Cooperation Ddos attack detection system based on svm-som combination and method thereof

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103150609A (en) * 2013-02-18 2013-06-12 健雄职业技术学院 Modeling method for short time traffic flow predicting model
CN103730006A (en) * 2014-01-26 2014-04-16 吉林大学 Short-time traffic flow combined forecasting method
CN106781465A (en) * 2016-12-06 2017-05-31 广州市科恩电脑有限公司 A kind of road traffic Forecasting Methodology
CN106837678A (en) * 2017-03-15 2017-06-13 大连大学 Based on the turbine-generator units PID governor parameters optimization for improving TLBO algorithms
CN108229352A (en) * 2017-12-21 2018-06-29 上海交通大学 A kind of standing detection method based on deep learning

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于改进教学优化算法的Hermite正交基神经网络混沌时间序列预测;李瑞国等;《物理学报》;20151031;第64卷(第20期);正文第1-13页 *

Also Published As

Publication number Publication date
CN109215344A (en) 2019-01-15

Similar Documents

Publication Publication Date Title
CN109215344B (en) Method and system for urban road short-time traffic flow prediction
CN110751318B (en) Ultra-short-term power load prediction method based on IPSO-LSTM
CN105354646B (en) Power load forecasting method for hybrid particle swarm optimization and extreme learning machine
CN108897719A (en) Meteorological data missing values complementing method based on self-adapted genetic algorithm
CN112733417B (en) Abnormal load data detection and correction method and system based on model optimization
CN109934422A (en) Neural network wind speed prediction method based on time series data analysis
CN111311001B (en) Bi-LSTM network short-term load prediction method based on DBSCAN algorithm and feature selection
CN111008790A (en) Hydropower station group power generation electric scheduling rule extraction method
CN113743538A (en) Intelligent building energy consumption prediction method, equipment and medium based on IPSO-BP neural network
CN116187835A (en) Data-driven-based method and system for estimating theoretical line loss interval of transformer area
CN115099511A (en) Photovoltaic power probability estimation method and system based on optimized copula
CN110097236A (en) A kind of short-term load forecasting method based on FA optimization Elman neural network
CN107301478A (en) A kind of cable run short-term load forecasting method
CN112348352B (en) Big data analysis-based automatic generation method for electric power budget proposal scheme
CN113762591B (en) Short-term electric quantity prediction method and system based on GRU and multi-core SVM countermeasure learning
CN115395502A (en) Photovoltaic power station power prediction method and system
CN115459982A (en) Power network false data injection attack detection method
CN112836876B (en) Power distribution network line load prediction method based on deep learning
CN113516163B (en) Vehicle classification model compression method, device and storage medium based on network pruning
CN115619028A (en) Clustering algorithm fusion-based power load accurate prediction method
CN115409317A (en) Transformer area line loss detection method and device based on feature selection and machine learning
CN115186882A (en) Clustering-based controllable load spatial density prediction method
CN105050096B (en) The complex network coverage method evolved based on Snowdrift game
CN113283638A (en) Load extreme curve prediction method and system based on fusion model
CN111626465A (en) New energy power short-term interval prediction method and system

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